Processing

Please wait...

Settings

Settings

Goto Application

1. WO2020109820 - MOLECULAR SIGNATURE

Note: Text based on automatic Optical Character Recognition processes. Please use the PDF version for legal matters

[ EN ]

MOLECULAR SIGNATURE

Field of the Invention

The present invention relates to a method of identifying a pre-invasive lung lesion in an individual that will progress to an invasive lung cancer comprising:

identifying a tissue as comprising a pre-invasive lung lesion that will progress to an invasive lung cancer where a molecular signature is present or identifying the tissue as comprising a pre-invasive lung lesion that will not progress to an invasive lung cancer where the molecular signature is absent.

The present invention also relates to a method of treating a pre-invasive lung lesion and/or treating and/or preventing an invasive lung cancer in an induvial comprising:

identifying a tissue as comprising a pre-invasive lung lesion that will progress to an invasive lung cancer where a molecular signature is present or identifying the tissue as comprising a pre-invasive lung lesion that will not progress to an invasive lung cancer where the molecular signature is absent.

The present invention further relates to a molecular signature, and uses thereof, for identifying whether or not an individual has a pre-invasive lung lesion that will progress to an invasive lung cancer.

Background to the Invention

Lung cancer is the most common cause of cancer death worldwide with 1.5 million deaths per year [1] Lung squamous cell carcinoma (LUSC) is the most common subtype in parts of Europe and second in the U.S.A [2]. Before progression to invasive LUSC, there is step-wise evolution of ever more disordered pre-invasive lesions, ranging from mild and moderate dysplasia (low-grade lesions) to severe dysplasia and carcinoma-in-situ (CIS; high-grade lesions) [3] The accessibility of the proximal airways allows detection and monitoring of these lesions using high-resolution diagnostic approaches such as autofluorescence bronchoscopy (AFB) [4] This technique enables the acquisition of tissue throughout the natural history of LUSC, providing an excellent model to study early tumorigenesis in human patients.

The molecular alterations that occur in cells before cancer is manifest are largely uncharted. Lung carcinoma-in-situ (CIS) lesions are the pre-invasive precursor to squamous cell carcinoma. While microscopically identical, their future is in equipoise with half progressing to invasive cancer and half regressing or remaining static. The cellular basis of this clinical observation is unknown.

Clinically, the optimal management of pre-invasive airway lesions remains unclear, despite the availability of surgery, radiotherapy and ablative techniques [5] AFB with biopsy allows assessment of the size, gross morphology and histopathology of pre-invasive lesions but cannot distinguish lesions that will ultimately progress to invasive tumours from those that will spontaneously regress. As such, indiscriminate surgical resection of pre-invasive lesions or external beam radiotherapy represents over treatment: lesions will spontaneously regress in 30% of cases, patient co-morbidity and poor lung function impart considerable risk, and the presence of field cancerization means independent lung cancers frequently emerge at sites outside resection or therapy margins [6]

Whilst molecular techniques have revolutionised the understanding of cancer biology, the key steps from normal cell to the point of cancer (uncontrolled growth and invasion) remain unclear. Thus, improved assays for the accurate diagnosis and management of lung CIS and/or lung cancer are sought and would be of significant clinical and economic benefit, particular assays which are minimally invasive.

Summary of the Invention

Provided is a new understanding of cancer precursor biology, based on a unique collection of high-grade pre-invasive lung lesions which were followed-up under conservative clinical management. Genomic, transcriptomic and epigenomic landscape of CIS have been profiled in a unique patient cohort with longitudinally monitored pre-invasive disease. Predictive modelling identifies which lesions will progress with remarkable accuracy. Progression- specific methylation changes on a background of widespread heterogeneity, alongside a strong chromosomal instability signature have been identified. Mutations and copy number changes characteristic of cancer and chart their emergence, offering a window into early carcinogenesis have also been identified.

This has enabled the provision of a novel molecular signature, with particular utility in the diagnosis of and monitoring of lung CIS and lung cancer. Methods of the invention allow more efficient patient risk stratification for the purposes of providing better treatment and/or to help plan and manage patient care.

The present disclosure delineates changes in the genomic architecture, genome wide gene expression and DNA methylation of pre-invasive cancers with known histological evidence of subsequent disease progression or regression. The CIS genome shares many of the hallmarks of advanced, invasive LUSC but marked genomic, transcriptomic and epigenetic differences exist between lesions that are benign and those that will progress to cancer. The disclosure demonstrate the use of these differences in predicting outcome over current clinical practice.

One of the pathways associated with progression is chromosomal instability (CIN), defined as a high rate of gain or loss of whole (or parts of) chromosomes. CIN is implicated in many human cancers, including lung, and has been suggested both as a prognostic marker and therapeutic target [30], [31]. Regressive lesions do not have the wholesale genomic instability of those that will progress and their epigenetic and transcriptional profiles more closely resemble normal bronchial epithelium than invasive cancers. Despite this, CIS lesions that spontaneously regress are genuine neoplasms; they harbour many somatic mutations, which can include known potential driver mutations. The mechanism of regression remains mysterious: it is unclear whether clones become exhausted and die out, potentially abetted by immune surveillance, or whether clones persist but phenotypically revert to an architecturally normal, physiological epithelium. Likewise the mechanisms of CIN are not well understood.

This disclosure represents the first whole genome sequencing data of pre-invasive lung lesions and offers the first insight into the molecular map of early lung squamous cancer pathogenesis, foretelling an era in which molecular profiling will enable personally tailored therapeutic decisions for patients with, for example, pre-invasive lung disease.

Thus, the invention provides a method of identifying whether or not an individual has a pre-invasive lung lesion that will progress to an invasive lung cancer, the method comprising:

(a) providing a sample of nucleic acid which has been taken from a tissue of the individual, wherein the tissue is suspected of harbouring a pre-invasive lung lesion;

(b) performing an assay to determine a progression score for the sample; and

(c) identifying whether or not the individual has a pre-invasive lung lesion that will progress to an invasive lung cancer by comparing the progression score to a threshold value;

wherein the progression score is determined using a molecular signature selected from:

i) a differentially expressed gene (DEG) signature; ii) a differentially methylated position (DMP) signature; iii) a copy number variation (CNV) signature; and

iv) combinations of (i) to (iii).

The individual may be identified as having a pre-invasive lung lesion that will progress to an invasive lung cancer if the progression score determined for the sample is higher than the threshold value; or

the individual may be identified as not having a pre-invasive lung lesion that will progress to an invasive lung cancer if the progression score determined for the sample is lower than the threshold value.

In some embodiments of the present invention, the DEG signature comprises the expression level of each of at least 10, at least 20, at least 30, at least 40, at least 50, at least 60, at least 75, at least 100, at least 150, at least 200, at least 250, at least 300, at least 325, least 350, at least 375, or 397 genes identified in Table 1, optionally wherein the method achieves an ROC AUC of at least 0.6, at least 0.60, at least 0.61, at least 0.62, at least 0.63, at least 0.64, at least 0.65, at least 0.66, at least 0.67, at least 0.68, at least 0.69, at least 0.7, at least 0.70, at least 0.71 , at least 0.72, at least 0.73, at least 0.74, at least 0.75, at least 0.76, at least 0.77, at least 0.78, at least 0.79, at least 0.8, at least 0.80, at least 0.81 , at least 0.82, at least 0.83, at least 0.84, at least 0.85, at least 0.86, or at least 0.87, optionally wherein the method achieves a sensitivity of at least about 95 % and/or a specificity of at least about 55 %, preferably wherein the threshold value is from about 0.02 to about 0.6.

The DEG signature may comprise the expression level of each of the genes identified in:

(i) Table 5, optionally wherein the threshold value is about 0.3, preferably wherein the method also achieves an ROC AUC of at least about 0.6 or about 0.64 and/or achieves a sensitivity of at least about 95 % and/or a specificity of at least about 55 %;

(ii) Table 4, optionally wherein the threshold value is about 0.035, preferably wherein the method also achieves an ROC AUC of at least about 0.65 or about 0.69 and/or achieves a sensitivity of at least about 95 % and/or a specificity of at least about 55 %;

(iii) Table 3, optionally wherein the threshold value is about 0.04, preferably wherein the method also achieves an ROC AUC of at least about 0.7 or about 0.76 and/or achieves a sensitivity of at least about 95 % and or a specificity of at least about 55 %;

(iv) Table 2, optionally wherein the threshold value is about 0.105, preferably wherein the method also achieves an ROC AUC of at least about 0.75 or about 0.81 and/or achieves a sensitivity of at least about 95 % and/or a specificity of at least about 55 %; or

(v) Table 1, optionally wherein the threshold value is about 0.14, preferably wherein the method achieves an ROC AUC of at least about 0.85 or about 0.87, optionally wherein the method achieves a sensitivity of at least about 95 % and/or a specificity of at least about 55 %.

In some embodiments of the methods of the present invention, the DMP signature comprises the methylation status (b) of least two, at least three, at least four, at least five, at least six, at least seven, at least eight, at least nine, at least 10, at least 12, least 14, at least 16, at least 18, at least 20, at least 25, at least 50, at least 75, at least 100, at least 125, or differentially methylated positions (DMPs) selected from Table 11, optionally wherein the method achieves an ROC AUC of at least 0.9, at least 0.90, at

least 0.91, at least 0.92, at least 0.93, at least 0.94, at least 0.95, at least 0.96, at least 0.97, at least 0.98, or at least 0.99, optionally wherein the method achieves a sensitivity of at least about 95 % and/or a specificity of at least about 50 %, preferably wherein the threshold value is from about 0.3 to about 0.6.

The DMP signature may comprise the b of each of the DMPs identified in:

(i) Table 16 or Table 17, optionally wherein the threshold value is about 0.43 or about 0.44, preferably wherein the method achieves an ROC AUC of at least about 0.9 or about 0.94 and/or achieves a sensitivity of at least about 95 % and/or a specificity of at least about 60 %;

(ii) Table 15, optionally wherein the threshold value is about 0.45, preferably wherein the method achieves an ROC AUC of at least about 0.93 or about 0.96 and/or achieves a sensitivity of at least about 95 % and/or a specificity of at least about 50 %;

(iii) Table 14, optionally wherein the threshold value is about 0.45, preferably wherein the method achieves an ROC AUC of at least about 0.95 or about 0.99 and/or achieves a sensitivity of at least about 95 % and/or a specificity of at least about 55 %;

(iv) Table 13, optionally wherein the threshold value is about 0.46, preferably wherein the method achieves an ROC AUC of at least about 0.95 or about 0.996 and/or achieves a sensitivity of at least about 95 % and/or a specificity of at least about 55 %;

(v) Table 12, optionally wherein the threshold value is about 0.48, preferably wherein the method achieves an ROC AUC of at least about 0.95 or about 0.999 and/or achieves a sensitivity of at least about 95 % and/or a specificity of at least about 55 %; or

(vi) Table 11, optionally wherein the threshold value is about 0.5, preferably wherein the method achieves an ROC AUC of at least about 0.95 or about 0.998 and/or achieves a sensitivity of at least about 95 % and/or a specificity of at least about 55 %.

In some embodiments of the methods of the present invention, the CNV signature comprises the amplification or loss of at least 5, at least 6, least 7, at least 8, at least 9, at least 10, at least 25, at least 50, at least 75, at least 100, at least 125, at least 150, at least 175, or at least 200 CNV bands identified in Table 19, optionally wherein the method achieves an ROC AUC of at least about 0.9, at least about 0.91, at least about 0.92, at least about 0.93, at least about 0.94, at least about 0.95, at least about 0.96, at least about 0.97, at least about 0.98, at least about 0.99, optionally wherein the method achieves a sensitivity of at least about 75 %, at least about 80 %, at least about 85 %, at least about 90 %, or at least about 95 %, and/or a specificity of at least about 65 %, at least about 70 %, at least about 15 %, at least about 80 %, at least about 85 %, at least about 90 %, or at least about 95 %, preferably wherein the threshold value is from about 0.05 to about 0.6.

The CNV signature may comprises the amplification or loss of the CNV bands identified in:

(i) Table 34, optionally wherein the threshold value is about 0.66, preferably wherein the method achieves an ROC AUC of at least about 0.9 or about 0.95 and/or achieves a sensitivity of at least about 95 % and/or a specificity of at least about 65 %;

(ii) Table 33, optionally wherein the threshold value is about 0.65, preferably wherein the method achieves an ROC AUC of at least about 0.9 or about 0.95 and/or achieves a sensitivity of at least about 95 % and/or a specificity of at least about 70 %;

(iii) Table 32, optionally wherein the threshold value is about 0.6, preferably wherein the method achieves an ROC AUC of at least about 0.9 or about 0.95 and/or achieves a sensitivity of at least about 95 % and/or a specificity of at least about 85 %;

(iv) Table 31, optionally wherein the threshold value is about 0.56, preferably wherein the method achieves an ROC AUC of at least about 0.9 or about 0.96 and/or achieves a sensitivity of at least about 95 % and/or a specificity of at least about 85 %;

(v) Table 30, optionally wherein the threshold value is about 0.48, preferably wherein the method achieves an ROC AUC of at least about 0.95 or about 0.96 and/or achieves a sensitivity of at least about 90 % and or a specificity of at least about 90 %;

(vi) Table 29, optionally wherein the threshold value is about 0.41, preferably wherein the method achieves an ROC AUC of at least about 0.95 or about 0.96 and/or achieves a sensitivity of at least about 90 % and/or a specificity of at least about 90 %;

(vii) Table 28, optionally wherein the threshold value is about 0.31 , preferably wherein the method achieves an ROC AUC of at least about 0.95 or about 0.95 and/or achieves a sensitivity of at least about 95 % and/or a specificity of at least about 90 %; (viii) Table 27, optionally wherein the threshold value is about 0.19, preferably wherein the method achieves an ROC AUC of at least about 0.95 or about 0.97 and/or achieves a sensitivity of at least about 80 % and/or a specificity of at least about 95 %;

(ix) Table 26, optionally wherein the threshold value is about 0.12, preferably wherein the method achieves an ROC AUC of at least about 0.95 or about 0.97 and/or achieves a sensitivity of at least about 90 % and/or a specificity of at least about 95 %;

(x) Table 25, optionally wherein the threshold value is about 0.06, preferably wherein the method achieves an ROC AUC of at least about 0.95 or about 0.98 and/or achieves a sensitivity of at least about 90 % and/or a specificity of at least about 90 %;

(xi) Table 24, optionally wherein the threshold value is about 0.02, preferably wherein the method achieves an ROC AUC of at least about 0.95 or about 0.98 and/or achieves a sensitivity of at least about 90 % and/or a specificity of at least about 95 %;

(xii) Table 23, optionally wherein the threshold value is about 0.01 , preferably wherein the method achieves an ROC AUC of at least about 0.95 or about 0.98 and/or achieves a sensitivity of at least about 95 % and/or a specificity of at least about 95 %;

(xiii) Table 22, optionally wherein the threshold value is about 0.01, preferably wherein the method achieves an ROC AUC of at least about 0.95 or about 0.98 and/or achieves a sensitivity of at least about 95 % and/or a specificity of at least about 90 %;

(xiv) Table 21, optionally wherein the threshold value is about 0.02, preferably wherein the method achieves an ROC AUC of at least about 0.95 or about 0.98 and/or achieves a sensitivity of at least about 95 % and/or a specificity of at least about 90 %;

(xv) Table 20, optionally wherein the threshold value is about 0.01, preferably wherein the method achieves an ROC AUC of at least about 0.95 or about 0.98 and/or achieves a sensitivity of at least about 95 % and/or a specificity of at least about 90 %; or

(xvi) Table 19, optionally wherein the threshold value is about 0.02, preferably wherein the method achieves an ROC AUC of at least about 0.95 or about 0.98 and/or achieves a sensitivity of at least about 95 % and/or a specificity of at least about 85 %.

The present invention also provides a method of treating a pre-invasive lung lesion and/or treating and/or preventing an invasive lung cancer in an individual comprising:

identifying a pre-invasive lung lesion in an individual that will progress to an invasive lung cancer by performing a method described herein for identifying whether or not an individual has a pre-invasive lung lesion that will progress to an invasive lung cancer; and

administering a lung cancer therapy to the individual, optionally wherein the therapy comprises surgical intervention.

The present invention also provides a method for determining whether or not to provide a therapeutic method of treatment to an individual, the method comprising performing a method described herein for identifying whether or not an individual has a pre-invasive lung lesion that will progress to an invasive lung cancer;

providing to the individual a therapeutic method of treatment if the individual is identified as having a pre-invasive lung lesion that will progress to an invasive lung cancer, optionally wherein the therapeutic method of treatment comprises administering to the individual a therapeutically effective amount of a lung cancer therapy.

In any of the methods of the present invention described herein, the sample may be from an individual who:

(a) is not suspected of having cancer;

(b) is suspected of having a pre-invasive lung lesion but not suspected of having cancer;

(c) has a pre-invasive lung lesion but is not suspected of having cancer;

(d) has a pre-invasive lung lesion and is suspected of having cancer;

(e) is suspected of having cancer; or

(f) has cancer.

In any of the methods of the present invention described herein, the tissue from which the sample of nucleic acid has been taken may:

(i) have been obtained from a biopsy;

(iii) be processed by laser-capture micro-dissection (LCM); and/or

(ii) be fresh-frozen tissue or formalin-fixed paraffin-embedded (FFPE) tissue.

The nucleic acid may be a DNA. In any of the methods of the present invention described herein, the progression score may be determined using the PAM method.

Where the progression score is determined using a DEG signature, the assay in step (b) may comprise performing a hybrisation assay. Where the progression score is determined using a DMP signature, the assay in step (b) may comprise bisulphite conversion of the DNA, optionally step (b) may comprise performing a sequencing step to determine the sequence of the DNA molecules, optionally wherein before sequencing an amplification step is performed; and/or optionally wherein step (b) comprises

(i) hybridising the DNA to an array comprising probes capable of discriminating between methylated and non- methylated forms of DNA and applying a detection system to the array to discriminate methylated and non-methylated forms of DNA, optionally wherein before hybridisation an amplification step is performed; or

(ii) performing an amplification step using methylation-specific primers, wherein the methylation status of the DNA is determined by the presence or absence of an amplified product, optionally wherein the amplification step is performed by PCR.

Where the progression score is determined using a CNV signature, the assay in step (b) may comprise whole genome sequencing.

The nucleic acid may be an RNA. Where the nucleic acid is an RNA, the progression score may be determined using a DEG signature and the assay in step (b) may comprise reverse transcriptase quantitative PCT (RT- qPCR).

The present invention also provides a molecular signature as described herein for identifying whether or not an individual has a pre-invasive lung lesion that will progress to an invasive lung cancer.

The present invention further provides use of a molecular signature as described herein for identifying whether or not an individual has a pre-invasive lung lesion that will progress to an invasive lung cancer. The use of the molecular signature may be in an ex-vivo method for identifying whether or not an individual has a pre-invasive lung lesion that will progress to an invasive lung cancer. Further provided by the present invention are the in vitro, in silico or ex-vivo use of a molecular signature as described herein for identifying whether or not an individual has a pre-invasive lung lesion that will progress to an invasive lung cancer.

In any of the methods, molecular signatures or uses thereof provided by the present invention described herein, the pre-invasive lung lesion may be a solid lesion and/or the invasive lung cancer may be a solid tumour. In any of the methods, molecular signatures or uses thereof provided by the present invention, the pre-invasive lung lesion may be normal epithelium, tissue hyperplasia, dysplasia, or lung carcinoma in situ (CIS). In any of the methods, molecular signatures or uses thereof provided by the present invention, the lung cancer may be lung squamous cell carcinoma (LUSC).

Brief description of the Figures

Figure 1. Demographic and clinical characteristics of patients in the whole-genome sequencing, methylation discovery and validation, and gene expression discovery and validation datasets.

Figure 2. Analysis of pre-invasive lung carcinoma- in-situ (CIS) lesions. (A) Detection of bronchial pre-invasive CIS lesions by autofluorescence bronchoscopy. (B)

Histological outcomes of bronchial pre-invasive lesions. (C) Overview of the study protocol. Patients with identified CIS lesions underwent repeat bronchoscopy and re biopsy every 4 months. Definitive cancer treatment was only performed if pathological evidence of progression to invasive cancer was detected. The‘index biopsy’ profiled in this study refers to the biopsy immediately preceding progression to invasive cancer or regression to low-grade dysplasia or normal epithelium. (D) Venn diagram of different -omics analyses performed on laser capture microdissection (LCM) -captured CIS lesions. Due to the small size of bronchial biopsies, not all analyses were performed on all samples.

Figure 3. Genomic aberrations in pre-invasive lung carcinoma-in-situ (CIS) lesions. Circos diagram comparing CIS genomic profiles with The Cancer Genome Atlas (TCGA) LUSC data. The outer histogram (A), shows mutation frequencies of all genes in TCGA data. The inner histogram (D) shows mutation frequencies in the CIS data presented herein. Profiles appear similar and no statistically significant differences were identified between the two datasets. Genes previously identified as potential drivers of lung cancer are labelled. Between the two histograms, average copy number changes are shown for TCGA data (B) and CIS data (C). Copy number gains are shown in red, losses in blue. Although there were some differences between whole-genome and whole-exome sequencing techniques, it was observed that many similar features between the two; for example, gains in 3q and 5p, which are well recognised features of squamous cell lung cancer in the centre of the circos plot, 39 rings represent the copy number profiles of the 39 samples descried herein, illustrating the individual contribution of each sample to the average values presented (E).

Figure 4. Altered methylation and gene expression in lung carcinoma-in-situ (CIS) lesions. (A) Hierarchical clustering of 1335 significantly differentially expressed genes in progressive and regressive CIS lesions. Biological and clinical factors including age at diagnosis, gender, smoking history (pack years) and COPD status had no effect on CIS lesion gene expression profile (high expression = purple, e.g., the trend in the top left quadrant of (A), low expression = orange). (B) Hierarchical clustering of the top 1000 significantly differentially methylated positions (DMPs) between progressive and regressive CIS lesions and controls. Biological and clinical factors including age at diagnosis, gender and smoking history (pack years) status had no effect on the methylation profile (hypomethylated DMPs = blue, hypermethylated DMPs = orange, e.g., the trend in the bottom right quadrant of (B)). (C) Principle component analysis of all profiled genes in progressive and regressive CIS lesions showing a clear distinction between progressive and regressive groups (p=0.0017). (D) Principle component analysis of all methylation data in progressive, regressive and control CIS lesions showing a clear distinction between progressive and regressive groups (p=6.8x10-25). P values were calculated using multivariate ANOVA.

Figure 5. Carcinoma-in-situ (CIS) gene expression and methylation profiles are predictive of progression to cancer. (A) Probability plot based on a 291 -gene signature for correct class prediction (discovery set - red circles indicate progressive lesions, e.g., top right; green circles indicate regressive lesions, e.g., bottom left). (B) Challenging the 291 -gene signature on a CIS validation set. Area under the curve (AUC) is 1 using Receiver Operating Characteristic (ROC) analysis. (C) Application of the 291 -gene signature to TCGA LUSC data. The signature described herein classified TCGA LUSC vs TCGA controls samples with AUC of 0.81 (green circles indicate TCGA controls (left portion of graph), orange circles indicate TCGA LUSC (right portion of graph).

(D) Distribution of methylation beta values across the genome in TCGA controls, CIS regressive and progressive and TCGA LUSC samples. Most probes are regulated at 0 or 1 in normal tissue but this regulation is reduced in both regressive and progressive CIS and TCGA LUSC samples. (E) Methylation Heterogeneity Index (MHI), defined as counts of methylation probes with 0.26 < b < 0.88, for each sample. MHI is higher in regressive and progressive CIS and TCGA LUSC compared with TCGA controls and this can be used as an accurate predictor with AUC=0.96 for TCGA LUSC vs TCGA controls and AUC=0.74 for progressive vs regressive CIS. (F) Histogram of AUC values calculated by performing the same analysis used in (E) 10,000 times, with each run limited to a different random sample of 2,000 probes (AUC mean for TCGA LUSC vs TCGA controls is 0.95 (95% Cl 0.92-0.98)). This demonstrates that a random sample of methylation probes is an accurate predictor using this method.

Figure 6. Chromosomal instability is associated with progression to cancer. (A) Mean expression of CIN-associated genes in CIS samples. Progressive and regressive CIS samples are well differentiated with AUC=0.96. Green circles indicate regressive CIS lesions; red circles indicate progressive CIS. (B) Plot of NEK2 expression across CIS samples demonstrates increasing expression with progression to cancer. Expression of this gene alone classifies progressive vs regressive CIS with AUC=0.93. (C) Pathway analysis of gene expression data between progressive and regressive CIS shows a strong chromosomal instability (CIN) signal. This signal remains strong when cell cycle genes are removed from the CIN70 signature. (D) Pathway analysis of methylation data demonstrating several cancer-related pathways up-regulated in progressive CIS compared with regressive CIS.

Figure 7. Experimental workflow. Flow diagram illustrating which profiling techniques were applied to which samples. Biopsies taken from index CIS lesions were stored as fresh frozen (FF) and formalin-fixed paraffin embedded (FFPE). DNA was extracted from FF biopsies. The first 54 samples studied that had sufficient extracted DNA passing quality control (QC) underwent first methylation profiling, then whole-genome sequencing (WGS) when sufficient remaining DNA was available. Due to the low DNA quantity extracted from some biopsies, the methylation dataset (n=54) was larger than the WGS data set (n=29), therefore the subsequent 10 samples underwent WGS directly without methylation profiling. RNA was extracted from FFPE samples and underwent gene expression profiling when RNA passed QC. To ensure validity of our conclusions across orthogonal platforms we used Illumina microarrays to profile a discovery set of 33 samples, then subsequently used Affymetrix microarrays to profile an independent validation set of 18 further samples.

Figure 8. Mutational signatures of carcinoma-in-situ (CIS) lesions. (A-D) The contribution of each of five pre-selected mutational signatures to each lesion is shown. These five mutational signatures, associated with CpG deamination (1), APOBEC (2 and 13), tobacco (4) and unknown aetiology (5), were selected based on an initial run using all 30 mutational signatures, which showed that these were present in the data and in signature extractions from lung squamous cell cancer (LUSC) datasets. The number of substitutions attributed to each signature is shown (A-B) as well as the proportion of mutations attributed to each mutational signature (C-D). Samples from the same patient share the same identifier except for the final letter; for example, PD21883a and

PD21883d are two samples from the same patient (e) Comparison of the mutational signatures of CIS lesions to those found in lung squamous cell cancer (LUSC). LUSC data were downloaded from TCGA and mutations called with our algorithms. All mutations from all samples from each cancer type were pooled for this analysis. The colour scale indicates the proportion of substitutions in each sample that are attributed to each signature. (F-J) Comparison of the relative proportion of mutations attributed to each signature between progressive (right-hand side) and regressive (left-hand side) CIS samples. P values were calculated using likelihood ratio tests of a mixed effects model with outcome (progressive or regressive) included as a fixed effect versus a model that was identical but for the fact that outcome was not included as a fixed effect. Only signature 4 (smoking-associated) was significantly different between the two groups.

Figure 9. Genome-wide copy number changes of carcinoma-in-situ (CIS) lesions. Visualisation of copy number changes for 39 whole-genome-sequenced CIS samples. Rows represent samples, genomic position is represented on the x-axis. Local copy number gains are illustrated in red, losses in blue. Widespread changes were observed in progressive CIS samples and a subset of regressive samples.

Figure 10. Documentation of biopsy history and chronology of lesion appearance in three misclassified regressive cases. (A) Case 1 (PD21893a) appeared to regress from a CIS lesion (07/2012) to squamous metaplasia (SqM; 1 1/2012). However, again, CIS was subsequently reconfirmed by biopsy (05/2013). (B) Case 2 (PD21884a) had a lobectomy for T1N0 lung squamous cell cancer (LUSC) in the left upper lobe (LUL) and was under surveillance for carcinoma-in-situ (CIS) at the resection margins. A subsequent, high-grade CIS lesion (08/2009) profiled for genome-wide DNA

methylation changes was considered regressive since a follow-up biopsy on the same anatomical site demonstrated the presence of a low-grade, moderately dysplastic (MoD) lesion (11/2009). A subsequent biopsy, however, was classified as CIS (02/201 1) and the lesion then remained static for 26 months but eventually progressed into invasive cancer (04/2014). (C) Case 3 (PD38326a) had an initial diagnosis of CIS (1 1/2015) followed by regression to normal epithelium (03/2016). CIS was subsequently identified at the same site (03/2017), with invasive cancer diagnosed on subsequent biopsy (07/2017).

Figure 11. Genomic aberrations in pre-invasive lung carcinoma-in-situ (CIS) lesions. Comparisons of the number of substitutions (A), small insertions and deletions (B), genome rearrangements (C) and copy number changes (D), showing significantly more genomic changes in progressive than regressive lesions. Although there were more

clonal substitutions in progressive than regressive lesions (E), the proportion of substitutions that were clonal and the number of clones were similar (F-G). Progressive lesions had more putative driver mutations (H). Telomere lengths (base pairs) were similar between the two groups (E). All P values were calculated using likelihood ratio tests of a mixed effects model with outcome (progressive or regressive) included as a fixed effect versus a model that was identical but for the fact that outcome was not included as a fixed effect.

Figure 12. Subclonal mutational structure in progressive and regressive CIS lesions. Heatmap showing the proportion of overlapping mutations between samples taken from the same patient. For four patients with lesions that would ultimately progress to cancer (denoted‘P’), over half the mutations were shared between any two given samples, suggesting that the lesions were derived from a common ancestral clone. By contrast, for two patients with lesions that would ultimately regress (denoted‘R’), almost no mutations were shared, suggesting that the lesions arose independently. Samples from the same patient are shown in the same colour; PD38321a and PD38322a do belong to the same patient and were mislabelled during processing.

Figure 13. Differential molecular changes between progressive and regressive lesions. Visualisation of differential changes across the genome. (A) shows all identified differentially methylated regions (DMRs) (hypermethyl ated regions in yellow, hypomethylated in blue) alongside a similar analysis comparing cancer and control samples from The Cancer Genome Atlas. It was observed that 58% of DMRs identified in the progressive vs regressive analysis are also identified in cancer vs control. (B) shows copy number changes across the genome in regressive CIS, progressive CIS and TCGA cancer samples. Congruency of copy number change was observed, suggesting similar processes in the two cohorts.

Figure 14. Principal component analysis investigating effect of various biological, clinical and technical factors affecting correct case segregation for all differentially methylated positions (DMPs) and gene expression data. (A-F) Principal component

analysis for all DMPs. (A) Smoking history (pack years). (B) Chronic obstructive pulmonary disease (COPD) status. (C) Previous lung cancer history referring to the presence of lung squamous cell cancer (LUSC) prior to identification of pre-invasive lesions. (D) Age at bronchoscopy (years); age of individual when pre-invasive lesion was first biopsied. (E) Gender. (F) Sentix ID. (G-K) Principal component analysis for all gene expression data. (G) Smoking history (pack years). (H) Chronic obstructive pulmonary disease (COPD) status. (I) Previous lung cancer history referring to the presence of lung squamous cell cancer (LUSC) prior to identification of pre-invasive lesions. (J) Age at bronchoscopy (years); age of individual when pre-invasive lesion was first biopsied. (K) Gender. P-values were calculated using multivariate ANOVA.

Figure 15. ROC analytics of gene expression predictive model. ROC and precision-recall curves for the predictive model based on gene expression data shown in Figure 5A-C. Curves are shown for the CIS discovery set (A-B), CIS validation set (C-D) and application to TCGA LUSC data (E-F).

Figure 16. Predictive modelling of methylation data. In addition to the predictive modelling based on probe variation shown in Figure 6, differentially expressed methylation probes were used to create a predictor using a Prediction Analysis for Microarrays (PAM) method. The model was trained on a training set (A-C) consisting of 26 progressive samples, 11 regressive samples and 23 control samples, shown in red, green and blue, respectively. A predictor based on 141 DMPs was created. This was applied to a validation set of 10 progressive, 7 regressive and 10 control samples (D-F), predicting outcome with AUC=0.99. (G-I) Application of the predictive model to TCGA methylation data. Samples were correctly classified into TCGA LUSC and TCGA control samples with AUC=0.99. (J-M) ROC analytics and precision-recall curves for Methylation Heterogeneity Index (MHI) model presented in Figure 5.

Curves apply to cancer vs control (J-K) and progressive vs regressive (L-M), respectively. (N) Histogram of AUC values using MHI model with random samples of 2000 probes, applied to progressive vs regressive data. This demonstrates that a similar AUC is achieved with a random sample of probes as when using the entire array.

Figure 17. Predictive modelling of copy number alteration (CNA) data. Using an analogous method to gene expression and methylation copy number data derived from methylation arrays was used to predict lesion outcome. Probe-level copy number changes were aggregated over cytogenetic bands; these data were used as input to Prediction Analysis of Microarrays (PAM). (A-C) Probability plot based on a 154 cytogenetic band signature for correct class prediction (red circles indicate progressive lesions, green circles indicate regressive lesions). The area under the curve for the 154-cytogenetic band signature is 0.86. (D-F) Application of the predictive model to previously published data (van Boerdonk et al.) replicates those result, classifying all regressive and 9/12 progressive samples correctly. This dataset included pre-invasive samples of various histological grades, rather than only CIS. (G-I) Application of the predictive model to TCGA copy number data. Samples were correctly classified into TCGA LUSC and TCGA control samples with an AUC of 0.98.

Figure 18. wGII score correlates with mean CIN gene expression. To confirm an association between CIN gene expression and copy number change, Weighted Genome Integrity Index (wGII) was correlated with mean CIN gene expression for the 1 1 CIS samples where gene expression and whole-genome sequencing data was available. Pearson correlation coefficient r2=0.473.

Detailed Description of the Invention

The early detection and treatment of pre-invasive lung disease, and the prevention of invasive lung cancers, such as LUSC, remains a major unmet need. The present inventors have extensively profiled the genome-wide gene expression and DNA methylation patterns associated with progressive pre-invasive lung disease as compared to regressive pre-invasive lung disease. Using these results, the present inventors have discovered molecular signatures that can be used to predict the clinical outcome of a pre-invasive lung lesion with a high degree of accuracy. The predictive molecular signatures provided by the present invention can be used to identify whether or not an individual has a progressive pre-invasive lung lesion i.e. a pre-invasive lesion that will progress or develop to an invasive lung cancer. Such an individual will benefit from preventative treatment of the pre-invasive lung lesion and/or treatment of the resultant invasive lung cancer. Conversely, the predictive molecular signatures provided by the present invention can be used to identify an individual as having a regressive pre-invasive lesion i.e a pre-invasive lesion that will not progress or develop into an invasive lung cancer. Such an individual would not benefit from preventative treatment of the pre-invasive lung lesion and/or treatment for lung cancer and therefore treatment and potentially harmful side-effects of e.g. chemotherapy / radiotherapy can be avoided.

The present invention relates to methods of identifying whether or not an individual has a pre-invasive lung lesion that will progress to an invasive lung cancer.

The present invention relates to a method of treating a pre-invasive lung lesion and/or treating and/or preventing an invasive lung cancer in an individual comprising: identifying a pre-invasive lung lesion in an individual that will progress to an invasive lung cancer by performing a method according to any one of claims 1-1 1; and administering a lung cancer therapy to the individual, optionally wherein the therapy comprises surgical intervention.

The present invention also relates to a molecular signature described herein for identifying whether or not an individual has a pre-invasive lung lesion that will progress to an invasive lung cancer.

The present invention further relates to the use of a molecular signature described herein for identifying whether or not an individual has a pre-invasive lung lesion that will progress to an invasive lung cancer.

Definitions

The term“ comprises” (comprise, comprising) should be understood to have its normal meaning in the art, i.e. that the stated feature or group of features is included, but that the term does not exclude any other stated feature or group of features from also being present. For example, a method comprising steps (a), (b) and (c) includes steps (a), (b) and (c) but may also include other steps.

The term“ consists of should also be understood to have its normal meaning in the art, i.e. that the stated feature or group of features is included, to the exclusion of further features. For example a method consisting of steps (a), (b) and (c) includes steps

(a), (b) and (c) and no other steps.

For every embodiment in which“ comprises” or“ comprising” is used, the present invention provides a further embodiment in which“ consists of’ or“ consisting of is used. Thus, every disclosure of“ comprises" should be considered to be a disclosure of“ consists of.

Individual

In any of the methods of the invention disclosed herein, the term“ individual’ may be a human. The most preferred individual to which the methods of the invention are applicable are humans.

In any of the methods of the invention disclosed herein, the individual may be a non-human animal. For example, methods of the invention disclosed herein may be applied to non-human animals to determine the efficacy of new therapeutics, new therapeutic strategies, new modes of administration of pre-existing therapeutic strategies, or surgical methods. Thus, in any of the methods of the invention disclosed herein the individual may be a rodent, such as a rat or a mouse. In any of the methods of the invention disclosed herein, the individual may be a non-human mammal, such as a primates, cats or pigs.

In any of the methods of the invention described herein, the individual can be one who:

(a) is not suspected of having cancer;

(b) is suspected of having a pre-invasive lung lesion but not suspected of having cancer;

(c) has a pre-invasive lesion but is not suspected of having cancer;

(d) has a pre-invasive lesion and is suspected of having cancer;

(e) is suspected of having cancer; or

(f) has cancer.

The individual may be suspected of having a pre-invasive lung lesion and/or cancer on the basis of a clinical presentation, a diagnostic test and/or family history.

The individual may have previously had a lung disease, a pre-invasive lung lesion

and/or cancer. The individual may be in remission from a lung disease, a pre-invasive lung lesion and/or cancer e.g. an invasive lung cancer, such as LUSC. The individual may be, or have been, a smoker. The individual may be a non-smoker. The individual may be male. The individual may be female. The individual may be an infant. The individual may be an adult. The individual may be elderly.

Samyle

The methods of the invention described herein comprise the step of providing a sample of nucleic acid which has been taken from a tissue of the individual, wherein the tissue is suspected of harbouring a pre-invasive lung lesion.

The“ nucleic acid sample which has been take from a tissue” may refer to a sample of nucleic acid which has been obtained from a material derived from a tissue, for example from whole blood, a blood fraction, plasma, serum, a bloodspot, a lung tissue biopsy, a lung tissue sample, lung mucus, sputum, or phlegm. Said sample of nucleic acid may be processed in any way that the user deems appropriate, such that a progression score can be determined using a molecular signature described herein.

The“ nucleic acid' may be a DNA. The“ nucleic acid' may be an RNA, such as a messenger RNA (mRNA).

Statistical parameters for predictive molecular signature

Sensitivity and specificity metrics for identification of pre-invasive lung lesions that will progress to an invasive lung may be defined using standard receiver operating characteristic (ROC) statistical analysis. In ROC analysis, 100 % sensitivity

corresponds to a finding of no false negatives, and 100 % specificity corresponds to a finding of no false positives.

As used herein the term“ sensitivity” (also referred to as the true positive rate) refers to a measure of the proportion of actual positives that are correctly identified as such. In other words, the sensitivity of a diagnostic test may be expressed as the number of true positives i.e. individuals correctly identified as having a disease as a proportion of all the individuals having the disease in the test population {i.e. the sum of true positive and false negative outcomes). Thus, a high sensitivity diagnostic test is desirable as it rarely misidentifies individuals having the disease. This means that a negative result obtained by a highly sensitive test has a high likelihood of ruling out the disease.

In the field of medical diagnostics, and as used herein, the term“ specificity”

(also referred to as the true negative rate) refers to a measure of the proportion of actual negatives that are correctly identified as such. In other words, the specificity of a diagnostic test may be expressed as the number of true negatives (i.e. healthy individuals correctly identified as not having a disease) as a proportion of all the healthy individuals in the test population (i.e. the sum of true negative and false positive outcomes). Thus, a high specificity diagnostic test is desirable as it rarely misidentifies healthy individuals. This means that a positive result obtained by a highly specific test has a high likelihood of ruling in the disease.

In the field of medical diagnostics, and as used herein, a“ Receiver Operating Characteristic (ROC) curve” refers to a plot of true positive rate (sensitivity) against the false positive rate (1 - specificity) for all possible cut-off values. These terms are well known in the art and to the skilled person.

The specificity and/or sensitivity of a method may be determined by performing said method on a validation set of samples. For samples in the validation set it is known which samples are positive samples e.g. samples derived from pre-invasive lung lesions known to have progressed to an invasive lung cancer, such as a CIS that progressed to a LUSC. It is also know which samples of the validation set are negative samples e.g. samples derived from pre-invasive lung lesions which did not progress to an invasive cancer (i.e. pre-invasive lung lesions that regressed). The extent to which the method correctly identifies the known positive samples (i.e. the sensitivity / true positive rate of the method) and/or the known negative samples (i.e. the specificity / true negative rate of the method) can thus be determined.

A further metric which can be employed to classify the accuracy of the methods of the present invention is ROC AUC. In ROC analysis, the area under the curve of a ROC plot (AUC) is a metric for binary classification. In a random binary classifier the number of true positives and false positives will be approximately equal. In this situation the AUC score for the ROC plot will be 0.5. In a perfect binary classifier the number of true positives will be 100 % and the number of false positives will be 0 %.

In this situation the AUC score for the ROC plot will be 1 which is therefore the highest AUC score a predictive classifier can achieve.

Predictive molecular signatures

The present invention relates to a method of identifying whether or not an individual has a pre-invasive lung lesion that will progress to an invasive lung cancer, the method comprising:

(a) providing a sample of nucleic acid which has been taken from a tissue of the individual, wherein the tissue is suspected of harbouring a pre-invasive lung lesion;

(b) performing an assay to determine a progression score for the sample; and

(c) identifying whether or not the individual has a pre-invasive lung lesion that will progress to an invasive lung cancer by comparing the progression score to a threshold value;

wherein the progression score is determined using a molecular signature selected from:

i) a differentially expressed gene (DEG) signature; ii) a differentially methylated position (DMP) signature;

iii) a copy number variation (CNV) signature; and

iv) combinations of (i) to (iii).

Thus, in some of the methods of the present invention described herein, the individual is identified as having a pre-invasive lung lesion that will progress to an invasive lung cancer based on a comparison of a progression score and a threshold value.

For example, the individual may be identified as having a pre-invasive lung lesion that will progress to an invasive lung cancer if the progression score determined for the sample is higher than the threshold; or the individual may be identified as not having a pre-invasive lung lesion that will progress to an invasive lung cancer if the progression score determined for the sample is lower than the threshold.

In any of the methods of the present invention described herein, the progression score may be determined using any of the molecular signatures described herein and

combinations thereof, for example any of the DEG signatures provided in Tables 1-9 and/or any of the DMP signatures provided in Tables 11 -17 and/or any of the CNV signatures provided in Tables 19-36. The gene weights provided in Tables 1-9 and/or the DMP weights provided Tables 11 -17 and/or the CNV band weights provided in Tables 19-36 may be used as part of the molecular signature to determine a progression score for a sample.

Thus, the present invention provides a molecular signature for use in identifying whether or not an individual has a pre-invasive lung lesion that will progress to an invasive lung cancer. The molecular signature of the present invention may be a DEG signature as defined in any one of Tables 1-9 or combinations thereof. The molecular signature of the present invention may be a DMP signature as defined in any one of Tables 11-17 or combinations thereof. The molecular signature of the present invention may be a CNV signature as defined in any one of Tables 19-36 or combinations thereof.

As used herein, a“ progression score’’ is a measure of the probability that an individual has a pre-invasive lung lesion that will progress to an invasive lung cancer.

In the context of the present invention, a progression score has value between 0 and 1.

The PAM (prediction analysis of microarray data) method may be used to generate a progression score from differential gene expression data using a DEG signature described herein, from differential methylation data using a DMP signature described herein, or from copy number band variation data using a CNV signature described herein. Thus, in any of the methods of the present invention described herein, the progression score may be determined using the PAM method. In any of the methods of the present invention described herein, the progression score may be determined using the PAM method and molecular signature selected from:

i) a differentially expressed gene (DEG) signature; ii) a differentially methylated position (DMP) signature;

iii) a copy number variation (CNV) signature; and

iv) combinations of (i) to (iii).

The PAM method is described Tibshirani, et al. " Diagnosis of multiple cancer types by shrunken centroids of gene expression " PNAS 2002 99:6567-6572 (the contents of which are incorporated herein by reference) and Tibshirani et al. (2002) “ Class prediction by nearest shrunken centroids, with applications to DNA

microarrays” Stanford tech report (the contents of which are incorporated herein by reference). A PAM software package (PAMR) is publically available at

http://www.bioconductor.Org/packages//2.7/bioc/html/pamr.html.

Briefly, by applying a molecular signature described herein to differential gene expression data or differential methylation data, discriminant scores can be calculated from which the probability of progression {i.e. the progression score) can be calculated, by analogy to Gaussian linear discriminant analysis, as implemented in the PAMR package (see equations [6]-[8] of Tibshirani, et al; PNAS 2002 99:6567-6572).

The progression score may be compared to a“ threshold value”, which is a numerical value between 0 and 1 , in order to make a binary classification for a given sample. For example, where a threshold value of X is applied, samples having a progression score less than X should be classified as regressive i.e. the individual from which the sample was obtained will be identified as not having a pre-invasive lung lesion that will progress to an invasive cancer. Conversely, samples having a progression score greater than X should be classified as progressive i.e. the individual from which the sample was obtained should be identified as having a pre-invasive lung lesion that will progress to an invasive lung cancer.

As shown in Figure 5, an example method using a DEG signature described herein (the DEG signature described in Table 2) generated progression scores of less than about 0.1 for the majority of regressive CIS lesions (plotted left of the vertical line in Figure 5B) and the majority of non-cancerous control samples (plotted left of the vertical line in Figure 5C). The majority of progressive CIS samples (plotted right of the vertical line in Figure 5B) and cancerous samples (plotted right of the vertical line in Figure 5B) had progression scores greater than about 0.1. Accordingly, based on this analysis, an appropriate“ threshold value” would be about 0.1. Thus, in this situation, where a sample is determined to have a progression score of less than 0.1 the individual is identified as not having a pre-invasive lesion that will progress to an invasive lung cancer. Alternatively, where a sample is determined to have a progression score of greater than 0.1 , the individual from which the sample was obtained will be identified as having a progressive pre-invasive lesion that will progress to an invasive lung cancer.

For any particular combination of molecular signature and threshold value, as applied to a differential gene expression dataset, a differential methylation dataset, or a copy number variation dataset, the skilled person using the PAMR methodology may determine an ROC AUC, sensitivity and/or specificity metrics. These metrics may be used to evaluate the usefulness of a particular method for a particular application. For example, a method that achieves a high sensitivity (few false negatives) at the expense of specificity (greater number of false positives) may be desirable if the method is to be used as a screening assay.

The skilled person will be able to select an appropriate an appropriate threshold value for use in a particular method. For example, where the method of the present invention involves determining a progression score using a DEG signature described herein, e.g. a DEG signature comprising the expression level of each of at least 10, at least 20, at least 30, at least 40, at least 50, at least 60, at least 75, at least 100, at least 150, at least 200, at least 250, at least 300, at least 325, least 350, at least 375, or 397 genes identified in Table 1, the threshold value may be from about 0.02 to about 0.06. Thus, in methods of the present invention where a progression score is determined using a DEG signature described herein e.g. a DEG signature comprising the expression level of each of at least 10, at least 20, at least 30, at least 40, at least 50, at least 60, at least 75, at least 100, at least 150, at least 200, at least 250, at least 300, at least 325, least 350, at least 375, or 397 genes identified in Table 1, the threshold value may be about 0.02, about 0.04, about 0.06, about 0.08, about 0.1, about 0.12, about 0.14, about 0.16, about 0.18, about 0.2, about 0.22, about 0.24, about 0.26, about 0.28, about 0.3, about 0.32, about 0.34, about 0.36, about 0.38, about 0.4, about 0.42, about 0.44, about 0.46, about 0.48, about 0.5, about 0.52, about 0.54, about 0.56, about 0.58, or about 0.6.

In other methods of the present invention, a progression score is determined using a DMP signature. Where the method of the present invention involves

determining a progression score using a DMP signature described herein, e.g. a DMP signature comprising the methylation status (b) of least two, at least three, at least four, at least five, at least six, at least seven, at least eight, at least nine, at least 10, at least 12, least 14, at least 16, at least 18, at least 20, at least 25, at least 50, at least 75, at least 100, at least 125, or differentially methylated positions (DMPs) selected from Table 11,

the threshold value may be from about 0.3 to 0.6. Thus, in methods of the present invention where a progression score is determined using a DMP signature described herein, e.g. a DMP signature comprising the methylation status (b) of least two, at least three, at least four, at least five, at least six, at least seven, at least eight, at least nine, at least 10, at least 12, least 14, at least 16, at least 18, at least 20, at least 25, at least 50, at least 75, at least 100, at least 125, or differentially methylated positions (DMPs) selected from Table 11, the threshold value may be about 0.3, about 0.31, about 0.32, about 0.33, about 0.34, about 0.35, about 0.36, about 0.37, about 0.38, about 0.39, about 0.4, about 0.41 , about 0.42, about 0.43, about 0.44, about 0.45, about 0.46, about 0.47, about 0.48, about 0.49, about 0.5, about 0.51 , about 0.52, about 0.53, about 0.54, about 0.55, about 0.56, about 0.57, about 0.58, about 0.59, or about 0.6.

In other methods of the present invention, a progression score is determined using a CNV signature. Where the method of the present invention involves determining a progression score using a CNV signature described herein, e.g. a CNV signature comprising the amplification or loss of at least 5, at least 6, least 7, at least 8, at least 9, at least 10, at least 25, at least 50, at least 75, at least 100, at least 125, at least 150, at least 175, or at least 200 CNV bands identified in Table 19, the threshold value may be from about 0.05 to 0.6. Thus, in methods of the present invention where a progression score is determined using a CNV signature described herein, e.g. a CNV signature comprising the amplification or loss of at least 5, at least 6, least 7, at least 8, at least 9, at least 10, at least 25, at least 50, at least 75, at least 100, at least 125, at least 150, at least 175, or at least 200 CNV bands identified in Table 19, the threshold value may be about 0.05, about 0.06, about 0.07, about 0.08, about 0.09, about 0.1, about 0.1 1, about 0.12, about 0.13, about 0.14, about 0.15, about 0.16, about 0.17, about 0.18, about 0.19, about 0.2, about 0.21, about 0.22, about 0.23, about 0.24, about 0.25, about 0.26, about 0.27, about 0.28, about 0.29, about 0.3, about 0.31, about 0.32, about 0.33, about 0.34, about 0.35, about 0.36, about 0.37, about 0.38, about 0.39, about 0.4, about 0.41, about 0.42, about 0.43, about 0.44, about 0.45, about 0.46, about 0.47, about 0.48, about 0.49, about 0.5, about 0.51 , about 0.52, about 0.53, about 0.54, about 0.55, about 0.56, about 0.57, about 0.58, about 0.59, or about 0.6.

Assessment of differentially expressed senes (DEGs)

In some of the methods of the present invention described herein, a progression score is determined using a DEG signature which comprises the expression levels of a plurality of differentially expressed genes (DEGs).

In any of the methods disclosed herein, the term“expression level” may refer to any measurable indicator of abundance of any number of the DEGs defined herein in the sample of nucleic acid which has been taken from a tissue of an individual. Thus,

“ expression level” of a DEG may refer an amount or quantity of a mRNA transcribed from said DEG defined herein or an amount or quantity of a cDNA reverse transcribed from said mRNA. Accordingly, the measurable indicator of abundance may, for example, be the concentration of a given mRNA or cDNA as determined using any technique suitable for use in a method of the invention. The measurable indicator of abundance may also be level of fluorescence, densitometry, colorimetry, or any assay indicator suitable for use in a method of the invention for providing a measurement of DEG, mRNA or cDNA expression level derived from the sample of nucleic acid.

A variety of techniques are available for determining the expression level of a DEG, as will be outlined briefly below. The methods described herein encompass any suitable technique for the determining the expression level of a differentially expressed gene.

Differential gene expression levels may be determined by a hybridisation assay using probes specific for a DEG of interest. The expression levels of multiple DEGs may be assayed in parallel using a plurality of probes, e.g. a plurality of probes provided together in a microarray. Typically, such microarrays will comprises a plurality of probes specific for labelled cDNAs prepared from mRNA isolated from a tissue. When hybridised to the microarray, the labelled cDNA generates a plurality of signals each of which is indicative of the concentration of a particular mRNA transcribed from a particular DEG.

The mRNA used to generate the cDNA by reverse transcription may be extracted from a formalin-fixed paraffin-embedded (FFPE) tissue sample. The expression level of DEGs may be assessed using a commercially available Human Whole-Genome DASL (cDNA-mediated Annealing, Selection, extension and Ligation)

beadarray (Illumina). The expression level of DEGs may be assayed using a

commercially available Clariom™ D Transcriptome Human Pico Assay 2.0

(Affymetrix).

Quantitative reverse transcription PCR (RT -qPCR) may also be used to determine the expression level of DEGs.

RNA sequencing (RNA-seq) methods may also be used to determine the expression level of DEGs. In such methods, total RNA may be isolated from the tissue sample taken from the individual. The total RNA sample may be purified to enrich for mRNAs prior to preparing an RNA library for sequencing. Library preparation may involve such steps as reverse transcription to cDNA, PCR amplification and may or may not preserve strandedness information. Next generation sequencing (NGS) may be used to sequence the cDNA library generated from the enriched mRNA in order to provide transcriptome information which can be compared against a reference in order to determine the expression levels of DEGs.

The skilled person may select any suitable available RNA-seq method available in the art. For example, any RNA-seq method described in Comey and Basturea (Materials and Methods 2013; 3:203; doi: 10.13070/mm. en.3.203; available at

https://www.labome.com/method/RNA-seq-Using-Next-Generation-Sequencing.html. which is incorporated herein by reference). Various RNA-seq methods are

commercially available. For example, an overview of RNA-seq methods available from Illumina is available at: https://www.illumina.com/content/dam/illumina-marketing/documents/products/other/ma-sequencing-workflow-buvers-guide-476-2015-003.pdf The skilled person is able to select an appropriate RNA-seq method depending on the type of sample, for example a particular RNA-seq method may be appropriate for interrogation of fresh frozen (FF) samples and another RNA-seq method may be better suited for interrogation of formalin-fixed paraffin- embedded samples (FFPE).

Nevertheless, it is routine for the skilled person to select a suitable RNA-seq method for evaluation of the expression levels of DEGs in any of the methods of the present invention disclosed herein.

Assessment of differentially methylated position (DMP) methylation status (b)

In some of the methods of the present invention, described herein a progression score is determined using a DMP signature which comprises the methylation status (b) of a plurality of differentially methylated positions (DMPs), which may also be referred to as methylation variable positions (MVPs).

Methylation status (b) is a well know term in the art and the skilled person is readily able to determine the b value of given DMP. A single DMP on a single DNA molecule will either be methylated (M) or unmethylated (U). Thus, the b value for a single DMP, in a single DNA, is binary, i.e. M = 1 or U = 0. However, for a population comprising a plurality of DNA molecules, there will be multiple copies of the same DMP. Thus, there may be variation in methylation status between copies of the same DMP on different DNA molecules. Thus, for a sample comprising a population of DNA molecules, the b value is measure of the average methylation status across the entire population and can therefore have any value between 1 and 0.

Where b values are determined on the basis of signals (e.g. fluorescent signals) associated with methylated or unmethylated DMPs. For each sample, the b value may be calculated as the ratio of the signal intensity of the methylated (M) and unmethylated (U) DMPs. For example, according to the following formula:

intensity of signal from M DMPs

^ instesity of signal from [U DMPs + M DMPs] + 100

Methylation of DNA is a recognised form of epigenetic modification which has the capability of altering the expression of genes and other elements such as microRNAs. In cancer development and progression, methylation may have the effect of e.g. silencing tumor suppressor genes and/or increasing the expression of oncogenes. Other forms of dysregulation may occur as a result of methylation. Methylation of DNA occurs at discrete loci which are predominately dinucleotide consisting of a CpG motif, but may also occur at CHH motifs (where H is A, C, or T). During methylation, a methyl group is added to the fifth carbon of cytosine bases to create methylcytosine.

Methylation can occur throughout the genome and is not limited to regions with respect to an expressed sequence such as a gene. Methylation typically, but not always, occurs in a promoter or other regulatory region of an expressed sequence.

A DMP as defined herein is any dinucleotide locus which may show a variation in its methylation status between phenotypes, e.g. between a progressive pre-invasive lung lesion and a regressive pre-invasive lung lesion. A DMP is preferably a CpG or a CHH dinucleotide motif. A DMP as defined herein is not limited to the position of the locus with respect to a corresponding expressed sequence.

Typically, an assessment of DNA methylation status involves analysing the presence or absence of methyl groups in DNA, for example methyl groups on the 5th position of one or more cytosine nucleotides. Preferably, the methylation status of one or more cytosine nucleotides present as a CpG dinucleotide (where C stands for Cytosine, G for Guanine and p for the phosphate group attached to the backbone between the two) is assessed.

A variety of techniques are available for determining the methylation status (i.e. determine the b value) of a DMP, as will be outlined briefly below. The methods described herein encompass any suitable technique for the determination of DMP methylation status.

Methyl groups are lost from a starting DNA molecule during conventional in vitro handling steps such as PCR. To avoid this, techniques for the detection of methyl groups commonly involve the preliminary treatment of DNA prior to subsequent processing, in a way that preserves the methylation status information of the original DNA molecule. Such preliminary techniques involve three main categories of processing, i.e. bisulphite modification, restriction enzyme digestion and affinity-based analysis. Products of these techniques can then be coupled with sequencing or array-based platforms for subsequent identification or qualitative assessment of DMP methylation status.

Techniques involving bisulphite modification of DNA have become the most common methods for detection and assessment of methylation status of CpG

dinucleotide. Treatment of DNA with bisulphite, e.g. sodium bisulphite, converts cytosine bases to uracil bases, but has no effect on 5-methylcytosines. Thus, the

presence of a cytosine in bisulphite-treated DNA is indicative of the presence of a cytosine base which was previously methylated in the starting DNA molecule. Such cytosine bases can be detected by a variety of techniques. For example, primers specific for unmethylated versus methylated DNA can be generated and used for PCR-based identification of methylated CpG dinucleotides. A separation/capture step may be performed, e.g. using binding molecules such as complementary oligonucleotide sequences. Standard and next-generation DNA sequencing protocols can also be used.

In other approaches, methylation-sensitive enzymes can be employed which digest or cut only in the presence of methylated DNA. Analysis of resulting fragments is commonly carried out using microarrays.

Affinity-based techniques exploit binding interactions to capture fragments of methylated DNA for the purposes of enrichment. Binding molecules such as anti-5-methylcytosine antibodies are commonly employed prior to subsequent processing steps such as PCR and sequencing.

Olkhov-Mitsel and Bapat (2012) [46] provide a comprehensive review of techniques available for the identification and assessment of DMP-based biomarkers involving methylcytosine.

For the purposes of assessing the methylation status of the DMP-based biomarkers characterised and described herein, any suitable method can be employed.

Preferred methods involve bisulphite treatment of DNA, including amplification of the identified DMP loci for methylation specific PCR and/or sequencing and/or assessment of the methylation status of target loci using methylation-discriminatory microarrays.

Amplification of DMP loci can be achieved by a variety of approaches.

Preferably, DMP loci are amplified using PCR. DMPs may also be amplified by other techniques such as multiplex ligation-dependent probe amplification (MLPA). A variety of PCR-based approaches may be used. For example, methylation-specific primers may be hybridized to DNA containing the DMP sequence of interest. Such primers may be designed to anneal to a sequence derived from either a methylated or non-methylated DMP locus. Following annealing, a PCR reaction is performed and the presence of a subsequent PCR product indicates the presence of an annealed DMP of identifiable sequence. In such methods, DNA is bisulphite converted prior to amplification. Such techniques are commonly referred to as methylation specific PCR (MSP) [47]

In other techniques, PCR primers may anneal to the DMP sequence of interest independently of the methylation status, and further processing steps may be used to determine the status of the DMP. Assays are designed so that the DMP site(s) are located between primer annealing sites. This method scheme is used in techniques such as bisulphite genomic sequencing [48], COBRA [49], Ms-SNuPE [50] In such methods, DNA can be bisulphite converted before or after amplification.

Preferably, small-scale PCR approaches are used. Such approaches commonly involve mass partitioning of samples ( e.g . digital PCR). These techniques offer robust accuracy and sensitivity in the context of a highly miniaturised system (pico -liter sized droplets), ideal for the subsequent handling of small quantities of DNA obtainable from the potentially small volume of cellular material present in biological samples, particularly urine samples. A variety of such small-scale PCR techniques are widely available. For example, microdroplet-based PCR instruments are available from a variety of suppliers, including RainDance Technologies, Inc. (Billerica, MA;

http://raindancetech.com/) and Bio-Rad, Inc. (http://www.bio-rad.com/). Microarray platforms may also be used to carry out small-scale PCR. Such platforms may include microfluidic network-based arrays e.g. available from Fluidigm Corp.

(www.fluidigm.com).

Following amplification of DMP loci, amplified PCR products may be coupled to subsequent analytical platforms in order to determine the methylation status of the DMPs of interest. For example, the PCR products may be directly sequenced to determine the presence or absence of a methylcytosine at the target DMP or analysed by array-based techniques.

Any suitable sequencing techniques may be employed to determine the sequence of target DNA. In the methods of the present invention the use of high-throughput, so-called“second generation”,“third generation” and“next generation” techniques to sequence bisulphite-treated DNA are preferred.

In second generation techniques, large numbers of DNA molecules are sequenced in parallel. Typically, tens of thousands of molecules are anchored to a given location at high density and sequences are determined in a process dependent upon DNA synthesis. Reactions generally consist of successive reagent delivery and washing steps, e.g. to allow the incorporation of reversible labelled terminator bases, and scanning steps to determine the order of base incorporation. Array-based systems of this type are available commercially e.g. from Illumina, Inc. (San Diego, CA;

http://www.illumina.com/).

Third generation techniques are typically defined by the absence of a requirement to halt the sequencing process between detection steps and can therefore be viewed as real-time systems. For example, the base-specific release of hydrogen ions, which occurs during the incorporation process, can be detected in the context of microwell systems (e.g. see the Ion Torrent system available from Life Technologies; http://www.lifetechnologies.com/). Similarly, in pyrosequencing the base-specific release of pyrophosphate (PPi) is detected and analysed. In nanopore technologies, DNA molecules are passed through or positioned next to nanopores, and the identities of individual bases are determined following movement of the DNA molecule relative to the nanopore. Systems of this type are available commercially e.g. from Oxford Nanopore (https://www.nanoporetech.com/). In an alternative method, a DNA polymerase enzyme is confined in a“zero-mode waveguide” and the identity of incorporated bases are determined with florescence detection of gamma-labeled phosphonucleotides (see e.g. Pacific Biosciences; http://www.pacificbiosciences.com/).

In other methods in accordance with the invention sequencing steps may be omitted. For example, amplified PCR products may be applied directly to hybridization arrays based on the principle of the annealing of two complementary nucleic acid strands to form a double-stranded molecule. Hybridization arrays may be designed to include probes which are able to hybridize to amplification products of a DMP and allow discrimination between methylated and non-methylated loci. For example, probes may be designed which are able to selectively hybridize to an DMP locus containing thymine, indicating the generation of uracil following bisulphite conversion of an unmethylated cytosine in the starting template DNA. Conversely, probes may be designed which are able to selectively hybridize to a DMP locus containing cytosine, indicating the absence of uracil conversion following bisulphite treatment. This corresponds with a methylated DMP locus in the starting template DNA.

Following the application of a suitable detection system to the array, computer-based analytical techniques can be used to determine the methylation status of an DMP. Detection systems may include, e.g. the addition of fluorescent molecules following a methylation status-specific probe extension reaction. Such techniques allow DMP status determination without the specific need for the sequencing of DMP amplification products. Such array-based discriminatory probes may be termed methylation-specific probes.

Any suitable methylation-discriminatory microarrays may be employed to assess the methylation status of the DMPs described herein. A preferred methylation-discriminatory microarray system is provided by Illumina, Inc. (San Diego, CA;

http://www.illumina.com/).

In particular, the Infinium HumanMethylation450 BeadChip array system may be used to assess the methylation status of DMPs as described herein. Such a system exploits the chemical modifications made to DNA following bisulphite treatment of the starting DNA molecule. Briefly, the array comprises Type I beads to which are coupled oligonucleotide probes specific for DNA sequences corresponding to the unmethylated form of a DMP, as well as separate Type I beads to which are coupled oligonucleotide probes specific for DNA sequences corresponding to the methylated form of a DMP.

The Infinium HumanMethylation450 BeadChip array system also comprises Type II beads to which are coupled two different types of oligonucleotide probe: a first probe specific for DNA sequences corresponding to the unmethylated form of a DMP and a second probe specific for DNA sequences corresponding to the methylated form of the same DMP.

Candidate DNA molecules are applied to the array and selectively hybridize, under appropriate conditions, to the oligonucleotide probe corresponding to the relevant epigenetic form. Thus, a DNA molecule derived from a DMP which was methylated in the corresponding genomic DNA will selectively attach to methylation-specific oligonucleotide probes, but will fail to attach to the non-methylation-specific

oligonucleotide probe. Single-base extension of only the hybridized probes

incorporates a labeled ddNTP, which is subsequently stained with a fluorescence reagent and imaged. The methylation status of the DMP may be determined by calculating the ratio of the fluorescent signal derived from the methylated and unmethylated sites.

Because the DMPs of the DMP signatures defined herein were initially identified using the Illumina Infinium HumanMethylation450 BeadChip array system, the same chip system can be used to interrogate those same DMPs in the methods described herein. Alternative or customised arrays could, however, be employed to interrogate the diagnostic DMPs defined herein, provided that they comprise means for interrogating all DMPs for a given method, as defined herein.

Techniques involving combinations of the above-described methods may also be used. For example, DNA containing DMP sequences of interest may be hybridized to microarrays and then subjected to DNA sequencing to determine the status of the DMP as described above.

In the methods described above, sequences corresponding to DMP loci may also be subjected to an enrichment process. DNA containing DMP sequences of interest may be captured by binding molecules such as oligonucleotide probes complementary to the DMP target sequence of interest. Sequences corresponding to DMP loci may be captured before or after bisulphite conversion or before or after amplification. Probes may be designed to be complementary to bisulphite converted DNA. Captured DNA may then be subjected to further processing steps to determine the status of the DMP, such as DNA sequencing steps.

Capture/separation steps may be custom designed. Alternatively a variety of such techniques are available commercially, e.g. the SureSelect target enrichment system available from Agilent Technologies f http://www.agilent.com/home). In this system biotinylated“bait” or“probe” sequences (e.g. RNA) complementary to the DNA containing DMP sequences of interest are hybridized to sample nucleic acids.

Streptavidin-coated magnetic beads are then used to capture sequences of interest hybridized to bait sequences. Unbound fractions are discarded. Bait sequences are then removed (e.g. by digestion of RNA) thus providing an enriched pool of DMP target

sequences separated from non-DMP sequences. In a preferred method of the invention, template DNA is subjected to bisulphite conversion and target loci are then amplified by small-scale PCR such as microdroplet PCR using primers which are independent of the methylation status of the DMP. Following amplification, samples are subjected to a capture step to enrich for PCR products containing the target DMP, e.g. captured and purified using magnetic beads, as described above. Following capture, a standard PCR reaction is carried out to incorporate DNA sequencing barcodes into DMP-containing amplicons. PCR products are again purified and then subjected to DNA sequencing and analysis to determine the presence or absence of a methylcytosine at the target genomic DMP [32]

The DMP biomarker loci defined herein are identified e.g. by Illumina® identifiers (IlmnID), which are also referred to as DMP identifiers (DMP ID). These DMP loci identifiers refer to individual DMP sites used in the commercially available Illumina® Infinium Human Methylation450 BeadChip kit. The identity of each DMP site represented by each DMP loci identifier is publicly available from the Illumina, Inc. website under reference to the DMP sites used in the Infinium Human Methylation450 BeadChip kit.

Further information regarding DMP loci identification used in Illumina, Inc products is found in the technical note entitled“Technical Note: Epigenetics. CpG Loci Identification. A guide to Illumina’ s method for unambiguous CpG loci identification and tracking for the Golden Gate® and Infinium® Assay for Methylation” published in 2010 and found at:

http://www.illumina.com/documents/products/technotes/technote cpg loci iden tification.pdf.

Further information regarding the Illumina® Infinium Human Methylation450 BeadChip system can be found at:

http : / / www. illumina. com/ content/ dam/ illumina-marketing/ documents/products/ datasheets/ datasheet humanmethylation45 O.pdf :

and at:

http://www.illumina.com/content/dam/illumina-marketing/documents/products/technotes/technote hm450 data analysis optimization. pdf.

To complement evolving public databases to provide accurate DMP/CpG loci identifiers and strand orientation, Illumina® has developed a method to consistently designate DMP/CpG loci based on the actual or contextual sequence of each individual DMP/CpG locus. To unambiguously refer to DMP/CpG loci in any species, Illumina® has developed a consistent and deterministic DMP loci database to ensure uniformity in the reporting of methylation data. The Illumina® method takes advantage of sequences flanking a DMP locus to generate a unique DMP locus cluster ID. This number is based on sequence information only and is unaffected by genome version. Illumina’ s standardized nomenclature also parallels the TOP/BOT strand nomenclature (which indicates the strand orientation) commonly used for single nucleotide polymorphism (SNP) designation.

Illumina® Identifiers for the Infinium Human Methylation450 BeadChip system are also available from public repositories such as Gene Expression Omnibus (GEO) (http://www.ncbi.nlm.nih.gov/geo/). For example, at

https://www.ncbi.nlm.nih.gov/geo/querv/acc.cgi?acc=GPL13534

Provided herein are DMP signatures comprising two or more of the DMPs listed in Tables 12-17. Each of the DMPs listed in Tables 11 -17 designated with a unique identifier termed a“ DMP ID’ (also known as an“ Illumina ID”) from which the skilled person can derive a genomic locus and nucleotide sequence said DMP, using for example the GEO database accessible at

https://www.ncbi.nlm.nih. gov/geo/querv/acc.cgi?acc=GPLl 3534. By way of example, the first DMP listed in Table 11 is designated“cg07716946” which corresponds to a DMP within chr 15:67325986-67326256 and specifically at position 67326118 of chromosome 15. The second DMP listed in Table 11 is designated“ cg0478628T’ which corresponds to a DMP within chrl 6:54970301 -54972846 and specifically at position 54970523 of chromosome 16. The nucleotide sequences corresponding to DMP cg07716946 (with the CpG island designated using the notation [CG]) is:

CCGGG ACGCT GCT GGAGGCGCCGTCGCTCCGCGGCGGAGGCGACCC AGTTT CCCAGCTCT[CG]TCCTCGCCACTTCCTCTGCATGGGCTTCCAGGAGACTCGG CCTCCGTCGGCGACGCTGGC

(SEQ ID NO: 1)

The nucleotide sequences corresponding to DMP cg04786287 (with the CpG island designated using the notation [CG]) is:

GCTGTCCGCTGCCCGCATCCCTTCCGCCCTGGGCCTCTGCACGGTCTGCGGT TTTCTGTG[CG] C ACTTGGT CTT C AGT ACT AGC ACCC AATT ACGTCTGGGTTTT TCTTCTTT AC AG AGCT GG

(SEQ ID NO: 2)

Assessment of copy number variation ( CNV )

In some of the methods of the present invention described herein, a progression score is determined using a CNV signature which comprises the amplification or loss of a plurality of CNV bands, which are also referred to in the art as“ cytogenetic bands”.

CNV bands identified herein, such as the CNV bands identified in Tables 19-36, are designated using the standard nomenclature for cytogenetic bands. In this regard, each human chromosome has a short arm designated "p" and long arm designated "q", separated by a centromere. The ends of the chromosome are called telomeres. The telomere at the end of the p arm is referred to as“ pteF while the telomere at the end of the q arm designated“qtel”. Each chromosome arm is divided into cytogenetic bands (i.e. CNV bands as referred to herein), which can be differentially stained and visualised using a microscope. The cytogenetic bands are labelled pi, p2, p3, etc. or ql, q2, q3, etc., counting from the centromere out toward the telomeres of either the p arm or the q arm. Each CNV band may comprise sub-bands, which in turn may comprise sub-sub-bands. The sub-bands and sub-sub-bands are also numbered from the centromere out toward the telomere. Accordingly, by way of example, if a CNV band has the designation 7q31.2 this indicates that the CNV band is on chromosome 7, q arm, band 3, sub-band 1 , and sub-sub-band 2.

A variety of techniques are available for determining the status of a CNV band, as will be outlined briefly below. The methods described herein encompass any

suitable technique for the determining the status of a CNV band, e.g. determining whether a particular CNV band has been amplified or lost.

Thus, in any of the methods of the present invention described herein, the amplification or loss of CNV bands may be determined using a method selected from the group consisting of next generation sequencing (NGS), the nCounter system

(nanoString; https://www.nanostring.com/download file/view/323/3778). Multiplex Ligation-Dependent Probe Amplification (MLPA), real-time quantitative PCR, comparative genomic hybridization (CGH), Fluorescent In Situ Hybridization (FISH), and combinations thereof.

Pre-invasive luns lesions and lune cancer

The methods of the present invention described herein may be used to determine whether or not an individual has a pre-invasive lung lesion that will progress to an invasive lung cancer.

As used herein the term“lung lesiori’ may refer to any population of cells derived from damaged or diseased lung tissue. The lesion may be cancerous. The lesion may be non-cancerous. Thus, in the context of the present invention the lesion lung lesion may present in the normal epithelium, tissue hyperplasia, dysplasia, or lung carcinoma in situ (CIS).

As used herein the term“ pre-invasive” refers to a lesion or cell population which has not spread, or is not capable of spreading, beyond the tissue region in which said lesion or cell population originated. Conversely, as used herein, the term “ invasive” refers to a cancer that has spread, or is capable of spreading, beyond the tissue region in which said cancer originated. The term“ infiltrating” is also used in the art to refer to invasive cancer. The invasive lung cancer may be a lung squamous cell carcinoma (LUSC).

Methods of treatment

The present invention also provides a method of treating a pre-invasive lung lesion and/or treating and/or preventing an invasive lung cancer in an individual comprising:

identifying a pre-invasive lung lesion in an individual that will progress to an invasive lung cancer by performing a method of the present invention; and

administering a lung cancer therapy to the individual.

Thus, for example, the present invention provides a method of treating a pre-invasive lung lesion and/or treating and/or preventing an invasive lung cancer in an individual comprising:

identifying a pre-invasive lung lesion in an individual that will progress to an invasive lung cancer by performing an identification method; and

administering a lung cancer therapy to the individual,

wherein the identification method is a method of identifying whether or not an individual has a pre-invasive lung lesion that will progress to an invasive lung cancer, the method comprising:

(a) providing a sample of nucleic acid which has been taken from a tissue of the individual, wherein the tissue is suspected of harbouring a pre-invasive lung lesion;

(b) performing an assay to determine a progression score for the sample; and

(c) identifying whether or not the individual has a pre-invasive lung lesion that will progress to an invasive lung cancer by comparing the progression score to a threshold value;

wherein the progression score is determined using a molecular signature selected from:

i) a differentially expressed gene (DEG) signature;

ii) a differentially methylated position (DMP) signature;

iii) a copy number variation (CNV) signature; and

iv) combinations of (i) to (iii).

The present invention also provides a method for determining whether or not to provide a therapeutic method of treatment to an individual, the method comprising performing a method of identifying whether or not an individual has a pre- invasive lung lesion that will progress to an invasive lung cancer according to the present invention;

providing to the individual a therapeutic method of treatment if the individual is identified as having a pre-invasive lung lesion that will progress to an invasive lung cancer, optionally wherein the therapeutic method of treatment comprises administering to the individual a therapeutically effective amount of a lung cancer therapy.

The lung cancer therapy may comprise any known treatment or procedure known in the art.

Thus, the invention encompasses administration of one or more surgical procedures, one or more chemotherapeutic agents, one or more immunotherapeutic agents, one or more radiotherapeutic agents, one or more hormonal therapeutic agents or any combination of the above following the identification of a pre-invasive lung lesion in an individual that will progress to an invasive lung cancer.

Surgical procedures the removal or one or more lung lobe (lobectomy), removal of two lung lobes (bilobectomy) removal of a lung, a lung transplant, a

lymphadenectomy, a wedge resection, a segmentectomy, and a sleeve resection.

Chemotherapeutic agents include the following. Alkylating agents, which include the nitrogen mustards, nitrosoureas, tetrazines, aziridines, cisplatin and platinum based derivatives, as well as the non-classical alkylating agents. Antimetabolites, which include the anti-folates, fluoropyrimidines, deoxynucleoside analogues and thiopurines. Microtubule disrupting agents, which include the vinca alkaloids and taxanes, as well as dolastatin 10 and derivatives thereof. Topoisomerase inhibitors, which include camptothecin, irinotecan and topotecan. Topoisomerase II poisons, which include etoposide, doxorubicin, mitoxantrone and teniposide. Topoisomerase II catalytic inhibitors, which include novobiocin, merbarone, and aclarubicin. Cytotoxic antibiotics, which include anthracyclines, actinomycin, bleomycin, plicamycin, and mitomycin.

Immunotherapeutics include monoclonal antibodies, antibody-drug conjugates, immune checkpoint inhibitors. For example, the lung cancer therapy may be selected from monoclonal antibodies directed against the VEGF/VEGFR pathway such as bevacizumab (Avastin), mononclonal antibodies directed against the EGFR pathway, such as Necitumumab (Portrazza), monoclonal antibodies that inhbit the PD-1. PD-L1 pathway, such as Atezolizumab (Tecentriq), Durvalumab (Imfinzi) Nivolumab

(Opdivo) and Pembrolizumab (Keytruda).

Combination therapies include carboplatin-taxol and gemcitabine-cisplastin.

The lung cancer therapy may be selected from Abraxane (Paclitaxel Albumin-stabilized Nanoparticle Formulation), Afatinib Dimaleate, Afinitor (Everolimus), Alecensa (Alectinib), Alectinib, Alimta (Pemetrexed Disodium), Alunbrig (Brigatinib), Atezolizumab, Avastin (Bevacizumab), Bevacizumab, Brigatinib, Carboplatin,

Ceritinib, Crizotinib, Cyramza (Ramucirumab), Dabrafenib, Dacomitinib, Docetaxel, Durvalumab, Erlotinib Hydrochloride, Everolimus, Gefitinib, Gilotrif (Afatinib

Dimaleate), Gemcitabine Hydrochloride, Gemzar (Gemcitabine Hydrochloride), Imfinzi (Durvalumab), Iressa (Gefitinib), Keytruda (Pembrolizumab), Mechlorethamine Hydrochloride, Mekinist (Trametinib), Methotrexate, Mustargen (Mechlorethamine Hydrochloride), Navelbine (Vinorelbine Tartrate), Necitumumab, Nivolumab, Opdivo (Nivolumab), Osimertinib, Paclitaxel, Paclitaxel Albumin-stabilized Nanoparticle Formulation, Paraplat (Carboplatin), Paraplatin (Carboplatin), Pembrolizumab,

Pemetrexed Disodium, Portrazza (Necitumumab), Ramucirumab, Tafinlar (Dabrafenib), Tagrisso (Osimertinib), Tarceva (Erlotinib Hydrochloride), Taxol (Paclitaxel), Taxotere (Docetaxel), Tecentriq (Atezolizumab), Trametinib, Trexall (Methotrexate), Vizimpro (Dacomitinib), Vinorelbine Tartrate, Xalkori (Crizotinib), Zykadia (Ceritinib), and combinations thereof.

The lung cancer therapy may comprise proton beam therapy. The lung cancer therapy may comprise photodynamic therapy.

The lung cancer therapy may be administered to an individual already having an invasive lung cancer, in an amount sufficient to cure, alleviate or partially arrest the lung cancer or one or more of its symptoms. Such therapeutic treatment may result in a decrease in severity of disease symptoms, or an increase in frequency or duration of symptom- free periods. An amount adequate to accomplish this is defined as

" therapeutically effective amount" . Effective amounts for a given purpose will depend on the severity of the disease as well as the weight and general state of the individual.

The lung cancer therapy may also be administered to an individual identified as having a pre-invasive lung lesion that will progress to an invasive lung cancer but who has not yet developed an invasive lung cancer. Such preventative treatment may prevent the progression of the pre-invasive lung lesion to an invasive cancer, delay the progression of the pre-invasive lung lesion to an invasive cancer, and/or lessen the severity or extent an invasive cancer derived from the identified pre-invasive lung lesion.

The following Examples are provided to illustrate the invention but not to limit the invention.

Examples

Example 1 : General Study Design

It was reasoned that information on the future clinical trajectory of a pre-invasive lung lesion might be encoded in the genetic and epigenetic profile present at diagnosis. A prospective cohort study of patients with pre-invasive squamous airway lesions was therefore undertaken. Patients were managed conservatively, undergoing surveillance AFB with biopsy and CT scanning every 4 and 12 months, respectively, with definitive cancer treatment only performed at the earliest pathological evidence of progression to invasive tumours (Fig. 2A, B) [7] When a CIS lesion either progressed to invasive cancer or regressed to normal epithelium/low-grade disease, molecular profiling was performed on the preceding CIS biopsy from the same lesion - the‘index biopsy’ (Fig. 2C). Index biopsies all demonstrated histologically and morphologically indistinguishable CIS and were classified as either‘progressive’ or‘regressive’. All such index CIS biopsies were subjected to a predetermined combination of

transcriptomic, epigenetic and finally genomic profiling depending on DNA/RNA availability (Figure 1 ; Fig. 2D; Fig. 7).

Example 2: Patient Characteristics

Patients with pre-invasive lung cancer lesions were recruited through University College London Hospitals (UCLH) Early Lung Cancer Surveillance Programme (ELCSP). Full details of the surveillance protocol including eligibility criteria for patient inclusion have been previously described [7] Briefly, the programme has recruited 140 patients to date with pre-invasive lung cancer lesions of varying histological grades. 129 index CIS biopsies were obtained from 85 patients and subjected to molecular analysis. Dependent on stored tissue quantity, in total, 51 samples from 42 patients underwent gene expression profiling; 87 samples from 47 patients underwent methylation profiling; and 39 samples from 29 patients underwent whole genome sequencing. Methylation and gene expression datasets were divided into independent discovery and validation groups.

Clinical characteristics within each analysis group are shown in Figure 1. In comparing progressive and regressive samples, it was found that progressive samples were associated with a higher pack-year smoking history in the methylation discovery group only (p < 0.01) and with increased age in the WGS group (p = 0.01). No clinical differences were consistently observed across the different analysis groups.

Example 3: Characterisation of CIS genomic profiles

The 39 CIS lesions are the first pre-invasive LUSC lesions to be whole-genome sequenced, so the burden and spectrum of mutations in CIS was compared with publicly available LUSC exome sequencing data from The Cancer Genome Atlas (TCGA). Due to differences between whole-genome and exome sequencing, only broad comparisons were made. A similar mutation burden and copy number profile between CIS samples and TCGA LUSC tumours was observed (Fig. 3). There is congruency of type and prevalence of potential driver mutations, broadly defined as any mutation in a gene previously implicated as a driver of lung cancer, between CIS and LUSC samples [8] Frequent alterations in TP53, CDKN2A, SOX2 and AKT2, and less frequent alterations in FAT1, KMT2D, KEAP1, EGFR and NOTCH 1 in CIS lesions were observed (Fig. 3). CIS mutational signatures [9], [10] showed a strong tobacco-associated signal and were similar to those found in LUSC (Fig. 8).

Marked aneuploidy was observed in CIS lesions, with somatic copy number alterations (CNAs) present across the genome (Fig. 3; Fig. 7). The most frequent changes were associated with gain and amplification of multiple locations on distal 3q: this is known to be the most common genomic aberration in LUSC1 1. Other recognised copy number associations identified in our data include gain/amplification in 5p, 8q and 19q and regions of loss/deletion in 3p, 4q, 5q, 8p, 9p and 13q.12-18.

Whilst most CIS samples had the genomic appearance of neoplasms, six lesions were observed which showed markedly lower mutational load and fewer copy number alterations than the others (Fig. 7; PD21884c, PD21885a, PD21885c, PD21904d,

PD38317a, PD38319a). These samples had very few genomic changes, despite being CIS histologically. All of these six samples regressed to normal epithelium or low-grade dysplasia on subsequent biopsy. Four further samples met this end-point for regression, despite widespread mutational and copy number changes. However, with longer follow up one of these cases developed CIS recurrence (Fig. 10a; PD21893a), and two developed invasive cancer on further surveillance (Fig. 10B, C; PD21884a, PD38326a). Only one sample, PD21908a, showed sustained clinical regression after 9 years of follow up despite widespread molecular changes.

All but one progressive sample and all highly mutated regressive samples showed amplification in a small region of distal 3q (chr3: 172516434- 178440382). This region contains known driver genes (SOX2, PIK3CA), genes associated with chromosomal instability (ACTL6A) and methyltransferases (ECE2). Progressive sample PD38320a had little change outside this region and did not harbour a TP53 mutation, suggesting that this amplification may be a crucial early event in LUSC tumorigenesis.

Genomic features between the 29 progressive and 10 regressive lesions were compared. The three samples which showed evidence of progression after meeting the end-point for regression were excluded from this analysis. Comparisons of mutation burden between progressive and regressive lesions were performed by mixed effects modelling, allowing us to account for samples that come from the same patient. Even after correcting for patient age, smoking history and sample purity, progressive lesions had more somatically acquired mutations than those from regressive lesions, across base substitutions (p<0.001), indels (p=0.018), structural variants (p<0.001) and copy number changes (p<0.001) (Fig. 1 1 A-D). When the analysis was restricted only to substitutions that were fully clonal in each lesion, there were still substantially more substitutions in progressive than regressive lesions (p<0.001) (Fig. 1 IE), suggesting that the increase in mutation burden is not due to recent subclonal diversification in progressive lesions. All the mutational processes (or signatures [9], [10]) identified in the CIS lesions contribute to the excess of mutations in progressive compared to regressive samples; however, only tobacco-associated signature 4 showed

proportionally more mutations (p=0.017) (Fig. 8F-J). Progressive lesions contained

more putative driver mutations than regressive lesions (p=0.001) (Fig. 1 1H).

Importantly, no single cancer mutation perfectly discriminated between progressive and regressive lesions.

Within the biopsied lesions, clonal architecture was similar between progressive and regressive lesions (Fig. 11E-G). For four patients in whom we sequenced multiple progressive lesions, the lesions shared many somatic mutations despite their different locality in the bronchial tree, indicating their probable derivation from a common ancestral clone. By contrast, multiple regressive lesions from two further patients did not share common mutations and so are likely to have arisen independently (Fig. 12). There were no differences in telomere lengths between progressive and regressive lesions (p=0.59) (Fig. 1 II).

Example 4: CIS transcriptomie and epigenetic profiles

Gene expression microarrays were performed on a discovery set of 17 progressive and 16 regressive CIS lesions. Identified were 1335 genes with significant expression changes (FDR < 0.01); 657 genes were up-regulated and 678 down-regulated in progressive CIS lesions (Fig. 4A).

Differential analysis of methylation profiles was performed on a discovery set of 26 progressive, 11 regressive and 23 control samples. Widespread methylation changes were observed with 12,064 differentially methylated positions (DMPs), associated with 2,695 genes, at which methylation was significantly different between progressive and regressive samples (FDR < 0.01; |Db| > 0.3). 6,314 DMPs were hypermethylated and 5,750 hypomethylated in progressive CIS (Fig. 4B). 260 differentially methylated regions (DMRs) were identified, of which 151 (58%) overlap with DMRs between

TCGA cancer and control data (Fig. 13). Finally, identified were 36,620 differentially variable positions (DVPs) for which probe variance was markedly different between progressive and regressive groups.

Of the 1335 genes identified, TPM3, PTPRB, SLC34A2, KEAP1, NKX2-1, SMAD4 and SMARCA4 have previously been implicated as potential lung cancer drivers. Regarding methylation, the potential driver genes NKX2-1, TERT, DDR2,

LRIG3, CUX1, EPHA3, CSMD3, MET, ZNF479, GRIN2A, PTPRD, NOTCH1, CD74, NSD1 and CDKN2A contain at least one significant DMP. Several genes which are significant in our gene expression analysis are also identified in our methylation data, including multiple genes in the homeobox family (HOXC8, HOXC9, HOXC10, HOXD10, HOXA11 AS), previously implicated as an early epigenetic event in multiple cancers [19]. NKX2-1 (TTF-1) is the only putative driver gene to be identified in both gene expression and methylation analyses, and is also a member of the homeobox family. It is hypermethylated and underexpressed in progressive samples compared to regressive. This gene is widely used in diagnosis of lung adenocarcinoma and both underexpression and hypermethylation have been implicated in the development of this disease [20], [21]. NKX2-1 loss has been shown to drive squamous cancer formation in combination with SOX2 overexpression [22]; focal gains in the 3q region containing SOX2 are commonly observed in progressive CIS (Fig. 10).

Principal component analysis of all gene expression and methylation data showed a clear distinction between the progressive and regressive subgroups (p=0.0017 and p=6.8x10-25, respectively) (Fig. 4C,D). In the methylation dataset, the regressive lesions closely clustered with the control normal epithelial cells. A history of chronic obstructive pulmonary disease (COPD) had an effect on case segregation (p=1.2xl0-5) but all other clinical and technical variables analysed, including smoking status and history of lung cancer, had no effect (Fig. 14). This was also the case for PCA analysis of the gene expression data (Fig. 14G-K).

For methylation, one control and four regressive cases clustered with the progressive cases (Fig. 4D). Three of the four miss-classified regressive cases were subjected to whole-genome sequencing and were found to have more copy number alterations than other regressive samples (PD21884a, PD21893a, PD21908a). Two of these correspond to the samples discussed above, which showed signs of progression after meeting the clinical end point of regression (Fig. 10). For the control bronchial epithelium sample that was classified with the progressive lesions, CIS was detected in a biopsy specimen 12 months later from the same site. Thus, although these cases were formally treated as miss-classifications, it is likely that the molecular data underpinning the apparent errors indicate a cellular phenotype that is not consistent with a

straightforward regressive lesion.

Example 5: Molecular signatures predict CIS outcome

The ability to predict if a pre-invasive lesion will progress to cancer has important clinical implications. For gene expression, the above pre-defined discovery set to define our classifier were used (n=33; 17 progressive, 16 regressive; 10-fold cross-validation applied). This was applied to a separate validation set (n=l 8; 10 progressive, 8 regressive). All samples in the validation set were classified correctly. When applied to external data from TCGA (n=551 : 502 LUSC, 49 control), the 291 -gene model was able to classify LUSC vs control samples with AUC=0.81 (Fig. 5A-C; Fig. 15).

An analogous analysis was performed for methylation using a discovery set of 60 samples and a validation set of 27 samples. This classified validation samples with AUC=0.99 and classified external TCGA samples (n=412: 370 LUSC, 42 controls) into LUSC vs controls with AUC=0.99, based on a 141 -DMP classifier (Fig. 16A-I).

Observed as an increased number of methylation probes with intermediate methylation in TCGA LUSC cancer vs TCGA control samples (Fig. 5D), reflecting methylation heterogeneity in these samples. We therefore developed a methylation heterogeneity index (MHI), defined as the number of probes per sample with t|0 < B < thi. Optimisation based on the discovery set of 26 progressive and 11 regressive samples defined values of ti0 = 0.26 and thi = 0.88. Control samples were not used in this analysis. This model classified progressive vs regressive CIS samples in our validation set with AUC=0.74 and TCGA LUSC vs TCGA control samples with AUC=0.96 (Fig. 5E; Fig. 16J-N). Multivariate logistic regression in our CIS cohort demonstrated that this index was a predictor of progression status (p=0.017); previous history of lung cancer was also significantly associated (p=0.02), whereas smoking status, COPD status, age and gender were not.

Given the widespread nature of methylation changes, it was hypothesised that this increase in heterogeneity may be a genome-wide process rather than specific to functional pathways. To test this theory, the predictive value of MHI calculated from a sample of 2,000 probes, randomly selected from across the genome, was assessed. Running 10,000 simulations with each using a different random sample of 2,000 probes gave a mean AUC for TCGA LUSC vs TCGA control of 0.95 (95% Cl 0.92-0.98) (Fig. 5F), and for progressive vs regressive CIS of 0.75 (95% Cl 0.69-0.82) (Fig. 16N).

These results are similar to those obtained using the entire set of 450,000 probes, suggesting that methylation heterogeneity is a genome-wide process. However, these AUC values are lower than those obtained from our predictive model based on just 141 differentially methylated positions, suggesting that specific methylation changes may be important, on this background of generalised change.

To build a predictive classifier based on copy number, copy number derived from methylation data was used to increase sample size and classified 46 of 54 samples correctly (Fig. 17A-C). A predictive classifier based on 154 predictive cytogenetic bands (CNV bands) that we identified classified 24/24 regressive samples and 9/12 progressive samples correctly (Fig. 17D-F). When applied to external data from TCGA (n=763: 524 LUSC, 239 control), this CNV model was able to classify LUSC vs control samples with AUC=0.98 (Fig. 17G-I).

Further analyses was performed using only one sample per patient to

demonstrate that our results are not dependent on multiple sampling. The first available sample for each patient was selected, with CIS samples prioritised over control samples for methylation data. Results are similar to our analysis above, validating our initial results.

Although it cannot be fully excluded that lesions meeting the end point for regression will progress in future, most patients in this cohort now have several years of follow up. Of 35 regressive lesions undergoing molecular profiling, mean follow up was 67 months (median 57 months, range 11 -150 months).

Example 6: CIN is an early marker of progression to cancer

To investigate possible drivers of tumori genic progression, a differential analysis of gene expression data between the progressive and regressive groups was

performed. 5 of the top 100 genes identified have been previously associated with chromosomal instability (CIN) [24], as defined by the previously published CIN70 signature [25] (ACTL6A, ELAVL1, MAD2L1, NEK2, OIP5). All five are up-regulated in progressive compared with regressive samples. CIN-related genes can predict progression (Fig. 6A); NEK2 expression alone predicts progression with AUC=0.93 (Fig. 6B).

Pathway analysis was performed using the gage Bioconductor package [26] to compare the differentially expressed genes to KEGG gene sets. The CIN70 gene set was the most significant gene set identified (adjusted p value 8.9x10-32; up-regulated in progressive group), suggesting a role in early tumorigenesis. Cell cycle and DNA repair pathways were also implicated (Fig. 6C). Results were similar when cell-cycle associated genes were removed from the CIN70 signature, suggesting that this is a genuine CIN signal rather than a marker of proliferation.

Performing similar differential analysis of differentially methylated probes found widespread changes. The top probes identified were associated with cancer-associated cell signalling pathways, including TGF-beta, WNT and Hedgehog, as well as cell cycle and CIN-associated genes (Fig 6D).

This CIN signal is consistent with the observed pattern of widespread copy number change (Fig. 3). Overall copy number variation for a sample, as measured by Weighted Genome Integrity Index (wGII) [27], correlates with mean CIN-associated gene expression of that sample (Pearson r2=0.473) (Fig. 18). It was also observed that a correlation between local copy number of a gene and expression of that gene, consistent with previous results [28], [29]

Example 7: Materials and Methods

Ethical approval

All tissue and bronchial brushing samples were obtained under written informed patient consent and were fully anonymised. Study approval was provided by the UCL/UCLH Local Ethics Committee (REC references 06/Q0505/12 and 01/0148).

Data availability

Whole-genome sequencing data have been deposited at the European Genome Phenome Archive (https://www.ebi.ac.uk/ega/ at the EBI) with accession number EGAD00001003883. All gene expression and methylation microarray data reported in this study have been deposited in the National Center for Biotechnology Information Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/) public repository, and they are accessible through GEO accession number GSE108124.

Code availability

All code used in our analysis will be made available at http://github.com/ucl-respiratory/preinvasive on publication. All software dependencies, full version information, and parameters used in our analysis can be found here.

Unless otherwise specified, all analyses were performed in an R statistical environment (v3.5.0; www.r-project.org/) using Bioconductorl version 3.7.

Biological samples

All patients with pre-invasive lung cancer lesions were recruited through University College London Hospitals (UCLH) Early Lung Cancer Surveillance

Programme (ELCSP). Full details of the surveillance protocol including eligibility criteria for patient inclusion have been previously described [2] Briefly, the programme has recruited 140 patients to date with pre-invasive lung cancer lesions of varying histological grades. Patients undergo autofluorescence bronchoscopy (AFB) and CT/PET scans every four to six months during which multiple biopsy specimens are collected. This longitudinal sequential AFB procedure provides biopsies of the same lesion sampled repeatedly over time, allowing us to monitor whether the individual lesions have progressed, regressed or remained static [2].

For a given CIS lesion under surveillance, when a biopsy from the same site showed evidence of progression to invasive cancer or regression to normal epithelium or low-grade dysplasia, we define the preceding CIS biopsy as the‘index’ lesion. An index lesion was defined as progressive if the subsequent biopsy at the same site showed invasive cancer, or as regressive if the subsequent biopsy showed normal epithelium or low-grade disease (metaplasia, mild or moderate dysplasia). Lesions which do not satisfy one of these end-points were excluded from this study. Patients with multiple fresh-frozen (FF) and formal in -fixed, paraffin-embedded (FFPE) tissue biopsies were identified for DNA methylation and gene expression analysis, respectively. Laser-capture micro-dissection (LCM) was used to selectively isolate CIS cells for molecular analysis, reducing the extent of contamination by stromal cells.

The following protocol was used to determine which profiling methods were applied to a given CIS lesion during our initial data collection phase:

• If FFPE samples were available, gene expression profiling was performed. For the first 33 samples (17 progressive and 16 regressive), gene expression profiles were generated using Illumina microarrays. Our predictive models are trained on this discovery set. Subsequently, a further set of 10 progressive and 8 regressive samples from 18 patients were profiled using a different microarray platform

(Affymetrix) to validate our findings on an independent platform.

• If FF samples were available, DNA from these samples was first used for methylation profiling. Samples with sufficient DNA after DNA profiling were additionally subjected to whole-genome sequencing. After acquisition of sufficient samples for our methylation dataset (54 samples; 36 progressive, 18 regressive), only 29 samples had sufficient DNA for WGS, therefore we prioritised WGS over methylation for the subsequent 10 samples.

Tissue processing and laser-capture micro-dissection

FF or FFPE tissue sections (7-10mM thickness) were mounted on a

MembraneSlide 1.0 PEN. Prior to cryosectioning, the slides were heat-treated for 4 h at 180°C in a drying cabinet to inactivate nucleases. To overcome the membrane’s hydrophobic nature and to allow better section adherence, the slides were then UV-treated for 30 min at 254nm. Prior to laser-capture micro-dissection (LCM), the slides containing the FF tissue sections for DNA extraction were washed in serial ethanol dilutions (50, 75, 100%) to remove the freezing medium (OCT) and to avoid any interference with the laser’s efficiency. For RNA extraction, FFPE sections were dewaxed using the Arcturus® Paradise® PLUS Reagent System (Applied Biosystems,

Foster City, CA, USA). For each case, epithelial areas of pre-invasive disease were identified by haematoxylin and eosin staining of the corresponding cryosection (~7 mM thick). The presence of epithelial areas of interest was confirmed by histological assessment of each case by two histopathologists. LCM to isolate the tissue area/cells of interest was performed with the PALM MicrobeamTM system (Carl Zeiss

Microimaging, Munich, Germany) on unstained sections. The micro-dissected material was catapulted into a 500pl AdhesiveCap that allows capture of the isolated tissue without applying any liquid into the cap prior to LCM, thus minimizing the risk of nuclease activity. The captured cells were stored at -80°C until DNA extraction or processed immediately for RNA.

DNA extraction

DNA from the micro-dissected tissue and bronchial brushing samples was extracted using QIAGEN’s QIAmp DNA Mini and Micro kits, respectively (Crawley, UK). Soluble carrier RNA was used to increase tissue DNA yield. Concentration was measured using the Qubit® dsDNA High- Sensitivity assay and Qubit® 2.0 Fluorometer (Life Technologies, Paisley, UK). Nucleic acid quality and purity was estimated based on the A260/280 absorbance ratio readings using the NanoDrop-8000 UV-spectrophotometer (Thermo Scientific, Hertfordshire, UK). Only samples with an A260/280 ratio of 1.7-1.9 were included in the study.

RNA extraction

RNA was extracted using the High Pure FFPE RNA Kit (Roche Applied Science, West Sussex, UK) according to manufacturer’s protocol. Quantification was carried out using the Quant-iT RNA assay kit and the Qubit® 2.0 fluorometer (Life Technologies, Paisley, UK). RNA integrity was analysed using a BioAnalyzer 2100 (Agilent, Stockport, UK).

Bisulfite conversion

For each sample undergoing methylation profiling, 200 ng of DNA were bisulfite converted using the EZ DNA methylation kit (Zymo Research Corp., Orange, CA, USA) according to the manufacturer’s modified protocol for Illumina’s Infmium 450K assay. This protocol incorporates a cyclic denaturation step to improve the conversion efficiency [3] The 10 mΐ final conversion reaction was concentrated down to 4 mΐ with a vacufuge plus vacuum concentrator (Eppendorf AG, Hamburg, Germany) and sent to UCL’s Genomics Core Facility for hybridization on the 450K BeadArray according to Illumina’s Infmium HD protocol (Illumina Inc., San Diego, CA, USA) as previously described.4

Infmium HumanMethylation450K raw data extraction and pre-processing

Illumina’s iScan fluorescent system was used to scan and image the arrays.

DNA methylation data were extracted as raw intensity signals without any prior background subtraction or data normalization and were stored as ID AT files.

CpG-specific methylation levels (b-values; continuous value ranging from 0 to 1) for each sample were calculated as the ratio of the fluorescent signal intensity of the methylated (M) and unmethylated (U) alleles according to the following formula:

p _ intensity of methylated allele (M)


intensity of [unmethylated (U) + methylated (M) allele] +100

All subsequent raw b-value pre-processing, normalisation and down-stream analysis was performed using the Chip Analysis Methylation Pipeline (ChAMP) Bioconductor package with default settings [5]

Analysis of differentially variable positions (DVP) was performed using iEVORA6. Beta values from ChAMP were used as input to iEVORA following normalization and batch correction.

Genome-wide gene expression array

The extracted FFPE RNA used to generate the gene expression profiles on the discovery set was sent to UCL’s Genomics Core Facility for hybridisation on the Human Whole-Genome DASL (cDNA-mediated Annealing, Selection, extension and

Ligation) beadarrays according to Illumina’ s protocol (Illumina Inc., San Diego, CA, USA).

The extracted FFPE RNA used to generate the gene expression profiles on the validation set was sent to UK Bioinformatics Limited for hybridisation on the

Clariom™ D Transcriptome Human Pico Assay 2.0 according to Affymetrix’s protocol (Thermo Fisher Scientific Waltham, MA, USA).

Principal Component Analysis (PCA)

In order to identify any potential factors of variability affecting sample/group segregation, we applied principal component analysis on all probes passing filters defined above (implemented in the prcomp method of the R stats package). Technical and biological variation was investigated for batch arrays, smoking (pack-years), age at initial diagnosis, gender and previous lung cancer history. The ability of these features to predict the first principal component was quantified using ANOVA analysis, implemented in the R aov method p-values quoted are derived from this method.

Gene expression analysis

Raw gene expression data were expressed as log2 ratios of fluorescence intensities of the experimental samples. Quantile normalization was applied to Illumina data, using proprietory Illumina software. For Affymetrix data, RMA normalization was applied as defined in the affy Bioconductor package. For analyses utilizing both data sets, only genes represented on both arrays were included and ComBat7 was used to adjust for batch effects.

Differential expression analysis was performed using the limma8 Bioconductor package. Raw p-values were adjusted by the Benjamini-Hochberg procedure to give a FDR [9]. A significance threshold of FDR < 0.01 was used to select differentially expressed genes. Cluster analysis and visualization was performed using the pheatmap [10] Bioconductor package.

Real Time PCR Validation

For microarray validation, total RNA from the 33 pre-invasive LUSC lesions undergoing Illumina gene expression profiling was reverse transcribed using qScriptTM cDNA Super-Mix (Quanta Biosciences, Lutterworth, UK) according to the

manufacturer’s protocol. Real-time quantitative PCR was carried out in eight genes using the SYBR-green master mix (Applied BioSystems, Bleiswijk, Netherlands) in an Eppendorf real-time PCR Machine (Eppendorf, Stevenage, UK). Findings were validated using quantitative PCR (qPCR) for four up-regulated (GAGE5, GPNMB, MMP12 and STC2) and four down-regulated (SPDEF, LM07, OBSCN and MT1 E) genes. Gene-specific primers were designed inside or nearby the microarray sequence targeted, using Primer Express Software (PE Applied Biosystems, Bleiswijk,

Netherlands). Relative gene expression was quantified using the threshold cycle (Ct) method and normalized to the amount of CTBL and CEP250, which met the criteria of less variation between samples and compatible expression level with the studied genes. Each sample was tested in triplicate and a sample without template was included in each run as a negative control. Correlations between microarrays and real time PCR data were measured using the Pearson coefficient. From microarray and real time PCR data, we calculated the progressive/regressive ratio for each gene expression. All eight genes tested were significant in our differential microarray analysis with FDR < 0.05. A high degree of correlation (r=0.982) was observed between qPCR and array data.

Predictive modelling

For methylation, gene expression and copy number data we applied Prediction Analysis of Microarrays (PAM)[1 1] to predict whether a sample was progressive or regressive based on its molecular profile. The Bioconductor PAMR package was used. In all presented analyses we select a threshold which minimizes the number of data inputs required whilst maintaining the minimum possible number of classification errors.

PAM calculates the probability of each sample being progressive. We describe this value as a‘Progression Score’. ROC analytics were performed on these

progression scores to determine their value as a diagnostic test, using the pROC12 and PRROC13 Bioconductor packages.

For methyl ation and gene expression data a predictive model was trained on the training set and subsequently applied to an independent validation set. Regressive and control samples were grouped together for the methylation data analysis. ROC analytics were performed only on the validation set. Internal cross-validation was used for methylation-derived copy number data due to smaller sample size (control samples are used as a baseline to calculate copy number, therefore are excluded from predictive analysis).

When multiple lesions from one patient were included in an analysis, these were treated as independent events as they were always taken from different sites in the lung. The outcome of a lesion (whether it progressed or regressed) was determined on a per-lesion basis; the lesion was assigned to the progressive group only if cancer developed at the same site in the lung, and to the regressive group only if normal or low-grade dysplasia was obtained from the same site in the lung.

In some cases different technologies were used, for example our gene expression discovery set used Illumina microarrays whereas our validation set used Affymetrix. In such instances, both data sets were reduced to the subset of genes covered by probes in both platforms prior to creating a predictive model. The ComBat method from the sva Bioconductor package was used to correct for batch effects between the different platforms. In the case of RNAseq data, we used the voom transformation defined in the limma Bioconductor package to derive data comparable to expression data prior to batch correction with ComBat.

Copy number variation analysis

For samples with whole-genome sequencing available we used ASCAT14 to derive local copy number estimates as described below. To increase our sample size for comparative analyses, copy number variation (CNV) data were obtained from non-normalised methylated and unmethylated signal intensities of probes in the 450K array as previously described [15] using the ChAMP Bioconductor package with default settings. Copy number (CN) profiles for progressive and regressive cases were obtained using the control cases for baseline normalisation. A previously defined threshold of ±0.3 was used for the identification of single CNV. Probes associated with highly polymorphic regions (e.g. major histocompatibility complex) were removed from the analysis. The analysis generated group CN frequency plots and CN profiles for each sample. For samples with both methylation and sequencing data available we observed good correlation between copy numbers derived from the two different methods.

For comparison with previous results, the ChAMP pipeline was then modified to return CNV values per-probe. Probe locations were matched to cytogenetic bands using the Ensembl GRCh37 assembly, obtained from

http://grch37.rest.ensembl.org/info/assembly/homo_sapiens7content-type=application/json&bands=l, such that copy number variation could be assessed by cytogenetic band. The mean CNV value for each of 778 cytogenetic bands was calculated for each of our 54 samples. Limma analysis was used to identify bands that differed significantly between progressive and regressive samples with BH-adjusted p-value < 0.05. Predictive modelling was performed using PAM to find bands predictive of progression, using the same method as for gene expression data. Due to the low number of regressive samples, an internal cross-validation method was used rather than separate discovery and validation sets.

Following identification of predictive cytogenetic bands, PAM modelling was repeated with the dataset limited to only those bands identified by van Boerdonk et al: 3q26.2-29, 3p26.3— p 11.1 and 6p25.3-p24.3.16,17. This model was also accurate.

Finally, we applied our model to the validation data set of 24 regressive and 12 progressive samples used by van Boerdonk et al (GEO accession number GSE45287). These data were measured using a different microarray platform (arrayCGH). We assigned each probe to a cytogenetic band, and took the mean values to create a matrix of expression values by band. Our model was applied to the subset of chromosomal bands present in both data sets (760 of 778 bands). ComBat was used for batch correction between the two platforms. Our model correctly predicted 24/24 regressive samples and 9/12 progressive samples, replicating the results of van Boerdonk et al.

External validation using TCGA

Lung cancer methylation datasets publically available through The Cancer Genome Atlas (TCGA) were downloaded using GenomicDataCommons download tools [18]. We obtained the normalized b-values of 370 LUSC samples and 42 normal controls. ComBat was used to correct for batch effects between our data and TCGA data. These data were used as an external validation set to test our predictive models, and as input for our differential analysis of progression drivers from control through CIS to cancer.

Gene-expression microarray data sets comparable to our data were not publically available. RNAseq data was available from TCGA for 502 LUSC samples and 49 control samples. We applied a voom transformation [19] to these data, which uses normalized log-counts-per-million as an approximation for expression values, and hence allows comparison of RNAseq data with our gene expression pipeline. ComBat was used to correct for batch effects. The predictive model generated using PAM on our gene expression microarray data was applied to voom-transformed RNAseq data from TCGA and shown to be predictive (Fig. 5C). We therefore demonstrate the applicability of our model to this fully independent data set. These data were again used as input to our differential analysis of progression drivers.

Pathway analysis

For gene expression data, the GAGE Bioconductor package [20] was used with KEGG gene sets [21 ]-[23] to identify pathways associated with genes differentially expressed in our analysis of progression to cancer (BH-adjusted p- value <0.01). In addition to these pathways we use the CIN70 signature defined by Carter et al. [24] to assess for a chromosomal instability signal. We also use a subset of the CIN70 genes with cell-cycle associated genes [25] removed to ensure that our signal is genuinely CIN-related, rather than a measure of proliferation.

Methylation data was analysed in the same way, using beta values as input to GAGE. In cases where there are multiple methylation probes for a single gene we use the mean beta value over that gene as input to pathway analysis. We acknowledge that using mean signal may be insensitive to single-probe methylation changes, however given the scale of changes observed we believe it will identify areas of large

methylation change.

Genomic sequencing

We created genome-wide shotgun libraries (insert size 331 -367 bp) from native DNA using the Agilent Technologies Custom SureSelect Library Prep Kit library (cat no. 930075). 150 bp paired-end sequence data were generated using the Illumina HiSeq X Ten system.

Sequenced data were realigned to the human genome (NCBI build 37) using BWA-MEM. Unmapped reads and PCR duplicates were removed. A minimum sequencing depth of 40x was required.

Somatic mutation calling and annotation

Single base somatic substitutions were identified by our in-house algorithm Cancer Variants through Expectation Maximisation (CaVEMan:

https://github.com/cancerit/CaVEMan) [26] This algorithm compares the sequence data from each tumour sample to its matched normal and calculates a mutation probability at each locus. This calculation incorporates information from aberrant cell fraction and copy number estimates from the Allele-Specific Copy number Analysis of Tumours (ASCAT) algorithm (https://www.crick.ac.uk/peter-van-loo/software/ASCAT)

[14], [27] Additional post-processing as described previously [28] was implemented. Any putative driver mutations were visually inspected with Jbrowse [29] For every substitution that passed all filters in at least one sample, we counted the number of wild-type and mutant reads at the same position in all other samples from the same patient to see if that mutation was also present in related samples but had not been called.

Somatic small insertions and deletions

These were identified using our in-house algorithm Pindel [30], [31]. As with substitutions, all putative driver mutations were visualised with Jbrowse.

Somatic structural variant detection

Abnormally paired read pairs were grouped using an in-house tool,“Brass” [32] Read groups overlapping genomic repeats, reads from the matched normal, or from a panel of unmatched normals were ignored. Read pair clusters were then filtered by read remapping. Read pair clusters with >50% of the reads mapping to microbial sequences were removed. Finally, candidate SV breakpoints were matched to copy number breakpoints as defined by ASCAT within 10 kb. Candidate SVs that were not associated with copy number segmentation breakpoints and with a copy number change of at least 0.3 were removed. All putative driver rearrangements were visually inspected using IGV [33], [34]

Somatic copy number events, ploidy, and stromal contamination

Copy number changes were derived from whole-genome sequencing data using the ASCAT algorithm. This algorithm compares the relative representation of heterozygous SNPs and the total read depth at these positions to estimate the aberrant cell fraction and ploidy for each sample, and then to determine allele-specific copy number.

Weighted Genome Integrity Index

To estimate the overall chromosomal instability of a sample, we use the Weighted Genome Integrity Index (wGII) score [35] This is calculated by measuring the percentage of the genome which is abnormal, corrected such that each chromosome is equally weighted.

Mutation annotation

Lung cancer driver genes were selected from the COSMIC Cancer Gene Census (CGC) v85 (cancer.sanger.ac.uk) [36]. CGC data was downloaded on 20th June 2018. Genes annotated in the CGC as potential drivers in lung cancer or NSCLC were included. Those specific to adenocarcinoma were excluded as our samples are precursors to squamous cancers. Genes identified in two large studies of squamous cell cancer, and some additional genes based on expert curation of the literature (ARID 1 A, AKT2, FAT1, PTPRB) were included if they were present in the CGC - even if they were not annotated explicitly as implicated in lung cancer. Both Tier 1 and Tier 2 genes were included. A total of 96 genes were selected as putative lung squamous cell carcinoma drivers.

Mutations affecting these putative driver genes were annotated as driver mutations if they passed the following filters:

• The mutation type (e.g. missense, frameshift, amplification) must have been validated in the CGC for the affected gene.

• For genes annotated as tumour suppressors, mutations determined to have High or Moderate impact using Ensembl’s Variant Effect Predictor [37] were classed as driver mutations.

• For genes annotated as oncogenes, we checked the specific mutation against COSMIC mutation data for lung carcinomas. If the specific mutation occurred 3 or more times in this dataset it was classed as a driver mutation.

• For genes annotated as fusion proteins, translocations with a

translocation partner gene matching validated tranlocation partner genes in the CGC were classed as driver events.

• Copy number amplifications and deletions were all classed as driver events if amplifications/deletions in the affected gene have been previously validated in the CGC. We included homozygous deletions of tumour suppressor genes and amplifications to more than double the sample ploidy for oncogenes.

Driver mutation discovery was also attempted using dndscv [38]. This was underpowered, however, and only yielded TP53 and CDKN2A as genes under positive selection. This package was also used to estimate the global dNdS for both progressive and regressive lesions.

Subclonality analysis

The number of subclones contributing to a sample and their relative contribution was estimated by using a modified version of the sciClone Bioconductor package [39] sciClone uses a Bayesian method to allocate mutations to clusters based on their variant allele frequency (VAF). By default, sciClone only considers regions that are copy

number neutral and LOH-free. Given the significant aneuploidy in our data set we overcame this limitation by clustering on cancer cell fraction (CCF) rather than VAF. Briefly, cancer cell fraction represents the fraction of cancer cells in which a given mutation is present, therefore clonal mutations will have CCF=1. Following the method of McGranahan et al. [40], we estimated the CCF for each mutation with a 95% confidence interval. Mutations for which 1 lay within this confidence interval were labelled as‘clonal’, other mutations as‘subclonal’.

CCF values for each mutation were then used as input to sciClone in place of VAF values to quantify clusters present (divided by 2 such that clonal mutations have a value of 0.5). As CCF corrects for local copy number, all regions were assumed to have copy number of 2, allowing sciClone to group mutations based only on their CCF estimates. A minimum tumour sequencing depth of 10 was required for each mutation.

Where more than one sample from a given patient was available, both one dimensional and multi-dimensional clustering were performed. Results from one dimensional clustering were used in the comparison of numbers of clones and proportion of clonal mutations between progressive and regressive lesions, in order to provide as fair a comparison as possible.

Extraction of mutational signatures

To obtain an approximate estimate of the contribution of different known mutational signatures to each sample, we used the MutationalPattems Bioconductor package41. As a reference set of mutational signatures, we used a table with the relative frequency of each of the 96 trinucleotide substitutions across 30 known mutation signatures, [42], [43] available through the COSMIC website

(http://cancer.sanger.ac.uk/cosmic/signatures).

After a first run which indicated the most likely contribution of each signature, it seemed that the majority of substitutions were contributed by signatures 1, 2, 4, 5, and 13, which have been described to be the strongest signatures in lung squamous cell cancer [44]. Some contribution was identified from signatures 16, 8, 18 and 3 in our initial analysis; however, in this context it is likely that these represent overfitting given that signature 16 is similar to signature 5, and signatures 8, 18 and 3 are similar to

signature 4. We therefore ran the algorithm a second time, this time only using a 5x96 matrix of mutational signatures 1, 2, 4, 5 and 13. All mutations were thus forced to belong to one of these five mutational signatures.

For a comparison of the clonal vs subclonal mutational processes in each sample, substitutions were annotated as clonal or subclonal based on CCF as described above. These were then run through the MutationalPattems package.

Comparison of mutational burden and signatures with other cancer types

Signatures of mutations in our CIS dataset were compared with mutational signatures found in lung squamous cell cancer. Raw whole-exome sequencing data for this cancer type was downloaded from TCGA, and run through our substitution-calling algorithm CAVEMaN as described above. We then looked at the total number of subsitutions called, and estimated the contribution of each mutational signature using the methods described above. Only coding regions of the CIS whole-genome sequencing data were compared to these exomes.

Estimation of telomere lengths

Telomere lengths were estimated using telomerecat [45], and were compared in progressive and regressive groups. Telomerecat is a de novo method for the estimation of telomere length (TL) from whole-genome sequencing samples. The algorithm works by comparing the ratio of full telomere reads to reads on the boundary between telomere and subtelomere. This ratio is transformed to a measure of length by taking into account the fragment length distribution. Telomerecat also corrects for error in sequencing reads by modelling the observed distribution of phred scores associated with mismatches in the telomere sequence. Samples were analysed in two groups corresponding to two separate sequencing batches, as per the telomerecat

documentation.

All publications mentioned in the above specification are herein incorporated by reference. Various modifications and variations of the described aspects of the invention will be apparent to those skilled in the art without departing from the scope and spirit of the invention. Although the invention has been described in connection with specific preferred embodiments, it should be understood that the invention as claimed should not be unduly limited to such specific embodiments. Indeed, various modifications of the described modes for carrying out the invention which are obvious to those skilled in genetics, epigenetics, molecular biology, cell biology, oncology or related fields are intended to be within the scope of the following claims.

Tables

Table 1 - Example DEG signature comprising 397 genes (gene weights are given in columns XO. score and XI. score)

GENE ID XO.score X1.score GENE ID XO.score X1.score GENE ID XO.score X1.score

KLK7 0.4684 -0.4408 GHR -0.2436 0.2293 CBS -0.1736 0.1634

SPNS2 0.4353 -0.4097 LHX9 0.2426 -0.2284 HTR2A 0.173 -0.1628

CPVL -0.3878 0.365 ABCG4 0.2393 -0.2252 CD164L2 0.1705 -0.1604

FER1 L6 0.3719 -0.3501 GABRB1 0.2331 -0.2194 SLC29A4 -0.1704 0.1603

ATP12A 0.3663 -0.3448 ZIC2 -0.2298 0.2163 RPL7 -0.1687 0.1588

KLK12 0.3631 -0.3418 KRTAP13-1 0.2215 -0.2085 ZFP37 -0.1685 0.1586

ADCYAP1 0.344 -0.3238 LCE6A 0.2119 -0.1994 SLC6A11 0.1685 -0.1586

RASIP1 0.3429 -0.3227 MIOX 0.2066 -0.1944 PLAT 0.168 -0.1581

KLK6 0.341 -0.3209 CHST6 0.2056 -0.1935 PRSS3 0.1666 -0.1568

KLK5 0.3229 -0.3039 RBAK -0.2054 0.1933 ELAVL1 -0.1662 0.1564

NCCRP1 0.3167 -0.2981 CSN3 0.2006 -0.1888 SOX17 0.1659 -0.1562

RHCG 0.3159 -0.2973 ASCL2 -0.1953 0.1838 SLC13A5 0.1657 -0.156

IGFL1 0.3087 -0.2905 MPP6 -0.1947 0.1832 SLC6A14 0.1649 -0.1552

GDPD3 0.2877 -0.2708 LYPD3 0.1944 -0.1829 SMYD3 -0.1599 0.1505

MLPH 0.2863 -0.2695 ZNF804B 0.1939 -0.1825 ITPKA -0.159 0.1497

OR5B3 0.2846 -0.2679 RSRC1 -0.1937 0.1823 STAB2 0.1574 -0.1482

MT1 F -0.2805 0.264 PANX3 0.1936 -0.1822 IFF02 0.1559 -0.1467

OR52K1 0.2799 -0.2634 SPRR2B 0.1913 -0.1801 MYNN -0.1544 0.1453

KRT23 0.2789 -0.2625 TMPRSS11 D 0.1906 -0.1794 PDE3B -0.1539 0.1448

DHRS9 0.2766 -0.2603 UNC93A 0.1887 -0.1776 KLK10 0.1524 -0.1434

SULT2B1 0.2723 -0.2563 INO80B -0.1864 0.1754 HIST1 H2AA 0.1512 -0.1423

CST6 0.2614 -0.246 CLIC3 0.186 -0.175 ZNF614 -0.1508 0.1419

SBSN 0.2552 -0.2402 MAD2L1 -0.1854 0.1745 H19 0.1497 -0.1409

PRAME -0.2552 0.2402 DNAJC19 -0.1852 0.1743 GPNMB -0.149 0.1402

NEK2 -0.2518 0.237 SNORD2 -0.1757 0.1654 E2F7 -0.1486 0.1398

Table 1 (continued)

GENE ID XO.score X1.score GENE ID XO.score X1.score GENE ID XO.score X1.score

SERPINB4 0.1475 -0.1388 ZNF721 -0.1296 0.122 SP6 0.1098 -0.1034

SLC4A11 0.1468 -0.1382 SKAP2 -0.1266 0.1191 PBOV1 0.109 -0.1026

HES5 0.1458 -0.1372 IRF7 0.1253 -0.1179 ID1 0.1087 -0.1023

ALOX12B 0.1453 -0.1367 MEOX1 0.1252 -0.1178 ANLN -0.1077 0.1014

VAX2 -0.1445 0.136 TEX14 0.1248 -0.1175 MIPOL1 -0.1076 0.1013

FAM3D 0.1438 -0.1354 G6PC 0.124 -0.1167 C19orf33 0.1073 -0.101

OR52B6 0.1437 -0.1353 CCDC59 -0.1234 0.1162 SCEL 0.1059 -0.0997

OR2B11 0.1429 -0.1345 SCML2 -0.1223 0.1151 ZNF284 0.1051 -0.0989

KLK8 0.1427 -0.1343 MAGEA9B -0.1213 0.1142 MIR554 0.103 -0.0969

NKAIN2 -0.1425 0.1341 ZSCAN4 0.1206 -0.1135 SH3TC1 0.1023 -0.0963

NOS1 0.1412 -0.1329 CRABP2 0.1199 -0.1129 STC2 -0.1022 0.0962

PEG10 -0.1412 0.1329 TRIM16 0.1198 -0.1128 DEFB122 0.1019 -0.0959

HOXC8 -0.141 0.1327 TFAP4 -0.1197 0.1126 OR14J1 0.1 -0.0941

ABHD8 -0.1396 0.1314 C19orf54 -0.1179 0.1 1 1 ZNF777 -0.1 0.0941

CLDN5 0.1386 -0.1304 HIST1 H4F 0.117 0.1 101 NRSN1 0.0999 -0.094

MCM10 -0.1378 0.1297 IFNA21 0.1163 -0.1095 FRAS1 -0.0987 0.0929

PI3 0.1378 -0.1297 CA1 0.1156 -0.1088 ECM1 0.0986 -0.0928

ADCY4 0.1367 -0.1286 ZYG11A 0.115 -0.1082 ANGPT4 0.0984 -0.0926

HOXC9 -0.1365 0.1285 CDA 0.1146 -0.1079 GABRP 0.0979 -0.0921

FANCL -0.1351 0.1271 SERPINB3 0.1145 -0.1078 OIP5 -0.0967 0.0911

RDH13 0.1347 -0.1268 GRM3 0.1139 -0.1072 DNAJC5G 0.0965 -0.0908

NOD2 0.1329 -0.1251 KRTDAP 0.1129 -0.1062 HOXD10 -0.0965 0.0908

SORBS1 0.1328 -0.125 PRSS27 0.1109 -0.1044 B3GALT4 0.0965 -0.0908

KRTAP8-1 0.1328 -0.1249 VPS37D -0.1105 0.104 IGF2BP3 -0.0947 0.0891

E2F3 -0.1312 0.1235 MYL3 0.11 -0.1035 UBL4B 0.094 -0.0884

C2orf78 0.1305 -0.1228 EBF1 0.11 -0.1035 IL4R 0.0938 -0.0883

Table 1 (continued)

GENE ID XO.score X1.score GENE ID XO.score X1.score GENE ID XO.score X1.score

ACTL6A -0.0938 0.0882 MUC16 0.0833 -0.0784 KRT16 0.0655 -0.0616

PCDHB13 -0.0937 0.0882 HOXC10 -0.083 0.0781 RNF126P1 0.0654 -0.0616

NKX2-1 0.0934 -0.0879 ZNF124 -0.0822 0.0773 PFN2 -0.0641 0.0603

SSX2 0.0933 -0.0879 PTPRB 0.0813 -0.0765 PLD2 0.0634 -0.0597

HIST2H3A -0.0932 0.0877 TMEM41A -0.0806 0.0759 KCNS1 0.0631 -0.0594

MND1 -0.0931 0.0877 INPP4B 0.0806 -0.0759 KCNJ3 0.0626 -0.0589

SCN8A -0.0931 0.0876 FBX016 0.0796 -0.0749 CLCF1 0.062 -0.0584

OR5B12 0.0928 -0.0874 TMEM45B 0.0788 -0.0742 TFB2M -0.062 0.0583

CLDN16 -0.0923 0.0869 MED28 -0.0783 0.0737 ZNF131 -0.0619 0.0582

SLC17A4 0.092 -0.0866 SORCS2 0.0782 -0.0736 ECT2 -0.0617 0.058

RRM2B -0.0916 0.0862 GSTM5 0.0782 -0.0736 CKAP2 -0.061 1 0.0575

GATA4 -0.0914 0.086 OR51 Q1 0.0777 -0.0731 TMEM40 0.061 1 -0.0575

TGM5 0.0903 -0.085 OR4E2 0.0771 -0.0725 GADL1 0.0609 -0.0573

LOR 0.0901 -0.0848 CXXC5 0.0767 -0.0722 ZDHHC13 0.0601 -0.0566

KRT78 0.09 -0.0847 NGEF 0.0739 -0.0695 RAB36 0.0601 -0.0566

CRIP1 0.0896 -0.0843 SPACA5B 0.0735 -0.0691 ALPL 0.0601 -0.0566

ALDH7A1 0.0895 -0.0843 DUXAP3 0.0729 -0.0687 RAB26 0.0597 -0.0562

PALMD 0.089 -0.0837 UBE2W -0.0724 0.0682 DLG1 -0.0595 0.056

S100A7 0.0888 -0.0835 EHHADH -0.0713 0.0671 RAB33B -0.0593 0.0558

REG1 B 0.0883 -0.0831 SPRR2E 0.0686 -0.0645 PPOX 0.0592 -0.0557

MIDI -0.0878 0.0827 ZNF121 -0.0684 0.0644 CYMP 0.0589 -0.0555

CXCL10 -0.0877 0.0826 ECSCR 0.0678 -0.0638 SIM2 -0.0573 0.0539

HSD17B3 0.0876 -0.0824 FOXI2 0.0677 -0.0637 OR4F4 0.0572 -0.0538

HCN1 0.0852 -0.0802 FGD5 0.0665 -0.0626 PAK2 -0.0571 0.0538

PRSS22 0.0834 -0.0785 LY6G6C 0.0658 -0.0619 NMD3 -0.0563 0.053

SPP1 -0.0834 0.0785 INTS10 0.0658 -0.0619 FBN3 -0.0559 0.0526

Table 1 ( continued)

GENE ID XO.score X1.score GENE ID XO.score X1.score GENE ID XO.score X1.score

PSCA 0.0556 -0.0523 SNORD87 -0.0471 0.0443 NCBP2 -0.0345 0.0324

OR5AR1 0.0552 -0.0519 SCARA5 0.0461 -0.0434 TRIB1 0.0335 -0.0315

LCA5L 0.0548 -0.0516 DHRS2 -0.0455 0.0429 ZNF10 -0.0332 0.0312

MMP12 -0.0541 0.051 ANKRD35 0.0452 -0.0425 TRIO -0.033 0.031 1

TYW1 -0.0536 0.0504 LRP5 -0.0441 0.0415 SLC5A8 0.0322 -0.0303

MDH1 B 0.0535 -0.0503 OR52K2 0.044 -0.0414 FBXL18 -0.0322 0.0303

C15orf62 0.0533 -0.0502 RPL39L -0.0438 0.0412 MYH14 0.032 -0.0301

BTBD2 -0.0532 0.05 C1 orf1 16 0.043 -0.0405 OR2AT4 0.0313 -0.0295

ZNF703 -0.0531 0.05 HLA-DQA2 0.0428 -0.0402 PQLC2 -0.0312 0.0294

PPFIA2 0.0529 -0.0498 OR52R1 0.0419 -0.0394 LRIG1 0.0312 -0.0294

MRAP 0.0525 -0.0494 CEACAM5 0.0418 -0.0394 HLTF -0.0309 0.0291

MYOM3 0.0524 -0.0493 EIF2AK1 -0.041 0.0386 RAD51AP1 -0.0308 0.029

LCN2 0.0517 -0.0487 EPS8L1 0.0391 -0.0368 SDCBP2 0.0306 -0.0288

ZMYND15 0.0517 -0.0486 SLC22A8 0.039 -0.0367 GIPC2 0.0304 -0.0286

FUT3 0.0507 -0.0477 SNW1 -0.0385 0.0363 LACTB2 -0.0304 0.0286

GDF3 0.0505 -0.0475 C1 QTNF1 0.0384 -0.0362 FAM133A -0.0303 0.0285

REM1 0.0502 -0.0473 MPV17L 0.0382 -0.036 ACADVL 0.0303 -0.0285

NOL10 -0.05 0.0471 ATP8A2 0.0375 -0.0353 SFTA3 0.0294 -0.0277

RUNDC3B 0.05 -0.047 CD3EAP -0.0371 0.0349 BOC 0.0293 -0.0276

MYOCD 0.0495 -0.0466 ANXA1 0.037 -0.0348 CYP2E1 0.0287 -0.027

SPRR4 0.0494 -0.0465 B3GALNT1 -0.0367 0.0346 ZNF91 -0.0286 0.0269

KRT7 0.0487 -0.0459 TMEM183B -0.0367 0.0345 C1 D -0.0279 0.0263

HIST1 H2BH -0.0487 0.0458 GPR152 0.0361 -0.034 PNCK -0.0277 0.0261

MAB21 L2 0.0483 -0.0454 CTSG 0.0355 -0.0334 TM2D3 -0.0275 0.0259

MAS1 0.0477 -0.0449 RASL12 0.0354 -0.0333 OR1 1 H6 0.0267 -0.0252

FZD4 0.0471 -0.0444 CLEC4G 0.0347 -0.0326 RIOK1 -0.0263 0.0247

Table 1 (continued)

GENE ID XO.score X1.score GENE ID XO.score X1.score GENE ID XO.score X1.score

CRIP2 0.0262 -0.0246 SPINK5 0.0166 -0.0156 KRTAP4-7 0.0105 -0.0098

OR2L13 0.0261 -0.0246 CD300LD 0.0164 -0.0155 ANKS6 -0.0099 0.0093

PGBD5 0.0259 -0.0244 FXR1 -0.016 0.015 C3orf30 0.0098 -0.0092

SERPINB1 0.0257 -0.0242 TMC5 0.0152 -0.0143 SFRP5 0.0095 -0.009

MTIF3 0.0246 -0.0232 DEFB125 0.0141 -0.0132 HIST1 H3G -0.0092 0.0087

PIWIL3 0.0243 -0.0229 MIR548I1 0.014 -0.0131 USP1 1 -0.009 0.0085

VM01 0.024 -0.0226 SNX32 0.0139 -0.0131 FES 0.0072 -0.0068

TTC30A -0.0238 0.0224 IL33 0.0139 -0.0131 GPR68 0.0071 -0.0067

PN01 -0.0234 0.0221 CDCA7L -0.0129 0.0122 PLCD1 0.0071 -0.0066

TTYH3 -0.0223 0.021 KCNG1 -0.0129 0.0121 TIAM1 0.0061 -0.0057

HIST3H2BB 0.022 0.0207 CTNS 0.0128 0.0121 KPNA2 -0.0059 0.0055

PLEKHH3 0.0217 -0.0204 RAB23 -0.0128 0.012 ACAP2 -0.0059 0.0055

NES 0.0217 -0.0204 ETNK2 -0.0128 0.012 AGTPBP1 -0.0058 0.0055

SDK1 -0.0216 0.0204 BAG2 -0.0128 0.012 LMAN1 -0.0057 0.0054

GTF2H3 -0.0214 0.0201 DCUN1 D5 -0.0126 0.01 19 OR9I1 0.0055 -0.0052

SOX18 0.021 1 -0.0199 XKR3 0.0124 -0.01 16 CHI3L1 0.0053 -0.005

TRIM16L 0.0206 -0.0194 ANXA2 0.0123 -0.01 16 SLC6A5 0.0052 -0.0049

MY01 C 0.0201 -0.0189 ARHGAP30 0.012 -0.01 13 KLHL24 -0.0052 0.0049

TMEM139 0.0194 -0.0183 DSCC1 0.012 0.0112 GALE 0.005 -0.0047

LYPD2 0.0192 -0.0181 SERPINB8 0.01 17 0.01 1 MMP28 0.0046 -0.0044

KRT6B 0.0189 -0.0178 LARP4 -0.01 15 0.0108 CCL14 0.0046 -0.0043

KIF13B 0.0188 -0.0177 KLHL13 0.0112 0.0105 PROX2 0.004 -0.0038

ADAM29 0.0186 -0.0175 CSF2RB 0.0109 0.0102 IL23A 0.0039 -0.0036

PSAT1 -0.0183 0.0172 FGD6 -0.0108 0.0102 ICA1 0.0032 -0.003

PLTP -0.0174 0.0163 CYB5R2 0.0108 0.0102 PARP1 -0.0031 0.0029

LSM5 -0.0173 0.0163 BTN1A1 0.0107 0.0101 NEK3 0.0029 -0.0027

Table 1 (continued)

GENE ID XO.score X1.score

PON1 0.0026 -0.0024

MTHFD2L -0.0019 0.0018

WDHD1 -0.0013 0.0012

HOXC6 -0.0013 0.0012

S100P 0.0013 -0.0012

ALOX12P2 0.0007 -0.0006

SNORD1A -0.0006 0.0005

C6orf89 0.0005 -0.0005

BIRC5 -0.0003 0.0003

Table 2 - Example DEG signature comprising 291 genes (gene weights are given in columns XO.score and XL score)

GENE ID XO.score X1.score GENE ID XO.score X1.score GENE ID XO.score X1.score

KLK7 0.4372 -0.4115 GHR -0.2124 0.1999 CBS -0.1424 0.134

SPNS2 0.4041 -0.3803 LHX9 0.2114 -0.199 HTR2A 0.1417 -0.1334

CPVL -0.3566 0.3356 ABCG4 0.2081 -0.1958 CD164L2 0.1392 -0.131

FER1 L6 0.3407 -0.3207 GABRB1 0.2019 -0.19 SLC29A4 -0.1391 0.1309

ATP12A 0.3351 -0.3154 ZIC2 -0.1986 0.1869 RPL7 -0.1375 0.1294

KLK12 0.3319 -0.3124 KRTAP13-1 0.1903 -0.1791 ZFP37 -0.1373 0.1292

ADCYAP1 0.3128 -0.2944 LCE6A 0.1806 -0.17 SLC6A11 0.1373 -0.1292

RASIP1 0.3117 -0.2934 MIOX 0.1754 -0.165 PLAT 0.1367 -0.1287

KLK6 0.3098 -0.2915 CHST6 0.1744 -0.1641 PRSS3 0.1353 -0.1274

KLK5 0.2917 -0.2745 RBAK -0.1742 0.1639 ELAVL1 -0.135 0.127

NCCRP1 0.2855 -0.2687 CSN3 0.1694 -0.1594 SOX17 0.1347 -0.1268

RHCG 0.2847 -0.2679 ASCL2 -0.164 0.1544 SLC13A5 0.1345 -0.1266

IGFL1 0.2774 -0.2611 MPP6 -0.1635 0.1539 SLC6A14 0.1336 -0.1258

GDPD3 0.2565 -0.2414 LYPD3 0.1632 -0.1536 SMYD3 -0.1287 0.121 1

MLPH 0.2551 -0.2401 ZNF804B 0.1627 -0.1531 ITPKA -0.1278 0.1203

OR5B3 0.2534 -0.2385 RSRC1 -0.1625 0.1529 STAB2 0.1262 -0.1188

MT1 F -0.2493 0.2346 PANX3 0.1624 -0.1528 IFF02 0.1247 -0.1173

OR52K1 0.2486 -0.234 SPRR2B 0.1601 -0.1507 MYNN -0.1232 0.116

KRT23 0.2477 -0.2331 TMPRSS11 D 0.1594 -0.15 PDE3B -0.1227 0.1155

DHRS9 0.2454 -0.2309 UNC93A 0.1574 -0.1482 KLK10 0.121 1 -0.114

SULT2B1 0.241 -0.2269 INO80B -0.1551 0.146 HIST1 H2AA 0.12 -0.1129

CST6 0.2302 -0.2166 CLIC3 0.1548 -0.1457 ZNF614 -0.1196 0.1126

SBSN 0.224 -0.2108 MAD2L1 -0.1541 0.1451 H19 0.1185 -0.1115

PRAME -0.2239 0.2108 DNAJC19 -0.154 0.1449 GPNMB -0.1178 0.1108

NEK2 -0.2205 0.2076 SNORD2 -0.1445 0.136 E2F7 -0.1173 0.1104

Table 2 (continued)

GENE ID XO.score X1.score GENE ID XO.score X1.score GENE ID XO.score X1.score

SERPINB4 0.1 162 -0.1094 C2orf78 0.0992 -0.0934 MYL3 0.0787 -0.0741

SLC4A1 1 0.1 156 -0.1088 ZNF721 -0.0984 0.0926 EBF1 0.0787 -0.0741

HES5 0.1 145 -0.1078 SKAP2 -0.0954 0.0898 SP6 0.0786 -0.074

ALOX12B 0.1 14 -0.1073 IRF7 0.094 -0.0885 PBOV1 0.0778 -0.0732

VAX2 -0.1 133 0.1066 MEOX1 0.0939 -0.0884 ID1 0.0775 -0.0729

FAM3D 0.1 126 -0.106 TEX14 0.0936 -0.0881 ANLN -0.0765 0.072

OR52B6 0.1 125 -0.1059 G6PC 0.0927 -0.0873 MIPOL1 -0.0764 0.0719

OR2B1 1 0.1 1 17 -0.1051 CCDC59 -0.0922 0.0868 C19orf33 0.0761 -0.0716

KLK8 0.1 1 14 -0.1049 SCML2 -0.091 0.0857 SCEL 0.0747 -0.0703

NKAIN2 -0.1 1 13 0.1047 MAGEA9B -0.0901 0.0848 ZNF284 0.0739 -0.0695

NOS1 0.1 1 -0.1035 ZSCAN4 0.0894 -0.0841 MIR554 0.0718 -0.0676

PEG10 -0.1 1 0.1035 CRABP2 0.0887 -0.0835 SH3TC1 0.071 1 -0.0669

HOXC8 -0.1098 0.1033 TRIM16 0.0886 -0.0834 STC2 -0.0709 0.0668

ABHD8 -0.1083 0.102 TFAP4 -0.0885 0.0833 DEFB122 0.0707 -0.0665

CLDN5 0.1073 -0.101 C19orf54 -0.0867 0.0816 OR14J1 0.0688 -0.0647

MCM10 -0.1066 0.1003 HIST1 H4F 0.0858 -0.0807 ZNF777 -0.0687 0.0647

PI3 0.1066 -0.1003 IFNA21 0.0851 -0.0801 NRSN1 0.0687 -0.0646

ADCY4 0.1055 -0.0993 CA1 0.0844 -0.0794 FRAS1 -0.0675 0.0635

HOXC9 -0.1053 0.0991 ZYG1 1A 0.0838 -0.0788 ECM1 0.0674 -0.0634

FANCL -0.1038 0.0977 CDA 0.0834 -0.0785 ANGPT4 0.0671 -0.0632

RDH13 0.1035 -0.0974 SERPINB3 0.0833 -0.0784 GABRP 0.0667 -0.0627

NOD2 0.1017 -0.0957 GRM3 0.0827 -0.0778 OIP5 -0.0655 0.0617

SORBS1 0.1016 -0.0956 KRTDAP 0.0817 -0.0769 DNAJC5G 0.0653 -0.0614

KRTAP8-1 0.1015 -0.0956 PRSS27 0.0797 -0.075 HOXD10 -0.0653 0.0614

E2F3 -0.1 0.0941 VPS37D -0.0792 0.0746 B3GALT4 0.0652 -0.0614

Table 2 (continued)

GENE ID XO.score X1.score GENE ID XO.score X1.score GENE ID XO.score X1.score

IGF2BP3 -0.0635 0.0597 HSD17B3 0.0563 -0.053 ECSCR 0.0365 -0.0344

UBL4B 0.0627 -0.0591 HCN1 0.0539 -0.0508 FOXI2 0.0364 -0.0343

IL4R 0.0625 -0.0589 PRSS22 0.0522 -0.0491 FGD5 0.0353 -0.0332

ACTL6A -0.0625 0.0589 SPP1 -0.0521 0.0491 LY6G6C 0.0346 -0.0326

PCDHB13 -0.0625 0.0588 MUC16 0.0521 -0.049 INTS10 0.0345 -0.0325

NKX2-1 0.0622 -0.0585 HOXC10 -0.0518 0.0487 KRT16 0.0342 -0.0322

SSX2 0.0621 -0.0585 ZNF124 -0.0509 0.0479 RNF126P1 0.0342 -0.0322

HIST2H3A -0.0619 0.0583 PTPRB 0.05 -0.0471 PFN2 -0.0328 0.0309

MND1 -0.0619 0.0583 TMEM41A -0.0494 0.0465 PLD2 0.0322 -0.0303

SCN8A -0.0618 0.0582 INPP4B 0.0494 -0.0465 KCNS1 0.0319 -0.03

OR5B12 0.0616 -0.058 FBX016 0.0484 -0.0456 KCNJ3 0.0314 -0.0296

CLDN16 -0.061 1 0.0575 TMEM45B 0.0476 -0.0448 CLCF1 0.0308 -0.029

SLC17A4 0.0608 -0.0572 MED28 -0.0471 0.0443 TFB2M -0.0308 0.0289

RRM2B -0.0604 0.0568 SORCS2 0.047 -0.0442 ZNF131 -0.0307 0.0289

GATA4 -0.0602 0.0566 GSTM5 0.047 -0.0442 ECT2 -0.0304 0.0287

TGM5 0.0591 -0.0556 OR51 Q1 0.0465 -0.0437 CKAP2 -0.0299 0.0281

LOR 0.0589 -0.0554 OR4E2 0.0458 -0.0431 TMEM40 0.0298 -0.0281

KRT78 0.0588 -0.0553 CXXC5 0.0455 -0.0428 GADL1 0.0297 -0.0279

CRIP1 0.0584 -0.0549 NGEF 0.0427 -0.0402 ZDHHC13 0.0289 -0.0272

ALDH7A1 0.0583 -0.0549 SPACA5B 0.0422 -0.0398 RAB36 0.0289 -0.0272

PALMD 0.0577 -0.0543 DUXAP3 0.0417 -0.0393 ALPL 0.0289 -0.0272

S100A7 0.0575 -0.0541 UBE2W -0.0412 0.0388 RAB26 0.0285 -0.0268

REG1 B 0.0571 -0.0537 EHHADH -0.0401 0.0377 DLG1 -0.0283 0.0266

MIDI -0.0566 0.0533 SPRR2E 0.0373 -0.0351 RAB33B -0.0281 0.0264

CXCL10 -0.0565 0.0532 ZNF121 -0.0372 0.035 PPOX 0.028 -0.0263

Table 2 (continued)

GENE ID XO.score X1.score GENE ID XO.score X1.score GENE ID XO.score X1.score

CYMP 0.0277 -0.0261 MYOCD 0.0183 -0.0172 CD3EAP -0.0059 0.0055

SIM2 -0.0261 0.0245 SPRR4 0.0182 -0.0171 ANXA1 0.0058 -0.0054

OR4F4 0.0259 -0.0244 KRT7 0.0175 -0.0165 B3GALNT1 -0.0055 0.0052

PAK2 -0.0259 0.0244 HIST1 H2BH -0.0175 0.0165 TMEM183B -0.0055 0.0051

NMD3 -0.0251 0.0236 MAB21 L2 0.0171 -0.0161 GPR152 0.0049 -0.0046

FBN3 -0.0247 0.0232 MAS1 0.0165 -0.0155 CTSG 0.0042 -0.004

PSCA 0.0243 -0.0229 FZD4 0.0159 -0.015 RASL12 0.0042 -0.004

OR5AR1 0.0239 -0.0225 SNORD87 -0.0158 0.0149 CLEC4G 0.0034 -0.0032

LCA5L 0.0236 -0.0222 SCARA5 0.0149 -0.014 NCBP2 -0.0032 0.003

MMP12 -0.0229 0.0216 DHRS2 -0.0143 0.0135 TRIB1 0.0023 -0.0021

TYW1 -0.0223 0.021 ANKRD35 0.014 -0.0131 ZNF10 -0.0019 0.0018

MDH1 B 0.0222 -0.0209 LRP5 -0.0129 0.0121 TRIO -0.0018 0.0017

C15orf62 0.0221 -0.0208 OR52K2 0.0127 0.012 SLC5A8 0.001 -0.0009

BTBD2 -0.0219 0.0206 RPL39L -0.0125 0.01 18 FBXL18 -0.0009 0.0009

ZNF703 -0.0219 0.0206 Cl orfl 16 0.01 18 0.01 1 1 MYH14 0.0008 -0.0007

PPFIA2 0.0217 -0.0204 HLA-DQA2 0.01 15 -0.0109 OR2AT4 0.0001 -0.0001

MRAP 0.0212 -0.02 OR52R1 0.0107 0.01

MYOM3 0.0212 -0.0199 CEACAM5 0.0106 0.01

LCN2 0.0205 -0.0193 EIF2AK1 -0.0098 0.0092

ZMYND15 0.0205 -0.0193 EPS8L1 0.0079 -0.0074

FUT3 0.0195 -0.0183 SLC22A8 0.0077 -0.0073

GDF3 0.0193 -0.0181 SNW1 -0.0073 0.0069

REM1 0.019 -0.0179 C1 QTNF1 0.0072 -0.0068

NOL10 -0.0188 0.0177 MPV17L 0.007 -0.0066

RUNDC3B 0.0187 -0.0176 ATP8A2 0.0062 -0.0059

Table 3 - Example DEG signature comprising 211 genes (gene weights are given in columns XO.score and XI. score)

GENE ID XO.score X1.score GENE ID XO.score X1.score GENE ID XO.score X1.score

KLK7 0.4059 -0.3821 NEK2 -0.1893 0.1782 DNAJC19 -0.1227 0.1155

SPNS2 0.3728 -0.3509 GHR -0.1812 0.1705 SNORD2 -0.1133 0.1066

CPVL -0.3254 0.3062 LHX9 0.1802 -0.1696 CBS 0.1 1 1 1 0.1046

FER1 L6 0.3095 -0.2913 ABCG4 0.1768 -0.1664 HTR2A 0.1105 -0.104

ATP12A 0.3039 -0.286 GABRB1 0.1707 -0.1606 CD164L2 0.108 -0.1017

KLK12 0.3007 -0.283 ZIC2 -0.1674 0.1575 SLC29A4 -0.1079 0.1016

ADCYAP1 0.2816 -0.265 KRTAP13-1 0.159 -0.1497 RPL7 -0.1062 0.1

RASIP1 0.2805 -0.264 LCE6A 0.1494 -0.1406 ZFP37 -0.1061 0.0998

KLK6 0.2785 -0.2622 MIOX 0.1441 -0.1357 SLC6A11 0.106 -0.0998

KLK5 0.2604 -0.2451 CHST6 0.1431 -0.1347 PLAT 0.1055 -0.0993

NCCRP1 0.2543 -0.2393 RBAK -0.143 0.1345 PRSS3 0.1041 -0.098

RHCG 0.2535 -0.2386 CSN3 0.1382 -0.13 ELAVL1 -0.1037 0.0976

IGFL1 0.2462 -0.2317 ASCL2 -0.1328 0.125 SOX17 0.1035 -0.0974

GDPD3 0.2253 -0.212 MPP6 -0.1322 0.1245 SLC13A5 0.1033 -0.0972

MLPH 0.2239 -0.2107 LYPD3 0.1319 -0.1242 SLC6A14 0.1024 -0.0964

OR5B3 0.2222 -0.2091 ZNF804B 0.1315 -0.1237 SMYD3 -0.0974 0.0917

MT1 F -0.2181 0.2052 RSRC1 -0.1312 0.1235 ITPKA -0.0966 0.0909

OR52K1 0.2174 -0.2046 PANX3 0.1312 -0.1234 STAB2 0.095 -0.0894

KRT23 0.2165 -0.2037 SPRR2B 0.1289 -0.1213 IFF02 0.0935 -0.088

DHRS9 0.2141 -0.2015 TMPRSS11 D 0.1281 -0.1206 MYNN -0.092 0.0866

SULT2B1 0.2098 -0.1975 UNC93A 0.1262 -0.1188 PDE3B -0.0914 0.0861

CST6 0.199 -0.1872 INO80B -0.1239 0.1166 KLK10 0.0899 -0.0846

SBSN 0.1927 -0.1814 CLIC3 0.1235 -0.1163 HIST1 H2AA 0.0888 -0.0835

PRAME -0.1927 0.1814 MAD2L1 -0.1229 0.1157 ZNF614 -0.0884 0.0832

Table 3 (continued)

GENE ID XO.score X1.score GENE ID XO.score X1.score GENE ID XO.score X1.score

H19 0.0872 -0.0821 NOD2 0.0705 -0.0663 SERPINB3 0.0521 -0.049

GPNMB -0.0865 0.0815 SORBS1 0.0703 -0.0662 GRM3 0.0515 -0.0485

E2F7 -0.0861 0.081 KRTAP8-1 0.0703 -0.0662 KRTDAP 0.0504 -0.0475

SERPINB4 0.085 -0.08 E2F3 -0.0688 0.0647 PRSS27 0.0484 -0.0456

SLC4A1 1 0.0844 -0.0794 C2orf78 0.068 -0.064 VPS37D -0.048 0.0452

HES5 0.0833 -0.0784 ZNF721 -0.0672 0.0632 MYL3 0.0475 -0.0447

ALOX12B 0.0828 -0.0779 SKAP2 -0.0641 0.0604 EBF1 0.0475 -0.0447

VAX2 -0.0821 0.0773 IRF7 0.0628 -0.0591 SP6 0.0474 -0.0446

FAM3D 0.0814 -0.0766 MEOX1 0.0627 -0.059 PBOV1 0.0465 -0.0438

OR52B6 0.0813 -0.0765 TEX14 0.0624 -0.0587 ID1 0.0462 -0.0435

OR2B1 1 0.0805 -0.0757 G6PC 0.0615 -0.0579 ANLN -0.0453 0.0426

KLK8 0.0802 -0.0755 CCDC59 -0.061 0.0574 MIPOL1 -0.0452 0.0425

NKAIN2 -0.08 0.0753 SCML2 -0.0598 0.0563 C19orf33 0.0448 -0.0422

NOS1 0.0788 -0.0742 MAGEA9B -0.0589 0.0554 SCEL 0.0434 -0.0409

PEG10 -0.0788 0.0741 ZSCAN4 0.0581 -0.0547 ZNF284 0.0427 -0.0401

HOXC8 -0.0785 0.0739 CRABP2 0.0575 -0.0541 MIR554 0.0406 -0.0382

ABHD8 -0.0771 0.0726 TRIM16 0.0574 -0.054 SH3TC1 0.0399 -0.0375

CLDN5 0.0761 -0.0716 TFAP4 -0.0572 0.0539 STC2 -0.0397 0.0374

MCM10 -0.0754 0.0709 C19orf54 -0.0555 0.0522 DEFB122 0.0394 -0.0371

PI3 0.0753 -0.0709 HIST1 H4F 0.0545 -0.0513 OR14J1 0.0376 -0.0354

ADCY4 0.0742 -0.0699 IFNA21 0.0539 -0.0507 ZNF777 -0.0375 0.0353

HOXC9 -0.074 0.0697 CA1 0.0531 -0.05 NRSN1 0.0374 -0.0352

FANCL -0.0726 0.0683 ZYG1 1A 0.0525 -0.0494 FRAS1 -0.0363 0.0341

RDH13 0.0722 -0.068 CDA 0.0521 -0.0491 ECM1 0.0361 -0.034

Table 3 (continued)

GENE ID XO.score X1.score GENE ID XO.score X1.score GENE ID XO.score X1.score

ANGPT4 0.0359 -0.0338 CRIP1 0.0271 -0.0256 CXXC5 0.0143 -0.0134

GABRP 0.0354 -0.0333 ALDH7A1 0.0271 -0.0255 NGEF 0.01 14 -0.0108

OIP5 -0.0343 0.0323 PALMD 0.0265 -0.025 SPACA5B 0.01 1 -0.0104

DNAJC5G 0.034 -0.032 S100A7 0.0263 -0.0248 DUXAP3 0.0105 -0.0099

HOXD10 -0.034 0.032 REG1 B 0.0259 -0.0243 UBE2W 0.01 0.0094

B3GALT4 0.034 -0.032 MIDI -0.0254 0.0239 EHHADH -0.0088 0.0083

IGF2BP3 -0.0323 0.0304 CXCL10 -0.0253 0.0238 SPRR2E 0.0061 -0.0057

UBL4B 0.0315 -0.0297 HSD17B3 0.0251 -0.0236 ZNF121 -0.0059 0.0056

IL4R 0.0313 -0.0295 HCN1 0.0227 -0.0214 ECSCR 0.0053 -0.005

ACTL6A -0.0313 0.0295 PRSS22 0.021 -0.0197 FOXI2 0.0052 -0.0049

PCDHB13 -0.0313 0.0294 SPP1 -0.0209 0.0197 FGD5 0.0041 -0.0038

NKX2-1 0.031 -0.0291 MUC16 0.0209 -0.0196 LY6G6C 0.0034 -0.0032

SSX2 0.0309 -0.0291 HOXC10 -0.0206 0.0193 INTS10 0.0033 -0.0031

HIST2H3A -0.0307 0.0289 ZNF124 -0.0197 0.0185 KRT16 0.003 -0.0028

MND1 -0.0307 0.0289 PTPRB 0.0188 -0.0177 RNF126P1 0.003 -0.0028

SCN8A -0.0306 0.0288 TMEM41A -0.0182 0.0171 PFN2 -0.0016 0.0015

OR5B12 0.0304 -0.0286 INPP4B 0.0182 -0.0171 PLD2 0.0009 -0.0009

CLDN16 -0.0299 0.0281 FBX016 0.0172 -0.0162 KCNS1 0.0007 -0.0006

SLC17A4 0.0296 -0.0279 TMEM45B 0.0164 -0.0154 KCNJ3 0.0002 0.0002

RRM2B -0.0292 0.0274 MED28 -0.0159 0.0149

GATA4 -0.0289 0.0272 SORCS2 0.0158 -0.0148

TGM5 0.0279 -0.0262 GSTM5 0.0157 -0.0148

LOR 0.0277 -0.0261 OR51 Q1 0.0153 -0.0144

KRT78 0.0275 -0.0259 OR4E2 0.0146 -0.0137

Table 4 - Example DEG signature comprising 155 genes (gene weights are given in columns XO.score and Xl .score)

GENE ID XO.score X1.score GENE ID XO.score X1.score GENE ID XO.score Xl .score

KLK7 0.3747 -0.3527 NEK2 -0.1581 0.1488 DNAJC19 -0.0915 0.0861

SPNS2 0.3416 -0.3215 GHR -0.15 0.141 1 SNORD2 -0.082 0.0772

CPVL -0.2941 0.2768 LHX9 0.149 -0.1402 CBS -0.0799 0.0752

FER1 L6 0.2783 -0.2619 ABCG4 0.1456 -0.137 HTR2A 0.0793 -0.0746

ATP12A 0.2727 -0.2566 GABRB1 0.1395 -0.1313 CD164L2 0.0768 -0.0723

KLK12 0.2694 -0.2536 ZIC2 -0.1361 0.1281 SLC29A4 -0.0767 0.0722

ADCYAP1 0.2504 -0.2356 KRTAP13-1 0.1278 -0.1203 RPL7 -0.075 0.0706

RASIP1 0.2492 -0.2346 LCE6A 0.1 182 -0.1 1 12 ZFP37 -0.0748 0.0704

KLK6 0.2473 -0.2328 MIOX 0.1 129 -0.1063 SLC6A1 1 0.0748 -0.0704

KLK5 0.2292 -0.2157 CHST6 0.1 1 19 -0.1053 PLAT 0.0743 -0.0699

NCCRP1 0.2231 -0.2099 RBAK -0.1 1 17 0.1052 PRSS3 0.0729 -0.0686

RHCG 0.2222 -0.2092 CSN3 0.1069 -0.1006 ELAVL1 -0.0725 0.0682

IGFL1 0.215 -0.2023 ASCL2 -0.1016 0.0956 SOX17 0.0723 -0.068

GDPD3 0.1941 -0.1827 MPP6 -0.101 0.0951 SLC13A5 0.072 -0.0678

MLPH 0.1926 -0.1813 LYPD3 0.1007 -0.0948 SLC6A14 0.0712 -0.067

OR5B3 0.1909 -0.1797 ZNF804B 0.1002 -0.0944 SMYD3 -0.0662 0.0623

MT1 F -0.1869 0.1759 RSRC1 -0.1 0.0941 ITPKA -0.0654 0.0615

OR52K1 0.1862 -0.1752 PANX3 0.0999 -0.0941 STAB2 0.0637 -0.06

KRT23 0.1852 -0.1743 SPRR2B 0.0977 -0.0919 IFF02 0.0622 -0.0586

DHRS9 0.1829 -0.1722 TMPRSS1 1 D 0.0969 -0.0912 MYNN -0.0608 0.0572

SULT2B1 0.1786 -0.1681 UNC93A 0.095 -0.0894 PDE3B -0.0602 0.0567

CST6 0.1677 -0.1579 INO80B -0.0927 0.0872 KLK10 0.0587 -0.0552

SBSN 0.1615 -0.152 CLIC3 0.0923 -0.0869 HIST1 H2AA 0.0575 -0.0541

PRAME -0.1615 0.152 MAD2L1 -0.0917 0.0863 ZNF614 -0.0571 0.0538

Table 4 (continued)

GENE ID XO.score X1.score GENE ID XO.score X1.score GENE ID XO.score X1.score

H19 0.056 -0.0527 NOD2 0.0392 -0.0369 SERPINB3 0.0208 -0.0196

GPNMB -0.0553 0.0521 SORBS1 0.0391 -0.0368 GRM3 0.0203 -0.0191

E2F7 -0.0549 0.0517 KRTAP8-1 0.0391 -0.0368 KRTDAP 0.0192 -0.0181

SERPINB4 0.0538 -0.0506 E2F3 -0.0376 0.0353 PRSS27 0.0172 -0.0162

SLC4A1 1 0.0532 -0.05 C2orf78 0.0368 -0.0346 VPS37D -0.0168 0.0158

HES5 0.0521 -0.049 ZNF721 -0.0359 0.0338 MYL3 0.0163 -0.0153

ALOX12B 0.0516 -0.0486 SKAP2 -0.0329 0.031 EBF1 0.0163 -0.0153

VAX2 -0.0509 0.0479 IRF7 0.0316 -0.0297 SP6 0.0162 -0.0152

FAM3D 0.0502 -0.0472 MEOX1 0.0315 -0.0296 PBOV1 0.0153 -0.0144

OR52B6 0.0501 -0.0471 TEX14 0.0312 -0.0293 ID1 0.015 -0.0141

OR2B1 1 0.0493 -0.0464 G6PC 0.0303 -0.0285 ANLN -0.014 0.0132

KLK8 0.049 -0.0461 CCDC59 -0.0297 0.028 MIPOL1 -0.014 0.0131

NKAIN2 -0.0488 0.0459 SCML2 -0.0286 0.0269 C19orf33 0.0136 -0.0128

NOS1 0.0476 -0.0448 MAGEA9B -0.0277 0.026 SCEL 0.0122 -0.01 15

PEG10 -0.0476 0.0448 ZSCAN4 0.0269 -0.0253 ZNF284 0.01 14 -0.0108

HOXC8 -0.0473 0.0445 CRABP2 0.0262 -0.0247 MIR554 0.0093 -0.0088

ABHD8 -0.0459 0.0432 TRIM16 0.0261 -0.0246 SH3TC1 0.0087 -0.0082

CLDN5 0.0449 -0.0423 TFAP4 -0.026 0.0245 STC2 -0.0085 0.008

MCM10 -0.0441 0.0415 C19orf54 -0.0243 0.0228 DEFB122 0.0082 -0.0077

PI3 0.0441 -0.0415 HIST1 H4F 0.0233 -0.0219 OR14J1 0.0063 -0.006

ADCY4 0.043 -0.0405 IFNA21 0.0227 -0.0213 ZNF777 -0.0063 0.0059

HOXC9 -0.0428 0.0403 CA1 0.0219 -0.0206 NRSN1 0.0062 -0.0058

FANCL -0.0414 0.039 ZYG1 1A 0.0213 0.0201 FRAS1 -0.0051 0.0048

RDH13 0.041 -0.0386 CDA 0.0209 -0.0197 ECM1 0.0049 -0.0046

Table 4 (continued)

GENE ID XO.score X1.score

ANGPT4 0.0047 -0.0044

GABRP 0.0042 -0.004

OIP5 -0.0031 0.0029

DNAJC5G 0.0028 -0.0026

HOXD10 -0.0028 0.0026

B3GALT4 0.0028 -0.0026

IGF2BP3 0.001 0.001

UBL4B 0.0003 -0.0003

IL4R 0.0001 0.0001

ACTL6A 0.0001 0.0001

PCDHB13 0.0001 0.0001

Table 5 - Example DEG signature comprising 105 genes (gene weights are given in columns XO.score and XL score)

GENE ID XO.score X1.score GENE ID XO.score X1.score GENE ID XO.score X1.score

KLK7 0.3435 -0.3233 NEK2 -0.1269 0.1 194 DNAJC19 -0.0603 0.0567

SPNS2 0.3104 -0.2921 GHR -0.1 187 0.1 1 18 SNORD2 -0.0508 0.0478

CPVL -0.2629 0.2475 LHX9 0.1 177 -0.1 108 CBS -0.0487 0.0458

FER1 L6 0.247 -0.2325 ABCG4 0.1 144 -0.1077 HTR2A 0.0481 -0.0452

ATP12A 0.2414 -0.2272 GABRB1 0.1082 -0.1019 CD164L2 0.0456 -0.0429

KLK12 0.2382 -0.2242 ZIC2 -0.1049 0.0987 SLC29A4 -0.0455 0.0428

ADCYAP1 0.2191 -0.2062 KRTAP13-1 0.0966 -0.0909 RPL7 -0.0438 0.0412

RASIP1 0.218 -0.2052 LCE6A 0.087 -0.0818 ZFP37 -0.0436 0.041

KLK6 0.2161 -0.2034 MIOX 0.0817 -0.0769 SLC6A1 1 0.0436 -0.041

KLK5 0.198 -0.1863 CHST6 0.0807 -0.0759 PLAT 0.0431 -0.0405

NCCRP1 0.1918 -0.1806 RBAK -0.0805 0.0758 PRSS3 0.0417 -0.0392

RHCG 0.191 -0.1798 CSN3 0.0757 -0.0713 ELAVL1 -0.0413 0.0389

IGFL1 0.1837 -0.1729 ASCL2 -0.0703 0.0662 SOX17 0.041 -0.0386

GDPD3 0.1628 -0.1533 MPP6 -0.0698 0.0657 SLC13A5 0.0408 -0.0384

MLPH 0.1614 -0.1519 LYPD3 0.0695 -0.0654 SLC6A14 0.0399 -0.0376

OR5B3 0.1597 -0.1503 ZNF804B 0.069 -0.065 SMYD3 -0.035 0.0329

MT1 F -0.1556 0.1465 RSRC1 -0.0688 0.0647 ITPKA -0.0341 0.0321

OR52K1 0.155 -0.1458 PANX3 0.0687 -0.0647 STAB2 0.0325 -0.0306

KRT23 0.154 -0.1449 SPRR2B 0.0664 -0.0625 IFF02 0.031 -0.0292

DHRS9 0.1517 -0.1428 TMPRSS1 1 D 0.0657 -0.0618 MYNN -0.0295 0.0278

SULT2B1 0.1474 -0.1387 UNC93A 0.0637 -0.06 PDE3B -0.029 0.0273

CST6 0.1365 -0.1285 INO80B -0.0615 0.0579 KLK10 0.0274 -0.0258

SBSN 0.1303 -0.1226 CLIC3 0.061 1 -0.0575 HIST1 H2AA 0.0263 -0.0248

PRAME -0.1303 0.1226 MAD2L1 -0.0605 0.0569 ZNF614 -0.0259 0.0244

Table 5 (continued)

GENE ID XO.score X1.score GENE ID XO.score X1.score

H19 0.0248 -0.0233 NOD2 0.008 -0.0076

GPNMB -0.0241 0.0227 SORBS1 0.0079 -0.0074

E2F7 -0.0237 0.0223 KRTAP8-1 0.0079 -0.0074

SERPINB4 0.0226 0.0212 E2F3 -0.0063 0.006

SLC4A1 1 0.0219 -0.0206 C2orf78 0.0056 -0.0052

HES5 0.0208 -0.0196 ZNF721 -0.0047 0.0044

ALOX12B 0.0204 -0.0192 SKAP2 -0.0017 0.0016

VAX2 -0.0196 0.0185 IRF7 0.0004 -0.0003

FAM3D 0.0189 -0.0178 MEOX1 0.0003 0.0002

OR52B6 0.0188 -0.0177

OR2B1 1 0.018 -0.017

KLK8 0.0178 -0.0167

NKAIN2 -0.0176 0.0165

NOS1 0.0163 -0.0154

PEG10 -0.0163 0.0154

HOXC8 -0.0161 0.0151

ABHD8 -0.0147 0.0138

CLDN5 0.0137 -0.0129

MCM10 -0.0129 0.0121

PI3 0.0129 0.0121

ADCY4 0.01 18 0.01 1 1

HOXC9 -0.01 16 0.0109

FANCL 0.0102 0.0096

RDH13 0.0098 -0.0092

85

Table 6 - Example DEG signature comprising 66 genes (gene weights are given in columns XO.score and XL score)

GENE ID XO.score X1.score GENE ID XO.score X1.score GENE ID XO.score X1.score

KLK7 0.3123 -0.2939 NEK2 -0.0956 0.09 DNAJC19 -0.0291 0.0274

SPNS2 0.2792 -0.2627 GHR -0.0875 0.0824 SNORD2 -0.0196 0.0184

CPVL -0.2317 0.2181 LHX9 0.0865 -0.0814 CBS -0.0175 0.0164

FER1 L6 0.2158 -0.2031 ABCG4 0.0832 -0.0783 HTR2A 0.0168 -0.0158

ATP12A 0.2102 -0.1978 GABRB1 0.077 -0.0725 CD164L2 0.0143 -0.0135

KLK12 0.207 -0.1948 ZIC2 -0.0737 0.0693 SLC29A4 -0.0142 0.0134

ADCYAP1 0.1879 -0.1768 KRTAP13-1 0.0654 -0.0615 RPL7 -0.0126 0.01 18

RASIP1 0.1868 -0.1758 LCE6A 0.0557 -0.0525 ZFP37 -0.0124 0.01 16

KLK6 0.1849 -0.174 MIOX 0.0505 -0.0475 SLC6A1 1 0.0124 -0.01 16

KLK5 0.1668 -0.157 CHST6 0.0494 -0.0465 PLAT 0.01 18 0.0111

NCCRP1 0.1606 -0.1512 RBAK -0.0493 0.0464 PRSS3 0.0104 -0.0098

RHCG 0.1598 -0.1504 CSN3 0.0445 -0.0419 ELAVL1 0.0101 0.0095

IGFL1 0.1525 -0.1436 ASCL2 -0.0391 0.0368 SOX17 0.0098 -0.0092

GDPD3 0.1316 -0.1239 MPP6 -0.0386 0.0363 SLC13A5 0.0096 -0.009

MLPH 0.1302 -0.1225 LYPD3 0.0383 -0.036 SLC6A14 0.0087 -0.0082

OR5B3 0.1285 -0.1209 ZNF804B 0.0378 -0.0356 SMYD3 -0.0037 0.0035

MT1 F -0.1244 0.1 171 RSRC1 -0.0376 0.0354 ITPKA -0.0029 0.0027

OR52K1 0.1237 -0.1 165 PANX3 0.0375 -0.0353 STAB2 0.0013 0.0012

KRT23 0.1228 -0.1 156 SPRR2B 0.0352 -0.0331

DHRS9 0.1205 -0.1 134 TMPRSS1 1 D 0.0345 -0.0324

SULT2B1 0.1 161 -0.1093 UNC93A 0.0325 -0.0306

CST6 0.1053 -0.0991 INO80B -0.0302 0.0285

SBSN 0.099 -0.0932 CLIC3 0.0299 -0.0281

PRAME -0.099 0.0932 MAD2L1 -0.0292 0.0275

Table 7 - Example DEG signature comprising 45 genes (gene weights are given in columns XO.score and XL score)

GENE ID XO.score X1.score GENE ID XO.score X1.score

KLK7 0.281 -0.2645 NEK2 -0.0644 0.0606

SPNS2 0.2479 -0.2334 GHR -0.0563 0.053

CPVL -0.2005 0.1887 LHX9 0.0553 -0.052

FER1 L6 0.1846 -0.1737 ABCG4 0.0519 -0.0489

ATP12A 0.179 -0.1684 GABRB1 0.0458 -0.0431

KLK12 0.1758 -0.1654 ZIC2 -0.0424 0.04

ADCYAP1 0.1567 -0.1475 KRTAP13-1 0.0341 -0.0321

RASIP1 0.1556 -0.1464 LCE6A 0.0245 -0.0231

KLK6 0.1536 -0.1446 MIOX 0.0192 -0.0181

KLK5 0.1355 -0.1276 CHST6 0.0182 -0.0172

NCCRP1 0.1294 -0.1218 RBAK -0.0181 0.017

RHCG 0.1286 -0.121 CSN3 0.0133 -0.0125

IGFL1 0.1213 -0.1 142 ASCL2 -0.0079 0.0074

GDPD3 0.1004 -0.0945 MPP6 -0.0073 0.0069

MLPH 0.099 -0.0931 LYPD3 0.007 -0.0066

OR5B3 0.0973 -0.0915 ZNF804B 0.0066 -0.0062

MT1 F -0.0932 0.0877 RSRC1 -0.0063 0.006

OR52K1 0.0925 -0.0871 PANX3 0.0063 -0.0059

KRT23 0.0915 -0.0862 SPRR2B 0.004 -0.0038

DHRS9 0.0892 -0.084 TMPRSS1 1 D 0.0032 -0.003

SULT2B1 0.0849 -0.0799 UNC93A 0.0013 -0.0012

CST6 0.074 -0.0697

SBSN 0.0678 -0.0638

PRAME -0.0678 0.0638

Table 8 - Example DEG signature comprising 31 genes (gene weights are given in columns XO.score and XL score)

GENE ID XO.score X1.score GENE ID XO.score X1.score

KLK7 0.2498 -0.2351 NEK2 -0.0332 0.0312

SPNS2 0.2167 -0.204 GHR -0.0251 0.0236

CPVL -0.1692 0.1593 LHX9 0.024 -0.0226

FER1 L6 0.1534 -0.1443 ABCG4 0.0207 -0.0195

ATP12A 0.1477 -0.1391 GABRB1 0.0146 -0.0137

KLK12 0.1445 -0.136 ZIC2 -0.01 12 0.0106

ADCYAP1 0.1254 -0.1 181 KRTAP13-1 0.0029 -0.0027

RASIP1 0.1243 -0.1 17

KLK6 0.1224 -0.1 152

KLK5 0.1043 -0.0982

NCCRP1 0.0982 -0.0924

RHCG 0.0973 -0.0916

IGFL1 0.0901 -0.0848

GDPD3 0.0692 -0.0651

MLPH 0.0677 -0.0638

OR5B3 0.066 -0.0621

MT1 F -0.0619 0.0583

OR52K1 0.0613 -0.0577

KRT23 0.0603 -0.0568

DHRS9 0.058 -0.0546

SULT2B1 0.0537 -0.0505

CST6 0.0428 -0.0403

SBSN 0.0366 -0.0344

PRAME -0.0366 0.0344

Table 9 - Example DEG signature comprising 25 genes (gene weights are given in columns XO.score and Xl .score)

GENE ID XO.score X1.score GENE ID XO.score Xl .score

KLK7 0.2186 -0.2057 NEK2 -0.0019 0.0018

SPNS2 0.1855 -0.1746

CPVL -0.138 0.1299

FER1 L6 0.1221 -0.1 149

ATP12A 0.1 165 -0.1097

KLK12 0.1 133 -0.1066

ADCYAP1 0.0942 -0.0887

RASIP1 0.0931 -0.0876

KLK6 0.0912 -0.0858

KLK5 0.0731 -0.0688

NCCRP1 0.0669 -0.063

RHCG 0.0661 -0.0622

IGFL1 0.0588 -0.0554

GDPD3 0.0379 -0.0357

MLPH 0.0365 -0.0344

OR5B3 0.0348 -0.0328

MT1 F -0.0307 0.0289

OR52K1 0.0301 -0.0283

KRT23 0.0291 -0.0274

DHRS9 0.0268 -0.0252

SULT2B1 0.0225 0.021 1

CST6 0.01 16 -0.0109

SBSN 0.0054 -0.0051

PRAME -0.0054 0.005

89

Table 10 - Characteristic parameters and metrics of example DEG signatures presented in Tables 1-9

DEG Signature AUC Threshold Sensitivity Specificity Sensitivity Specificity

Genes AUC (TGCA)

Table (validation) value (validation) (validation) (TCGA) (TCGA)

1 397 1 0.868323 0.14 1 0.621514 1 0.578378378

2 291 1 0.812221 0.105 0.959184 0.651394 1 0.578378378

3 21 1 1 0.758964 0.04 0.693878 0.705179 1 0.578378378

4 155 1 0.698837 0.035 0.408163 0.756972 1 0.578378378

5 105 1 0.636393 0.3 0.55102 0.609562 1 0.578378378

6 66 1 0.576795 0.34 0.428571 0.611554 1 0.578378378

7 45 1 0.537523 0.355 0.326531 0.621514 1 0.578378378

8 31 1 0.520124 0.4124 0.326531 0.593625 1 0.578378378 9 25 1 0.510285 0.51 0.44898 0.505976 1 0.578378378

Table 11 - Example DMP signature comprising 141 DMPs (DMP weights are given in columns XO.score and XL score)

DMP ID XO.score X1.score DMP ID XO.score X1.score DMP ID XO.score X1.score cg07716946 -0.2803 0.3666 cg09968620 -0.0861 0.1 126 cg10364040 -0.053 0.0693 cg04786287 -0.2741 0.3584 eg 19664945 -0.0839 0.1097 cg22991 101 -0.0522 0.0683 cg03782157 -0.1916 0.2505 cg06685968 -0.0801 0.1048 cg03222323 -0.0516 0.0675 cg12368188 -0.1907 0.2494 cg1 1362010 -0.079 0.1033 cg21425842 -0.0502 0.0656 cg25829490 -0.171 1 0.2237 eg 10759602 -0.0754 0.0986 cg04233770 -0.0498 0.0651 cg09628195 -0.161 1 0.2107 cg26053832 -0.0752 0.0984 cg00503383 -0.0491 0.0642 cg20674701 -0.1581 0.2068 cg27071 152 -0.0749 0.098 cg06530490 -0.0466 0.0609 cg22549870 -0.1403 0.1835 eg 17975443 -0.0741 0.0969 cg2181 1 143 -0.0458 0.0598 cg02020945 -0.1384 0.181 cg03366986 -0.0731 0.0956 cg22674699 -0.0455 0.0595 cg14164044 -0.1363 0.1783 cg00332153 -0.071 1 0.093 cg00217080 0.0452 -0.0591 cg10210594 -0.1353 0.177 cg05317090 -0.0702 0.0918 cg16971668 -0.0447 0.0584 cg22974982 -0.1241 0.1622 cg00459623 -0.0692 0.0905 cg13406145 -0.0445 0.0582 eg 15545035 -0.1217 0.1591 cg27622679 -0.0691 0.0904 cg14290904 -0.0439 0.0574 cg03843000 -0.1 189 0.1555 cg04490714 -0.0691 0.0903 cg13294849 -0.0436 0.0571 eg 16332610 -0.1059 0.1385 cg20501518 -0.0689 0.09 cg06316886 -0.0435 0.0569 cg20627174 -0.1009 0.1319 cg04164058 -0.0661 0.0864 cg14765959 -0.0434 0.0567 cg07524679 -0.1008 0.1319 cg142391 1 1 0.0647 -0.0847 cg18235734 -0.0433 0.0567 cg09570682 -0.1002 0.131 1 cg25371634 -0.0643 0.0841 cg15281710 -0.0394 0.0516 cg18891712 -0.0994 0.13 cg01783662 -0.0636 0.0832 cg26666835 -0.0393 0.0514 cg03892356 -0.0991 0.1296 eg 15888290 -0.0595 0.0777 cg22669623 -0.039 0.051 1 cg26470798 -0.0984 0.1286 cg14631910 -0.0593 0.0775 cg12254845 -0.0384 0.0503 cg12144497 -0.097 0.1268 cg04689080 -0.0571 0.0746 cg14428048 -0.0382 0.05 cg04153740 -0.0916 0.1 198 cg1 1 123595 -0.054 0.0706 cg04018288 -0.0381 0.0499 cg04828133 -0.0891 0.1 165 cg22541735 -0.0539 0.0705 cg25023994 -0.0363 0.0475


tL O - O ^ O CO - CM CO O O - O iO OO ^ ^ O O OO O CD CO CM

- o o c¾ 5r 0 0) 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 O o o o o o o o o o o o o o o o o o o o o o o o o

H

Table 12 - Example DMP signature comprising 65 DMPs (DMP weights are given in columns XO.score and Xl .score)

DMP ID XO.score Xl .score DMP ID XO.score Xl .score DMP ID XO.score Xl .score cg07716946 -0.2403 0.3142 cg09968620 -0.0461 0.0603 cg10364040 -0.013 0.017 cg04786287 -0.234 0.306 eg 19664945 -0.0439 0.0573 cg22991 101 -0.0122 0.0159 cg03782157 -0.1515 0.1982 cg06685968 -0.0401 0.0524 cg03222323 -0.01 16 0.0151 cg12368188 -0.1507 0.197 cg1 1362010 -0.0389 0.0509 cg21425842 -0.0101 0.0133 cg25829490 -0.131 0.1713 eg 10759602 -0.0353 0.0462 cg04233770 -0.0097 0.0127 cg09628195 -0.121 1 0.1583 cg26053832 -0.0352 0.046 cg00503383 -0.0091 0.01 19 cg20674701 -0.1 181 0.1544 cg27071 152 -0.0349 0.0456 cg06530490 -0.0065 0.0085 cg22549870 -0.1003 0.131 1 eg 17975443 -0.034 0.0445 cg2181 1 143 -0.0057 0.0075 cg02020945 -0.0984 0.1287 cg03366986 -0.033 0.0432 cg22674699 -0.0054 0.0071 cg14164044 -0.0963 0.1259 cg00332153 -0.031 0.0406 cg00217080 0.0051 -0.0067 cg10210594 -0.0953 0.1246 cg05317090 -0.0302 0.0395 cg16971668 -0.0046 0.006 cg22974982 -0.084 0.1099 cg00459623 -0.0291 0.0381 cg13406145 -0.0045 0.0058 eg 15545035 -0.0816 0.1067 cg27622679 -0.0291 0.038 cg14290904 -0.0039 0.0051 cg03843000 -0.0788 0.1031 cg04490714 -0.029 0.0379 cg13294849 -0.0036 0.0047 eg 16332610 -0.0659 0.0862 cg20501518 -0.0288 0.0377 cg06316886 -0.0035 0.0046 cg20627174 -0.0608 0.0795 cg04164058 -0.026 0.034 cg14765959 -0.0033 0.0043 cg07524679 -0.0608 0.0795 cg142391 1 1 0.0247 -0.0323 cg18235734 -0.0033 0.0043 cg09570682 -0.0602 0.0787 cg25371634 -0.0243 0.0317

cg18891712 -0.0594 0.0777 cg01783662 -0.0236 0.0308

cg03892356 -0.059 0.0772 eg 15888290 -0.0194 0.0254

cg26470798 -0.0583 0.0763 cg14631910 -0.0192 0.0251

cg12144497 -0.0569 0.0744 cg04689080 -0.017 0.0222

cg04153740 -0.0516 0.0674 cg1 1 123595 -0.014 0.0183

cg04828133 -0.049 0.0641 cg22541735 -0.0139 0.0182

Table 13 - Example DMP signature comprising 27 DMPs (DMP weights are given in columns XO.score and Xl .score)

DMP ID XO.score X1.score DMP ID XO.score Xl .score

cg07716946 -0.2002 0.2619 cg09968620 -0.006 0.0079

cg04786287 -0.194 0.2536 cg19664945 -0.0038 0.005

cg03782157 -0.1115 0.1458 cg06685968 0 0

cg12368188 -0.1106 0.1447

cg25829490 -0.091 0.119

cg09628195 -0.081 0.106

cg20674701 -0.078 0.102

cg22549870 -0.0602 0.0788

cg02020945 -0.0583 0.0763

cg14164044 -0.0562 0.0736

cg10210594 -0.0552 0.0722

cg22974982 -0.044 0.0575

eg 15545035 -0.0416 0.0543

cg03843000 -0.0388 0.0507

eg 16332610 -0.0259 0.0338

cg20627174 -0.0208 0.0272

cg07524679 -0.0207 0.0271

cg09570682 -0.0201 0.0263

cg18891712 -0.0193 0.0253

cg03892356 -0.019 0.0248

cg26470798 -0.0183 0.0239

cg12144497 -0.0169 0.0221

cg04153740 -0.0115 0.0151

cg04828133 -0.009 0.0118

Table 14 - Example DMP signature comprising 13 DMPs (DMP weights are given in columns X0. score and Xl.score)

DMP ID XO.score Xl .score

cg07716946 -0.1602 0.2095

cg04786287 -0.1539 0.2013

cg03782157 -0.0714 0.0934

cg12368188 -0.0706 0.0923

cg25829490 -0.0509 0.0666

cg09628195 -0.041 0.0536

cg20674701 -0.038 0.0497

cg22549870 -0.0202 0.0264

cg02020945 -0.0183 0.0239

cg14164044 -0.0162 0.0212

cg10210594 -0.0152 0.0199

cg22974982 -0.0039 0.0051

eg 15545035 -0.0015 0.002

cg04828133 -0.009 0.0118

Table 15 - Example DMP signature comprising 6 DMPs (DMP weights are given in columns X0. score and XI. score)

DMP ID XO.score Xl .score

cg07716946 -0.1201 0.1571

cg04786287 -0.1139 0.1489

cg03782157 -0.0314 0.041

cg12368188 -0.0305 0.0399

cg25829490 -0.0109 0.0142

cg09628195 -0.0009 0.0012

Table 16 - Example DMP signature comprising 2 DMPs (DMP weights are given in columns XO.score and XL score)

DMP ID XO.score X1.score

cg07716946 -0.0801 0.1047

cg04786287 -0.0738 0.0965

Table 17 - Example DMP signature comprising 2 DMPs (DMP weights are given in columns XO.score and Xl .score)

DMP ID XO.score X1.score

cg07716946 -0.04 0.0524

cg04786287 -0.0338 0.0442

Table 18 - Summary of characteristic parameters and metrics of example DMP signatures presented in Tables 11-17

DMP

AUC AUV Threshold Sensitivity Specificity Sensitivity Specificity

Signature DMPs

(Validation) (TCGA) Value (Validation) (Validation) (TCGA) (TCGA)

Table

1 1 141 0.994118 0.998263 0.5 0.882353 1 1 0.583784

12 65 0.994118 0.998649 0.475 0.882353 1 1 0.581081

13 27 0.976471 0.995817 0.455 0.882353 1 1 0.578378

14 13 0.976471 0.987259 0.45 0.882353 1 1 0.564865

15 6 0.982353 0.956113 0.45 0.882353 1 1 0.545946

16 2 0.982353 0.939833 0.44 0.882353 1 1 0.559459 17 2 0.982353 0.940412 0.43 0.882353 1 1 0.613514

96

Table 19 - Example CNV signature comprising 219 CNV bands (CNV band weights are given in columns XO.score and XL score)

CNV band XO.score X1.score CNV band XO.score X1.score CNV band XO.score X1.score

3q26.32 -0.4346 0.2173 5q11.2 0.2415 -0.1208 5p15.33 -0.1984 0.0992

3q26.31 -0.4068 0.2034 3p24.1 0.2412 -0.1206 3p14.1 0.1955 -0.0978

3q26.2 -0.3878 0.1939 3p24.2 0.2373 -0.1187 19q13.12 -0.1922 0.0961

3q26.33 -0.3864 0.1932 3p23 0.2317 -0.1158 5q21.3 0.191 -0.0955

3q28 -0.3754 0.1877 3p26.3 0.2313 -0.1157 5q23.1 0.1904 -0.0952

5q13.2 0.3718 -0.1859 5p15.1 -0.2313 0.1156 5q14.1 0.1867 -0.0934

3q26.1 -0.3491 0.1745 3p22.1 0.2288 -0.1144 3q25.33 -0.184 0.092

5q12.3 0.3344 -0.1672 9p23 0.2246 -0.1123 3q25.32 -0.1839 0.092

3q27.3 -0.3254 0.1627 3p26.1 0.2238 -0.1119 3p25.1 0.1793 -0.0896

3q29 -0.3223 0.1612 19q13.13 -0.2172 0.1086 9p24.2 0.1792 -0.0896

3q27.2 -0.3162 0.1581 3p24.3 0.2163 -0.1081 5q33.3 0.1787 -0.0894

3q27.1 -0.3105 0.1552 3p22.2 0.2151 -0.1075 2p25.3 -0.1783 0.0891

5q 12.1 0.3078 -0.1539 2q34 0.2123 -0.1062 3q25.1 -0.1778 0.0889

5q 13.1 0.2884 -0.1442 3p21.33 0.2109 -0.1054 3q25.31 -0.1773 0.0886

5q22.1 0.2878 -0.1439 9p22.1 0.2085 -0.1042 19q 11 -0.1757 0.0879

21 q21 .2 0.2744 -0.1372 3p21.32 0.2077 -0.1038 3p14.2 0.1748 -0.0874

3p22.3 0.2592 -0.1296 9p24.1 0.207 -0.1035 3p21.1 0.1746 -0.0873

9p22.2 0.2584 -0.1292 3p14.3 0.2069 -0.1035 5p15.32 -0.1715 0.0857

5q13.3 0.2578 -0.1289 5q23.2 0.2068 -0.1034 3p21.31 0.1706 -0.0853

3p26.2 0.2566 -0.1283 5q15 0.2064 -0.1032 5p12 -0.17 0.085

5q14.3 0.2547 -0.1273 5q21.1 0.2063 -0.1031 5q33.1 0.1699 -0.0849

5q22.2 0.2538 -0.1269 5p14.1 -0.2036 0.1018 4q32.2 0.1698 -0.0849

5q22.3 0.2498 -0.1249 5p15.2 -0.2019 0.101 5p13.3 -0.1658 0.0829

9p22.3 0.2424 0.1212 3p13 0.2009 -0.1005 5p13.1 -0.1657 0.0829

Table 19 (continued)

CNV band XO.score X1.score CNV band XO.score X1.score CNV band XO.score X1.score

19q13.2 -0.165 0.0825 19q12 -0.1319 0.0659 6q11.1 -0.0949 0.0474

9p21.3 0.1632 -0.0816 4p15.31 0.1315 -0.0658 20p11.22 -0.0945 0.0473

4p15.2 0.1627 -0.0814 4q31.3 0.1274 -0.0637 2q33.3 0.0933 -0.0467

3p21.2 0.1626 -0.0813 22q11.1 -0.1265 0.0633 11 q13.3 -0.0911 0.0456

3p25.2 0.1618 -0.0809 8p22 0.1228 -0.0614 4q26 0.0879 -0.0439

5q14.2 0.161 -0.0805 8p21.2 0.1197 -0.0599 9p24.3 0.0853 -0.0426

3p25.3 0.1586 -0.0793 4q23 0.1163 -0.0582 2q36.2 0.0843 -0.0421

5q33.2 0.1573 -0.0787 5q31.1 0.1158 -0.0579 6p25.3 -0.0831 0.0416

4q31.23 0.1545 -0.0773 4p14 0.1128 -0.0564 4q31.21 0.0829 -0.0415

5p13.2 -0.1542 0.0771 8p21.1 0.1117 -0.0558 21 q21.3 0.08 -0.04

5p15.31 -0.1502 0.0751 4q34.2 0.1114 -0.0557 15q25.3 -0.0795 0.0397

4q32.1 0.1461 -0.0731 8p23.2 0.1 1 1 -0.0555 6p11.2 -0.0766 0.0383

4p15.1 0.1456 -0.0728 9p21.2 0.1104 -0.0552 15q22.32 -0.0764 0.0382

4q32.3 0.1442 -0.0721 3p12.2 0.1097 -0.0548 4q27 0.0757 -0.0378

3q24 -0.1419 0.0709 5q34 0.1085 -0.0542 10q21.1 0.0752 -0.0376

4q28.2 0.1412 -0.0706 4q22.3 0.1055 -0.0528 1 p13.3 0.075 -0.0375

8p12 0.1407 -0.0703 4q31.22 0.1035 -0.0517 5p14.3 -0.0715 0.0358

5q32 0.1375 -0.0687 5q31.2 0.1019 -0.051 5q31.3 0.0679 -0.034

15q26.3 -0.1366 0.0683 4q24 0.1003 -0.0502 15q26.1 -0.0649 0.0324

5q23.3 0.1347 -0.0673 4q28.1 0.1001 -0.05 2q36.3 0.0639 -0.032

3q25.2 -0.1346 0.0673 15q26.2 -0.0996 0.0498 8q24.3 -0.0638 0.0319

5q21.2 0.1345 -0.0673 17p11.1 -0.0995 0.0498 2p16.3 -0.0634 0.0317

4q33 0.1327 -0.0663 3p12.1 0.0976 -0.0488 2q36.1 0.0633 -0.0316

19q13.11 -0.1325 0.0663 1 p13.2 0.0973 -0.0487 8p21.3 0.0627 -0.0314

Table 19 (continued)

CNV band XO.score X1.score CNV band XO.score X1.score CNV band XO.score X1.score

2q35 0.0615 -0.0308 8q24.22 -0.0318 0.0159 2p16.2 -0.0154 0.0077 4q31.1 0.061 -0.0305 1 q22 -0.0309 0.0155 4p13 0.0149 -0.0075 4q25 0.058 -0.029 4p15.32 0.0286 -0.0143 1 p21.3 0.0133 -0.0066 1q42.13 -0.0563 0.0281 22q12.1 -0.0277 0.0139 2q32.3 -0.0132 0.0066 8q24.21 -0.0555 0.0277 13q21.2 0.0275 -0.0137 1 p32.2 0.0117 -0.0059 13q12.12 0.0546 -0.0273 1q41 -0.0273 0.0137 3q22.3 -0.0106 0.0053 4q34.1 0.054 -0.027 1q21.3 -0.027 0.0135 2p24.1 -0.0105 0.0052 5q11.1 0.0539 -0.0269 2q14.2 -0.0258 0.0129 22q11.22 -0.0099 0.005

20p11.1 -0.0527 0.0263 1q21.2 -0.0249 0.0124 15q24.3 -0.0085 0.0043

2q31.1 -0.0505 0.0253 1 p21.1 0.0247 -0.0123 1q42.2 -0.0076 0.0038 2q33.2 0.0499 -0.025 4q34.3 0.0243 0.0122 2q12.2 -0.0072 0.0036 2p25.1 -0.048 0.024 22q11.21 -0.0232 0.0116 2p11.2 -0.0071 0.0036

2p22.1 -0.0472 0.0236 3p11.2 0.023 -0.0115 1q42.3 -0.007 0.0035

2p14 -0.0465 0.0232 4q22.2 0.0227 -0.0113 2p16.1 -0.0069 0.0035

2p22.2 -0.046 0.023 8p23.3 0.0225 -0.0113 1 p22.1 0.0068 -0.0034

3p12.3 0.0453 -0.0227 8p23.1 0.0212 -0.0106 4q28.3 0.0062 -0.0031 1 p32.1 0.0432 -0.0216 1q32.3 0.02 0.01 1 p36.11 0.0055 -0.0028 13q12.11 0.0427 -0.0213 1 p31.1 0.0199 -0.0099 1q32.2 -0.0037 0.0018 2p25.2 -0.0414 0.0207 1q44 -0.0187 0.0093 3q23 -0.0035 0.0017 2p15 -0.0394 0.0197 2p21 -0.0176 0.0088 13q33.1 0.003 -0.0015 1 p22.3 0.0376 -0.0188 2p23.2 -0.0176 0.0088 2q11.2 -0.0029 0.0014 21 q22.11 0.0362 -0.0181 2p24.3 -0.0172 0.0086 2q12.3 -0.0026 0.0013 3p11.1 0.0337 -0.0168 2q31.3 -0.0165 0.0083 15q25.2 -0.0023 0.001 1 18q22.1 0.0333 -0.0166 6p22.3 -0.0159 0.0079 10q 11.23 0.0021 0.001 1

100

S>

o O O O

o o o o

o o o

CO o o o

X


Table 20 - Example CNV signature comprising 179 CNV bands (CNV band weights are given in columns XO.score and Xl.score)

CNV band XO.score X1.score CNV band XO.score X1.score CNV band XO.score X1.score

3q26.32 -0.4104 0.2052 5q11.2 0.2174 -0.1087 5p15.33 -0.1743 0.0871

3q26.31 -0.3827 0.1913 3p24.1 0.2171 -0.1085 3p14.1 0.1714 -0.0857

3q26.2 -0.3637 0.1818 3p24.2 0.2132 -0.1066 19q13.12 -0.168 0.084

3q26.33 -0.3623 0.1811 3p23 0.2075 -0.1038 5q21.3 0.1668 -0.0834

3q28 -0.3513 0.1756 3p26.3 0.2072 -0.1036 5q23.1 0.1662 -0.0831

5q13.2 0.3477 -0.1739 5p15.1 -0.2071 0.1036 5q14.1 0.1626 -0.0813

3q26.1 -0.325 0.1625 3p22.1 0.2046 -0.1023 3q25.33 -0.1598 0.0799

5q12.3 0.3102 -0.1551 9p23 0.2005 0.1002 3q25.32 -0.1598 0.0799

3q27.3 -0.3012 0.1506 3p26.1 0.1997 -0.0998 3p25.1 0.1551 -0.0776

3q29 -0.2982 0.1491 19q13.13 -0.1931 0.0965 9p24.2 0.1551 -0.0776

3q27.2 -0.292 0.146 3p24.3 0.1921 -0.0961 5q33.3 0.1546 -0.0773

3q27.1 -0.2863 0.1432 3p22.2 0.1909 -0.0955 2p25.3 -0.1541 0.0771

5q 12.1 0.2836 -0.1418 2q34 0.1882 -0.0941 3q25.1 -0.1536 0.0768

5q 13.1 0.2643 -0.1321 3p21.33 0.1867 -0.0934 3q25.31 -0.1531 0.0766

5q22.1 0.2637 -0.1318 9p22.1 0.1843 -0.0922 19q 11 -0.1516 0.0758

21 q21 .2 0.2502 -0.1251 3p21.32 0.1835 -0.0918 3p14.2 0.1507 -0.0753

3p22.3 0.2351 -0.1175 9p24.1 0.1829 -0.0915 3p21.1 0.1505 -0.0753

9p22.2 0.2343 -0.1171 3p14.3 0.1828 -0.0914 5p15.32 -0.1473 0.0737

5q13.3 0.2336 -0.1168 5q23.2 0.1826 -0.0913 3p21.31 0.1465 -0.0732

3p26.2 0.2325 -0.1162 5q15 0.1823 -0.0911 5p12 -0.1459 0.0729

5q14.3 0.2305 -0.1153 5q21.1 0.1821 -0.0911 5q33.1 0.1457 -0.0729

5q22.2 0.2296 -0.1148 5p14.1 -0.1795 0.0897 4q32.2 0.1457 -0.0728

5q22.3 0.2257 -0.1128 5p15.2 -0.1778 0.0889 5p13.3 -0.1417 0.0708

9p22.3 0.2182 -0.1091 3p13 0.1768 -0.0884 5p13.1 -0.1416 0.0708

Table 20 (continued)

CNV band XO.score X1.score CNV band XO.score X1.score CNV band XO.score X1.score

19q13.2 -0.1409 0.0704 19q12 -0.1077 0.0539 6q11.1 -0.0708 0.0354

9p21.3 0.139 -0.0695 4p15.31 0.1074 -0.0537 20p11.22 -0.0704 0.0352

4p15.2 0.1386 -0.0693 4q31.3 0.1033 -0.0516 2q33.3 0.0692 -0.0346

3p21.2 0.1384 -0.0692 22q11.1 -0.1024 0.0512 11 q13.3 -0.067 0.0335

3p25.2 0.1376 -0.0688 8p22 0.0986 -0.0493 4q26 0.0637 -0.0319

5q14.2 0.1369 -0.0684 8p21.2 0.0956 -0.0478 9p24.3 0.0611 -0.0306

3p25.3 0.1344 -0.0672 4q23 0.0922 -0.0461 2q36.2 0.0601 -0.0301

5q33.2 0.1332 -0.0666 5q31.1 0.0917 -0.0458 6p25.3 -0.059 0.0295

4q31.23 0.1304 -0.0652 4p14 0.0886 -0.0443 4q31.21 0.0588 -0.0294

5p13.2 -0.13 0.065 8p21.1 0.0876 -0.0438 21 q21.3 0.0558 -0.0279

5p15.31 -0.1261 0.063 4q34.2 0.0872 -0.0436 15q25.3 -0.0553 0.0277

4q32.1 0.122 -0.061 8p23.2 0.0868 -0.0434 6p11.2 -0.0525 0.0263

4p15.1 0.1215 -0.0607 9p21.2 0.0862 -0.0431 15q22.32 -0.0523 0.0262

4q32.3 0.1201 -0.06 3p12.2 0.0855 -0.0428 4q27 0.0516 -0.0258

3q24 -0.1177 0.0589 5q34 0.0843 -0.0422 10q21.1 0.051 -0.0255

4q28.2 0.117 -0.0585 4q22.3 0.0814 -0.0407 1 p13.3 0.0508 -0.0254

8p12 0.1165 -0.0583 4q31.22 0.0793 -0.0397 5p14.3 -0.0474 0.0237

5q32 0.1133 -0.0567 5q31.2 0.0778 -0.0389 5q31.3 0.0438 -0.0219

15q26.3 -0.1124 0.0562 4q24 0.0762 -0.0381 15q26.1 -0.0407 0.0204

5q23.3 0.1105 -0.0553 4q28.1 0.0759 -0.038 2q36.3 0.0398 -0.0199

3q25.2 -0.1104 0.0552 15q26.2 -0.0755 0.0377 8q24.3 -0.0396 0.0198

5q21.2 0.1104 -0.0552 17p11.1 -0.0754 0.0377 2p16.3 -0.0392 0.0196

4q33 0.1085 -0.0543 3p12.1 0.0735 -0.0367 2q36.1 0.0391 -0.0196

19q13.11 -0.1084 0.0542 1 p13.2 0.0732 -0.0366 8p21.3 0.0386 -0.0193

Table 20 (continued)

CNV band XO.score X1.score CNV band XO.score X1.score

2q35 0.0374 -0.0187 8q24.22 -0.0076 0.0038 4q31.1 0.0369 -0.0184 1 q22 -0.0068 0.0034 4q25 0.0339 -0.0169 4p15.32 0.0044 0.0022 1q42.13 -0.0321 0.0161 22q12.1 -0.0036 0.0018 8q24.21 -0.0313 0.0157 13q21.2 0.0033 -0.0017 13q12.12 0.0305 -0.0152 1q41 -0.0032 0.0016 4q34.1 0.0298 -0.0149 1q21.3 -0.0029 0.0015 5q11.1 0.0297 -0.0149 2q14.2 -0.0017 0.0008

20p11.1 -0.0285 0.0143 1q21.2 -0.0007 0.0004

2q31.1 -0.0264 0.0132 1 p21.1 0.0006 -0.0003 2q33.2 0.0258 -0.0129 4q34.3 0.0002 0.0001 2p25.1 -0.0239 0.0119

2p22.1 -0.0231 0.0115

2p14 -0.0223 0.0112

2p22.2 -0.0218 0.0109

3p12.3 0.0212 -0.0106

1 p32.1 0.0191 -0.0095

13q12.11 0.0185 -0.0093

2p25.2 -0.0172 0.0086

2p15 -0.0153 0.0076

1 p22.3 0.0135 -0.0067

21 q22.11 0.0121 -0.006

3p11.1 0.0095 -0.0048

18q22.1 0.0092 -0.0046

Table 21 - Example CNV signature comprising 155 CNV bands (CNV band weights are given in columns XO.score and XL score)

CNV band XO.score X1.score CNV band XO.score X1.score CNV band XO.score X1.score

3q26.32 -0.3863 0.1931 5q11.2 0.1932 -0.0966 5p15.33 -0.1501 0.0751

3q26.31 -0.3585 0.1793 3p24.1 0.1929 -0.0965 3p14.1 0.1472 -0.0736

3q26.2 -0.3395 0.1698 3p24.2 0.1891 -0.0945 19q13.12 -0.1439 0.0719

3q26.33 -0.3381 0.1691 3p23 0.1834 -0.0917 5q21.3 0.1427 -0.0713

3q28 -0.3272 0.1636 3p26.3 0.183 -0.0915 5q23.1 0.1421 -0.0711

5q13.2 0.3236 -0.1618 5p15.1 -0.183 0.0915 5q14.1 0.1384 -0.0692

3q26.1 -0.3008 0.1504 3p22.1 0.1805 -0.0902 3q25.33 -0.1357 0.0678

5q12.3 0.2861 -0.1431 9p23 0.1763 -0.0882 3q25.32 -0.1357 0.0678

3q27.3 -0.2771 0.1385 3p26.1 0.1755 -0.0878 3p25.1 0.131 -0.0655

3q29 -0.274 0.137 19q13.13 -0.1689 0.0845 9p24.2 0.131 -0.0655

3q27.2 -0.2679 0.1339 3p24.3 0.168 -0.084 5q33.3 0.1304 -0.0652

3q27.1 -0.2622 0.1311 3p22.2 0.1668 -0.0834 2p25.3 -0.13 0.065

5q 12.1 0.2595 -0.1297 2q34 0.164 -0.082 3q25.1 -0.1295 0.0647

5q 13.1 0.2401 0.1201 3p21.33 0.1626 -0.0813 3q25.31 -0.129 0.0645

5q22.1 0.2395 -0.1198 9p22.1 0.1602 -0.0801 19q 11 -0.1274 0.0637

21 q21 .2 0.2261 -0.113 3p21.32 0.1594 -0.0797 3p14.2 0.1265 -0.0633

3p22.3 0.2109 -0.1055 9p24.1 0.1588 -0.0794 3p21.1 0.1264 -0.0632

9p22.2 0.2101 -0.1051 3p14.3 0.1586 -0.0793 5p15.32 -0.1232 0.0616

5q13.3 0.2095 -0.1047 5q23.2 0.1585 -0.0792 3p21.31 0.1223 -0.0612

3p26.2 0.2083 -0.1042 5q15 0.1581 -0.0791 5p12 -0.1217 0.0609

5q14.3 0.2064 -0.1032 5q21.1 0.158 -0.079 5q33.1 0.1216 -0.0608

5q22.2 0.2055 -0.1027 5p14.1 -0.1554 0.0777 4q32.2 0.1215 -0.0608

5q22.3 0.2015 -0.1008 5p15.2 -0.1537 0.0768 5p13.3 -0.1175 0.0588

9p22.3 0.1941 -0.0971 3p13 0.1527 -0.0763 5p13.1 -0.1174 0.0587

Table 21 (continued)

CNV band XO.score X1.score CNV band XO.score X1.score CNV band XO.score X1.score

19q13.2 -0.1167 0.0584 19q12 -0.0836 0.0418 6q11.1 -0.0466 0.0233

9p21.3 0.1149 -0.0574 4p15.31 0.0832 -0.0416 20p11.22 -0.0462 0.0231

4p15.2 0.1145 -0.0572 4q31.3 0.0791 -0.0396 2q33.3 0.045 -0.0225

3p21.2 0.1143 -0.0571 22q11.1 -0.0783 0.0391 11 q13.3 -0.0428 0.0214

3p25.2 0.1135 -0.0567 8p22 0.0745 -0.0372 4q26 0.0396 -0.0198

5q14.2 0.1127 -0.0564 8p21.2 0.0714 -0.0357 9p24.3 0.037 -0.0185

3p25.3 0.1103 -0.0551 4q23 0.068 -0.034 2q36.2 0.036 -0.018

5q33.2 0.109 -0.0545 5q31.1 0.0675 -0.0338 6p25.3 -0.0348 0.0174

4q31.23 0.1062 -0.0531 4p14 0.0645 -0.0323 4q31.21 0.0346 -0.0173

5p13.2 -0.1059 0.0529 8p21.1 0.0634 -0.0317 21 q21.3 0.0317 -0.0159

5p15.31 -0.1019 0.051 4q34.2 0.0631 -0.0315 15q25.3 -0.0312 0.0156

4q32.1 0.0978 -0.0489 8p23.2 0.0627 -0.0313 6p11.2 -0.0284 0.0142

4p15.1 0.0973 -0.0487 9p21.2 0.0621 -0.031 15q22.32 -0.0282 0.0141

4q32.3 0.0959 -0.048 3p12.2 0.0614 -0.0307 4q27 0.0274 -0.0137

3q24 -0.0936 0.0468 5q34 0.0602 -0.0301 10q21.1 0.0269 -0.0134

4q28.2 0.0929 -0.0464 4q22.3 0.0572 -0.0286 1 p13.3 0.0267 -0.0133

8p12 0.0924 -0.0462 4q31.22 0.0552 -0.0276 5p14.3 -0.0232 0.0116

5q32 0.0892 -0.0446 5q31.2 0.0537 -0.0268 5q31.3 0.0197 -0.0098

15q26.3 -0.0883 0.0442 4q24 0.052 -0.026 15q26.1 -0.0166 0.0083

5q23.3 0.0864 -0.0432 4q28.1 0.0518 -0.0259 2q36.3 0.0156 -0.0078

3q25.2 -0.0863 0.0431 15q26.2 -0.0513 0.0257 8q24.3 -0.0155 0.0077

5q21.2 0.0862 -0.0431 17p11.1 -0.0512 0.0256 2p16.3 -0.0151 0.0075

4q33 0.0844 -0.0422 3p12.1 0.0493 -0.0247 2q36.1 0.015 -0.0075

19q13.11 -0.0842 0.0421 1 p13.2 0.0491 -0.0245 8p21.3 0.0144 -0.0072

O o


Table 22 - Example CNV signature comprising 136 CNV bands (CNV band weights are given in columns XO.score and XL score)

CNV band XO.score X1.score CNV band XO.score X1.score CNV band XO.score X1.score

3q26.32 -0.3621 0.1811 5q11.2 0.1691 -0.0845 5p15.33 -0.126 0.063

3q26.31 -0.3344 0.1672 3p24.1 0.1688 -0.0844 3p14.1 0.1231 -0.0615

3q26.2 -0.3154 0.1577 3p24.2 0.1649 -0.0825 19q13.12 -0.1197 0.0599

3q26.33 -0.314 0.157 3p23 0.1593 -0.0796 5q21.3 0.1185 -0.0593

3q28 -0.303 0.1515 3p26.3 0.1589 -0.0794 5q23.1 0.118 -0.059

5q13.2 0.2994 -0.1497 5p15.1 -0.1588 0.0794 5q14.1 0.1143 -0.0571

3q26.1 -0.2767 0.1383 3p22.1 0.1564 -0.0782 3q25.33 -0.1116 0.0558

5q12.3 0.262 -0.131 9p23 0.1522 -0.0761 3q25.32 -0.1115 0.0558

3q27.3 -0.2529 0.1265 3p26.1 0.1514 -0.0757 3p25.1 0.1068 -0.0534

3q29 -0.2499 0.1249 19q13.13 -0.1448 0.0724 9p24.2 0.1068 -0.0534

3q27.2 -0.2437 0.1219 3p24.3 0.1438 -0.0719 5q33.3 0.1063 -0.0531

3q27.1 -0.2381 0.119 3p22.2 0.1426 -0.0713 2p25.3 -0.1059 0.0529

5q 12.1 0.2353 -0.1177 2q34 0.1399 -0.07 3q25.1 -0.1053 0.0527

5q 13.1 0.216 -0.108 3p21.33 0.1384 -0.0692 3q25.31 -0.1048 0.0524

5q22.1 0.2154 -0.1077 9p22.1 0.136 -0.068 19q 11 -0.1033 0.0516

21 q21 .2 0.2019 0.101 3p21.32 0.1353 -0.0676 3p14.2 0.1024 -0.0512

3p22.3 0.1868 -0.0934 9p24.1 0.1346 -0.0673 3p21.1 0.1022 -0.0511

9p22.2 0.186 -0.093 3p14.3 0.1345 -0.0672 5p15.32 -0.099 0.0495

5q13.3 0.1854 -0.0927 5q23.2 0.1343 -0.0672 3p21.31 0.0982 -0.0491

3p26.2 0.1842 -0.0921 5q15 0.134 -0.067 5p12 -0.0976 0.0488

5q14.3 0.1822 -0.0911 5q21.1 0.1338 -0.0669 5q33.1 0.0975 -0.0487

5q22.2 0.1814 -0.0907 5p14.1 -0.1312 0.0656 4q32.2 0.0974 -0.0487

5q22.3 0.1774 -0.0887 5p15.2 -0.1295 0.0648 5p13.3 -0.0934 0.0467

9p22.3 0.17 -0.085 3p13 0.1285 -0.0643 5p13.1 -0.0933 0.0466

Table 22 (continued)

CNV band XO.score X1.score CNV band XO.score X1.score CNV band XO.score X1.score

19q13.2 -0.0926 0.0463 19q12 -0.0594 0.0297 6q11.1 -0.0225 0.0112

9p21.3 0.0907 -0.0454 4p15.31 0.0591 -0.0295 20p11.22 0.0221 0.01 1 1

4p15.2 0.0903 -0.0452 4q31.3 0.055 -0.0275 2q33.3 0.0209 -0.0104

3p21.2 0.0901 -0.0451 22q11.1 -0.0541 0.0271 11 q13.3 -0.0187 0.0093

3p25.2 0.0893 -0.0447 8p22 0.0504 -0.0252 4q26 0.0154 -0.0077

5q14.2 0.0886 -0.0443 8p21.2 0.0473 -0.0236 9p24.3 0.0128 -0.0064

3p25.3 0.0862 -0.0431 4q23 0.0439 -0.0219 2q36.2 0.0118 -0.0059

5q33.2 0.0849 -0.0425 5q31.1 0.0434 -0.0217 6p25.3 -0.0107 0.0054

4q31.23 0.0821 -0.041 4p14 0.0404 0.0202 4q31.21 0.0105 -0.0052

5p13.2 -0.0817 0.0409 8p21.1 0.0393 -0.0196 21 q21.3 0.0076 -0.0038

5p15.31 -0.0778 0.0389 4q34.2 0.0389 -0.0195 15q25.3 -0.007 0.0035

4q32.1 0.0737 -0.0369 8p23.2 0.0386 -0.0193 6p11.2 -0.0042 0.0021

4p15.1 0.0732 -0.0366 9p21.2 0.0379 -0.019 15q22.32 -0.004 0.002

4q32.3 0.0718 -0.0359 3p12.2 0.0373 -0.0186 4q27 0.0033 -0.0016

3q24 -0.0695 0.0347 5q34 0.036 -0.018 10q21.1 0.0028 -0.0014

4q28.2 0.0687 -0.0344 4q22.3 0.0331 -0.0165 1 p13.3 0.0026 -0.0013

8p12 0.0682 -0.0341 4q31.22 0.031 -0.0155

5q32 0.0651 -0.0325 5q31.2 0.0295 -0.0148

15q26.3 -0.0642 0.0321 4q24 0.0279 -0.0139

5q23.3 0.0622 -0.0311 4q28.1 0.0277 -0.0138

3q25.2 -0.0621 0.0311 15q26.2 -0.0272 0.0136

5q21.2 0.0621 -0.031 17p11.1 -0.0271 0.0135

4q33 0.0602 -0.0301 3p12.1 0.0252 -0.0126

19q13.11 -0.0601 0.03 1 p13.2 0.0249 -0.0125

Table 23 - Example CNV signature comprising 120 CNV bands (CNV band weights are given in columns XO.score and Xl.score)

CNV band XO.score X1.score CNV band XO.score X1.score CNV band XO.score X1.score

3q26.32 -0.338 0.169 5q11.2 0.145 -0.0725 5p15.33 -0.1018 0.0509

3q26.31 -0.3102 0.1551 3p24.1 0.1446 -0.0723 3p14.1 0.0989 -0.0495

3q26.2 -0.2912 0.1456 3p24.2 0.1408 -0.0704 19q13.12 -0.0956 0.0478

3q26.33 -0.2898 0.1449 3p23 0.1351 -0.0676 5q21.3 0.0944 -0.0472

3q28 -0.2789 0.1394 3p26.3 0.1347 -0.0674 5q23.1 0.0938 -0.0469

5q13.2 0.2753 -0.1376 5p15.1 -0.1347 0.0673 5q14.1 0.0901 -0.0451

3q26.1 -0.2525 0.1263 3p22.1 0.1322 -0.0661 3q25.33 -0.0874 0.0437

5q12.3 0.2378 -0.1189 9p23 0.1281 -0.064 3q25.32 -0.0874 0.0437

3q27.3 -0.2288 0.1144 3p26.1 0.1273 -0.0636 3p25.1 0.0827 -0.0413

3q29 -0.2258 0.1129 19q13.13 -0.1206 0.0603 9p24.2 0.0827 -0.0413

3q27.2 -0.2196 0.1098 3p24.3 0.1197 -0.0598 5q33.3 0.0821 -0.0411

3q27.1 -0.2139 0.107 3p22.2 0.1185 -0.0592 2p25.3 -0.0817 0.0409

5q 12.1 0.2112 -0.1056 2q34 0.1158 -0.0579 3q25.1 -0.0812 0.0406

5q 13.1 0.1919 -0.0959 3p21.33 0.1143 -0.0571 3q25.31 -0.0807 0.0403

5q22.1 0.1913 -0.0956 9p22.1 0.1119 -0.056 19q 11 -0.0792 0.0396

21 q21 .2 0.1778 -0.0889 3p21.32 0.1111 -0.0556 3p14.2 0.0783 -0.0391

3p22.3 0.1626 -0.0813 9p24.1 0.1105 -0.0552 3p21.1 0.0781 -0.039

9p22.2 0.1619 -0.0809 3p14.3 0.1104 -0.0552 5p15.32 -0.0749 0.0375

5q13.3 0.1612 -0.0806 5q23.2 0.1102 -0.0551 3p21.31 0.0741 -0.037

3p26.2 0.16 -0.08 5q15 0.1099 -0.0549 5p12 -0.0734 0.0367

5q14.3 0.1581 -0.079 5q21.1 0.1097 -0.0549 5q33.1 0.0733 -0.0367

5q22.2 0.1572 -0.0786 5p14.1 -0.1071 0.0535 4q32.2 0.0733 -0.0366

5q22.3 0.1532 -0.0766 5p15.2 -0.1054 0.0527 5p13.3 -0.0693 0.0346

9p22.3 0.1458 -0.0729 3p13 0.1044 -0.0522 5p13.1 -0.0691 0.0346

Table 23 (continued)

CNV band XO.score X1.score CNV band XO.score X1.score

19q13.2 -0.0684 0.0342 19q12 -0.0353 0.0176

9p21.3 0.0666 -0.0333 4p15.31 0.035 -0.0175

4p15.2 0.0662 -0.0331 4q31.3 0.0309 -0.0154

3p21.2 0.066 -0.033 22q11.1 -0.03 0.015

3p25.2 0.0652 -0.0326 8p22 0.0262 -0.0131

5q14.2 0.0644 -0.0322 8p21.2 0.0231 -0.0116

3p25.3 0.062 -0.031 4q23 0.0197 -0.0099

5q33.2 0.0608 -0.0304 5q31.1 0.0192 -0.0096

4q31.23 0.0579 -0.029 4p14 0.0162 -0.0081

5p13.2 -0.0576 0.0288 8p21.1 0.0151 -0.0076

5p15.31 -0.0536 0.0268 4q34.2 0.0148 -0.0074

4q32.1 0.0496 -0.0248 8p23.2 0.0144 -0.0072

4p15.1 0.049 -0.0245 9p21.2 0.0138 -0.0069

4q32.3 0.0476 -0.0238 3p12.2 0.0131 -0.0066

3q24 -0.0453 0.0227 5q34 0.0119 -0.0059

4q28.2 0.0446 -0.0223 4q22.3 0.0089 -0.0045

8p12 0.0441 0.022 4q31.22 0.0069 -0.0034

5q32 0.0409 -0.0205 5q31.2 0.0054 -0.0027

15q26.3 -0.04 0.02 4q24 0.0038 -0.0019

5q23.3 0.0381 -0.0191 4q28.1 0.0035 -0.0018

3q25.2 -0.038 0.019 15q26.2 -0.0031 0.0015

5q21.2 0.0379 -0.019 17p11.1 -0.0029 0.0015

4q33 0.0361 -0.018 3p12.1 0.001 -0.0005

19q13.11 -0.0359 0.018 1 p13.2 0.0008 -0.0004

Table 24 - Example CNV signature comprising 101 CNV bands (CNV band weights are given in columns XO.score and XL score)

CNV band XO.score X1.score CNV band XO.score X1.score CNV band XO.score X1.score

3q26.32 -0.3138 0.1569 5q11.2 0.1208 -0.0604 5p15.33 -0.0777 0.0388

3q26.31 -0.2861 0.1431 3p24.1 0.1205 -0.0602 3p14.1 0.0748 -0.0374

3q26.2 -0.2671 0.1335 3p24.2 0.1166 -0.0583 19q13.12 -0.0714 0.0357

3q26.33 -0.2657 0.1329 3p23 0.1 1 1 -0.0555 5q21.3 0.0702 -0.0351

3q28 -0.2547 0.1274 3p26.3 0.1106 -0.0553 5q23.1 0.0697 -0.0348

5q13.2 0.2511 -0.1256 5p15.1 -0.1106 0.0553 5q14.1 0.066 -0.033

3q26.1 -0.2284 0.1142 3p22.1 0.1081 -0.054 3q25.33 -0.0633 0.0316

5q12.3 0.2137 -0.1068 9p23 0.1039 -0.052 3q25.32 -0.0632 0.0316

3q27.3 -0.2047 0.1023 3p26.1 0.1031 -0.0516 3p25.1 0.0585 -0.0293

3q29 -0.2016 0.1008 19q13.13 -0.0965 0.0482 9p24.2 0.0585 -0.0293

3q27.2 -0.1955 0.0977 3p24.3 0.0956 -0.0478 5q33.3 0.058 -0.029

3q27.1 -0.1898 0.0949 3p22.2 0.0944 -0.0472 2p25.3 -0.0576 0.0288

5q 12.1 0.1871 -0.0935 2q34 0.0916 -0.0458 3q25.1 -0.0571 0.0285

5q 13.1 0.1677 -0.0839 3p21.33 0.0901 -0.0451 3q25.31 -0.0566 0.0283

5q22.1 0.1671 -0.0836 9p22.1 0.0878 -0.0439 19q 11 -0.055 0.0275

21 q21 .2 0.1537 -0.0768 3p21.32 0.087 -0.0435 3p14.2 0.0541 -0.0271

3p22.3 0.1385 -0.0692 9p24.1 0.0863 -0.0432 3p21.1 0.0539 -0.027

9p22.2 0.1377 -0.0689 3p14.3 0.0862 -0.0431 5p15.32 -0.0508 0.0254

5q13.3 0.1371 -0.0685 5q23.2 0.0861 -0.043 3p21.31 0.0499 -0.025

3p26.2 0.1359 -0.0679 5q15 0.0857 -0.0429 5p12 -0.0493 0.0246

5q14.3 0.1339 -0.067 5q21.1 0.0856 -0.0428 5q33.1 0.0492 -0.0246

5q22.2 0.1331 -0.0665 5p14.1 -0.0829 0.0415 4q32.2 0.0491 -0.0246

5q22.3 0.1291 -0.0645 5p15.2 -0.0812 0.0406 5p13.3 -0.0451 0.0226

9p22.3 0.1217 -0.0608 3p13 0.0802 -0.0401 5p13.1 -0.045 0.0225

Table 24 (continued)

CNV band XO.score X1.score CNV band XO.score X1.score

19q13.2 -0.0443 0.0222 19q12 0.0112 0.0056

9p21.3 0.0425 0.0212 4p15.31 0.0108 -0.0054

4p15.2 0.042 0.021 4q31.3 0.0067 -0.0034

3p21.2 0.0418 -0.0209 22q11.1 -0.0058 0.0029

3p25.2 0.0411 -0.0205 8p22 0.0021 0.001

5q14.2 0.0403 0.0201

3p25.3 0.0379 -0.0189

5q33.2 0.0366 -0.0183

4q31.23 0.0338 -0.0169

5p13.2 -0.0335 0.0167

5p15.31 -0.0295 0.0147

4q32.1 0.0254 -0.0127

4p15.1 0.0249 -0.0124

4q32.3 0.0235 -0.0117

3q24 0.0212 0.0106

4q28.2 0.0204 0.0102

8p12 0.02 0.01

5q32 0.0168 -0.0084

15q26.3 -0.0159 0.0079

5q23.3 0.014 -0.007

3q25.2 -0.0139 0.0069

5q21.2 0.0138 -0.0069

4q33 0.012 -0.006

19q13.11 -0.0118 0.0059

112

Table 25 - Example CNV signature comprising 85 CNV bands (CNV band weights are given in columns XO.score and XL score)

CNV band XO.score X1.score CNV band XO.score X1.score CNV band XO.score X1.score

3q26.32 -0.2897 0.1449 5q11.2 0.0967 -0.0483 5p15.33 -0.0536 0.0268

3q26.31 -0.262 0.131 3p24.1 0.0964 -0.0482 3p14.1 0.0507 -0.0253

3q26.2 -0.2429 0.1215 3p24.2 0.0925 -0.0462 19q13.12 -0.0473 0.0237

3q26.33 -0.2416 0.1208 3p23 0.0868 -0.0434 5q21.3 0.0461 -0.0231

3q28 -0.2306 0.1153 3p26.3 0.0865 -0.0432 5q23.1 0.0455 -0.0228

5q13.2 0.227 -0.1135 5p15.1 -0.0864 0.0432 5q14.1 0.0419 -0.0209

3q26.1 -0.2042 0.1021 3p22.1 0.0839 -0.042 3q25.33 -0.0391 0.0196

5q12.3 0.1895 -0.0948 9p23 0.0798 -0.0399 3q25.32 -0.0391 0.0195

3q27.3 -0.1805 0.0903 3p26.1 0.079 -0.0395 3p25.1 0.0344 -0.0172

3q29 -0.1775 0.0887 19q13.13 -0.0723 0.0362 9p24.2 0.0344 -0.0172

3q27.2 -0.1713 0.0857 3p24.3 0.0714 -0.0357 5q33.3 0.0339 -0.0169

3q27.1 -0.1656 0.0828 3p22.2 0.0702 -0.0351 2p25.3 -0.0334 0.0167

5q 12.1 0.1629 -0.0815 2q34 0.0675 -0.0337 3q25.1 -0.0329 0.0165

5q 13.1 0.1436 -0.0718 3p21.33 0.066 -0.033 3q25.31 -0.0324 0.0162

5q22.1 0.143 -0.0715 9p22.1 0.0636 -0.0318 19q 11 -0.0309 0.0154

21 q21 .2 0.1295 -0.0648 3p21.32 0.0628 -0.0314 3p14.2 0.03 -0.015

3p22.3 0.1144 -0.0572 9p24.1 0.0622 -0.0311 3p21.1 0.0298 -0.0149

9p22.2 0.1136 -0.0568 3p14.3 0.0621 -0.031 5p15.32 -0.0266 0.0133

5q13.3 0.1129 -0.0565 5q23.2 0.0619 -0.031 3p21.31 0.0258 -0.0129

3p26.2 0.1118 -0.0559 5q15 0.0616 -0.0308 5p12 -0.0251 0.0126

5q14.3 0.1098 -0.0549 5q21.1 0.0614 -0.0307 5q33.1 0.025 -0.0125

5q22.2 0.1089 -0.0545 5p14.1 -0.0588 0.0294 4q32.2 0.025 -0.0125

5q22.3 0.105 -0.0525 5p15.2 -0.0571 0.0285 5p13.3 0.021 0.0105

9p22.3 0.0975 -0.0488 3p13 0.0561 -0.028 5p13.1 -0.0209 0.0104


Table 26 - Example CNV signature comprising 70 CNV bands (CNV band weights are given in columns XO.score and XL score)

CNV band XO.score X1.score CNV band XO.score X1.score CNV band XO.score X1.score

3q26.32 -0.2656 0.1328 5q11.2 0.0725 -0.0363 5p15.33 -0.0294 0.0147

3q26.31 -0.2378 0.1189 3p24.1 0.0722 -0.0361 3p14.1 0.0265 -0.0133

3q26.2 -0.2188 0.1094 3p24.2 0.0684 -0.0342 19q13.12 -0.0232 0.0116

3q26.33 -0.2174 0.1087 3p23 0.0627 -0.0313 5q21.3 0.022 0.011

3q28 -0.2064 0.1032 3p26.3 0.0623 -0.0312 5q23.1 0.0214 -0.0107

5q13.2 0.2029 -0.1014 5p15.1 -0.0623 0.0311 5q14.1 0.0177 -0.0089

3q26.1 -0.1801 0.0901 3p22.1 0.0598 -0.0299 3q25.33 -0.015 0.0075

5q12.3 0.1654 -0.0827 9p23 0.0556 -0.0278 3q25.32 -0.015 0.0075

3q27.3 -0.1564 0.0782 3p26.1 0.0548 -0.0274 3p25.1 0.0103 -0.0051

3q29 -0.1533 0.0767 19q13.13 -0.0482 0.0241 9p24.2 0.0103 -0.0051

3q27.2 -0.1472 0.0736 3p24.3 0.0473 -0.0236 5q33.3 0.0097 -0.0049

3q27.1 -0.1415 0.0707 3p22.2 0.0461 -0.023 2p25.3 -0.0093 0.0046

5q 12.1 0.1388 -0.0694 2q34 0.0433 -0.0217 3q25.1 -0.0088 0.0044

5q 13.1 0.1194 -0.0597 3p21.33 0.0419 -0.0209 3q25.31 -0.0083 0.0041

5q22.1 0.1188 -0.0594 9p22.1 0.0395 -0.0197 19q 11 -0.0067 0.0034

21 q21 .2 0.1054 -0.0527 3p21.32 0.0387 -0.0193 3p14.2 0.0058 -0.0029

3p22.3 0.0902 -0.0451 9p24.1 0.0381 -0.019 3p21.1 0.0057 -0.0028

9p22.2 0.0894 -0.0447 3p14.3 0.0379 -0.019 5p15.32 -0.0025 0.0012

5q13.3 0.0888 -0.0444 5q23.2 0.0378 -0.0189 3p21.31 0.0016 -0.0008

3p26.2 0.0876 -0.0438 5q15 0.0374 -0.0187 5p12 0.001 0.0005

5q14.3 0.0857 -0.0428 5q21.1 0.0373 -0.0186 5q33.1 0.0009 -0.0004

5q22.2 0.0848 -0.0424 5p14.1 -0.0346 0.0173 4q32.2 0.0008 -0.0004

5q22.3 0.0808 -0.0404 5p15.2 -0.0329 0.0165

9p22.3 0.0734 -0.0367 3p13 0.032 -0.016

Table 27 - Example CNV signature comprising 50 CNV bands (CNV band weights are given in columns XO.score and XL score)

CNV band XO.score X1.score CNV band XO.score X1.score CNV band XO.score X1.score

3q26.32 -0.2414 0.1207 5q11.2 0.0484 -0.0242 5p15.33 -0.0053 0.0026

3q26.31 -0.2137 0.1068 3p24.1 0.0481 -0.024 3p14.1 0.0024 0.0012

3q26.2 -0.1947 0.0973 3p24.2 0.0442 0.0221

3q26.33 -0.1933 0.0966 3p23 0.0385 -0.0193

3q28 -0.1823 0.0912 3p26.3 0.0382 -0.0191

5q13.2 0.1787 -0.0894 5p15.1 -0.0381 0.0191

3q26.1 -0.156 0.078 3p22.1 0.0356 -0.0178

5q12.3 0.1413 -0.0706 9p23 0.0315 -0.0157

3q27.3 -0.1322 0.0661 3p26.1 0.0307 -0.0153

3q29 -0.1292 0.0646 19q13.13 -0.0241 0.012

3q27.2 -0.123 0.0615 3p24.3 0.0231 -0.0116

3q27.1 -0.1174 0.0587 3p22.2 0.0219 0.011

5q 12.1 0.1146 -0.0573 2q34 0.0192 -0.0096

5q 13.1 0.0953 -0.0476 3p21.33 0.0177 -0.0089

5q22.1 0.0947 -0.0473 9p22.1 0.0153 -0.0077

21 q21 .2 0.0812 -0.0406 3p21.32 0.0146 -0.0073

3p22.3 0.0661 -0.033 9p24.1 0.0139 -0.007

9p22.2 0.0653 -0.0326 3p14.3 0.0138 -0.0069

5q13.3 0.0646 -0.0323 5q23.2 0.0136 -0.0068

3p26.2 0.0635 -0.0317 5q15 0.0133 -0.0066

5q14.3 0.0615 -0.0308 5q21.1 0.0131 -0.0066

5q22.2 0.0606 -0.0303 5p14.1 -0.0105 0.0053

5q22.3 0.0567 -0.0283 5p15.2 -0.0088 0.0044

9p22.3 0.0493 -0.0246 3p13 0.0078 -0.0039

Table 28 - Example CNV signature comprising 33 CNV bands (CNV band weights are given in columns XO.score and Xl.score)

CNV band XO.score X1.score CNV band XO.score Xl .score

3q26.32 -0.2173 0.1086 5q11.2 0.0242 0.0121

3q26.31 -0.1895 0.0948 3p24.1 0.0239 0.012

3q26.2 -0.1705 0.0853 3p24.2 0.0201 0.01

3q26.33 -0.1691 0.0846 3p23 0.0144 -0.0072

3q28 -0.1582 0.0791 3p26.3 0.014 -0.007

5q13.2 0.1546 -0.0773 5p15.1 -0.014 0.007

3q26.1 -0.1318 0.0659 3p22.1 0.0115 -0.0058

5q12.3 0.1171 -0.0586 9p23 0.0073 -0.0037

3q27.3 -0.1081 0.054 3p26.1 0.0066 -0.0033

3q29 -0.105 0.0525

3q27.2 -0.0989 0.0494

3q27.1 -0.0932 0.0466

5q 12.1 0.0905 -0.0452

5q 13.1 0.0712 -0.0356

5q22.1 0.0705 -0.0353

21 q21 .2 0.0571 -0.0285

3p22.3 0.0419 0.021

9p22.2 0.0411 -0.0206

5q13.3 0.0405 -0.0203

3p26.2 0.0393 -0.0197

5q14.3 0.0374 -0.0187

5q22.2 0.0365 -0.0183

5q22.3 0.0325 -0.0163

9p22.3 0.0251 -0.0126

117

Table 29 - Example CNV signature comprising 25 CNV bands (CNV band weights are given in columns XO.score and XL score)

CNV band XO.score X1.score CNV band XO.score X1.score

3q26.32 -0.1931 0.0966 5q11.2 0.0001 0.0001

3q26.31 -0.1654 0.0827

3q26.2 -0.1464 0.0732

3q26.33 -0.145 0.0725

3q28 -0.134 0.067

5q13.2 0.1304 -0.0652

3q26.1 -0.1077 0.0538

5q12.3 0.093 -0.0465

3q27.3 -0.084 0.042

3q29 -0.0809 0.0405

3q27.2 -0.0748 0.0374

3q27.1 -0.0691 0.0345

5q 12.1 0.0663 -0.0332

5q 13.1 0.047 -0.0235

5q22.1 0.0464 -0.0232

21 q21 .2 0.0329 -0.0165

3p22.3 0.0178 -0.0089

9p22.2 0.017 -0.0085

5q13.3 0.0164 -0.0082

3p26.2 0.0152 -0.0076

5q14.3 0.0132 -0.0066

5q22.2 0.0124 -0.0062

5q22.3 0.0084 -0.0042

9p22.3 0.001 -0.0005

Table 30 - Example CNV signature comprising 16 CNV bands (CNV band weights are given in columns XO.score and XL score)

CNV band XO.score X1.score

3q26.32 -0.169 0.0845

3q26.31 -0.1413 0.0706

3q26.2 0.1222 0.0611

3q26.33 -0.1209 0.0604

3q28 -0.1099 0.0549

5q13.2 0.1063 -0.0531

3q26.1 -0.0835 0.0418

5q12.3 0.0688 -0.0344

3q27.3 -0.0598 0.0299

3q29 -0.0568 0.0284

3q27.2 -0.0506 0.0253

3q27.1 -0.0449 0.0225

5q 12.1 0.0422 0.021 1

5q 13.1 0.0229 -0.0114

5q22.1 0.0223 0.01 1 1

21 q21 .2 0.0088 -0.0044

Table 31 - Example CNV signature comprising 13 CNV bands (CNV band weights are given in columns XO.score and XL score)

CNV band XO.score X1.score

3q26.32 -0.1449 0.0724

3q26.31 -0.1171 0.0586

3q26.2 -0.0981 0.049

3q26.33 -0.0967 0.0484

3q28 -0.0857 0.0429

5q13.2 0.0821 -0.0411

3q26.1 -0.0594 0.0297

5q12.3 0.0447 -0.0223

3q27.3 -0.0357 0.0178

3q29 -0.0326 0.0163

3q27.2 -0.0265 0.0132

3q27.1 -0.0208 0.0104

5q12.1 0.0181 -0.009

Table 32 - Example CNV signature comprising 11 CNV bands (CNV band weights are given in columns XO. score and XI. score)

CNV band XO.score X1.score

3q26.32 -0.1207 0.0604

3q26.31 -0.093 0.0465

3q26.2 -0.074 0.037

3q26.33 -0.0726 0.0363

3q28 -0.0616 0.0308

5q13.2 0.058 -0.029

3q26.1 -0.0353 0.0176

5q12.3 0.0205 -0.0103

3q27.3 -0.0115 0.0058

3q29 -0.0085 0.0042

3q27.2 -0.0023 0.0012

Table 33 - Example CNV signature comprising 7 CNV bands (CNV band weights are given in columns XO.score and XL score)

CNV band XO.score X1.score

3q26.32 -0.0966 0.0483

3q26.31 -0.0688 0.0344

3q26.2 -0.0498 0.0249

3q26.33 -0.0484 0.0242

3q28 -0.0375 0.0187

5q13.2 0.0339 -0.0169

3q26.1 0.01 1 1 0.0056

Table 34 - Example CNV signature comprising 6 CNV bands (CNV band weights are given in columns XO.score and XL score)

CNV band XO.score X1.score

3q26.32 -0.0724 0.0362

3q26.31 -0.0447 0.0223

3q26.2 -0.0257 0.0128

3q26.33 -0.0243 0.0121

3q28 -0.0133 0.0067

5q13.2 0.0097 -0.0049

Table 35 - Example CNV signature comprising 4 CNV bands (CNV band weights are given in columns XO.score and XL score)

CNV band XO.score X1.score

3q26.32 -0.0483 0.0241

3q26.31 -0.0205 0.0103

3q26.2 -0.0015 0.0008

3q26.33 0.0001 0.0001

Table 36 - Example CNV signature comprising 1 CNV band (CNV band weights are given in columns XO.score and XL score)

CNV band XO.score X1.score

3q26.32 -0.0241 0.0121

Table 37 - Summary of characteristic parameters and metrics of example CNV signatures

CNV

CNV Threshold Sensitivity Specificity Sensitivity Specificity

Signature

Bands Validation) AUV <TCGA> Value (Validation) (Validation) (TCGA) (TCGA)

Table

19 219 0.885417 0.981012 0.015 1 0.75 1 0.908396947

20 179 0.881944 0.980237 0.01 1 0.75 1 0.923664122 21 155 0.878472 0.979527 0.015 1 0.75 0.9958159 0.929389313 22 136 0.881944 0.978976 0.01 0.791666667 0.833333333 0.991631799 0.946564885

23 120 0.885417 0.978473 0.01 0.083333333 0.916666667 0.966527197 0.958015267

24 101 0.888889 0.977706 0.02 0.041666667 1 0.920502092 0.961832061

25 85 0.885417 0.976492 0.06 0.041666667 1 0.941422594 0.958015267

26 70 0.895833 0.974456 0.12 0 1 0.907949791 0.959923664

27 50 0.920139 0.972013 0.185 0 1 0.820083682 0.961832061

28 33 0.930556 0.966735 0.31 0 1 0.907949791 0.942748092

29 25 0.954861 0.963174 0.405 0 1 0.912133891 0.940839695

30 16 0.958333 0.960123 0.48 0.041666667 1 0.933054393 0.927480916

31 13 0.965278 0.957281 0.555 0.5 1 0.979079498 0.893129771

32 1 1 0.958333 0.953919 0.6 0.75 0.916666667 0.987447699 0.879770992

33 7 0.958333 0.951715 0.65 1 0.666666667 1 0.753816794

34 6 0.888889 0.949871 0.66 1 0.666666667 1 0.709923664

35 4 0.861111 0.940896 0.545 0 1 0 1

36 1 0.84375 0.948861 0.62 0 1 0 1

References

1 Huber, W. et al. Orchestrating high-throughput genomic analysis with Bioconductor. Nat Methods 12, 1 15-121 , doi: 10.1038/nmeth.3252 (2015).

2 Jeremy George, P. et al. Surveillance for the detection of early lung cancer in patients with bronchial dysplasia. Thorax 62, 43-50, doi: 10.1136/thx.2005.052191 (2007).

3 Bibikova, M. et al. High density DNA methylation array with single CpG site resolution.

Genomics 98, 288-295, doi: 10.1016/j.ygeno.2011.07.007 (201 1).

4 Sandoval, J. et al. Validation of a DNA methylation microarray for 450,000 CpG sites in the human genome. Epigenetics 6, 692-702 (201 1).

5 Morris, T. J. et al. ChAMP: 450k Chip Analysis Methylation Pipeline. Bioinformatics 30, 428-430, doi: 10.1093/bioinformatics/btt684 (2014).

6 Teschendorff, A. E. et al. DNA methylation outliers in normal breast tissue identify field defects that are enriched in cancer. Nat Commun 7, 10478, doi:10.1038/ncommsl0478 (2016).

7 Johnson, W. E., Li, C. & Rabinovic, A. Adjusting batch effects in microarray expression data using empirical Bayes methods. Biostatistics 8, 118-127, doi: 10.1093/biostatistics/kxj037 (2007).

8 Ritchie, M. E. et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic acids research 43, e47, doi:10.1093/nar/gkv007 (2015).

9 Benjamini, Y., Drai, D., Elmer, G., Kafkafi, N. & Golani, I. Controlling the false discovery rate in behavior genetics research. Behav Brain Res 125, 279-284 (2001).

10 Kolde, R. Pheatmap: pretty heatmaps. R package version 61 (2012).

1 1 Tibshirani, R., Hastie, T., Narasimhan, B. & Chu, G. Diagnosis of multiple cancer types by shrunken centroids of gene expression. Proceedings of the National Academy of Sciences of the United States of America 99, 6567-6572, doi: 10.1073/pnas.082099299 (2002).

12 Robin, X. et al. pROC: an open-source package for R and S+ to analyze and compare ROC curves. BMC Bioinformatics 12, 77, doi: 10.1 186/1471 -2105-12-77 (201 1 ).

13 Keilwagen, J., Grosse, I. & Grau, J. Area under precision -recall curves for weighted and unweighted data. PLoS One 9, e92209, doi:10.1371/journal. pone.0092209 (2014).

14 Raine, K. M. et al. ascatNgs: Identifying Somatically Acquired Copy-Number Alterations from Whole-Genome Sequencing Data. Current protocols in bioinformatics 56, 15 19 11 -15 19 17, doi:10.1002/cpbi.l 7 (2016).

15 Feber, A. et al. Using high-density DNA methylation arrays to profile copy number alterations. Genome Biol 15, R30, dok lO.l 186/gb-2014-15-2-r30 (2014).

16 van Boerdonk, R. A. et al. DNA copy number alterations in endobronchial squamous metaplastic lesions predict lung cancer. American journal of respiratory and critical care medicine 184, 948-956, doklO.l 164/rccm.201 102-02180C (201 1).

17 van Boerdonk, R. A. et al. DNA copy number aberrations in endobronchial lesions: a validated predictor for cancer. Thorax 69, 451 -457, doi: 10.1 136/thoraxjnl-2013-203821 (2014).

18 Grossman, R. L. et al. Toward a Shared Vision for Cancer Genomic Data. N Engl J Med 375,

1 109- 1 1 12, doi : 10.1056/NEJMp 1607591 (2016).

19 Law, C. W., Chen, Y., Shi, W. & Smyth, G. K. voom: Precision weights unlock linear model analysis tools for RNA-seq read counts. Genome biology 15, R29, doi: 10.1 186/gb-2014-15-2-r29 (2014).

20 Luo, W., Friedman, M. S., Shedden, K., Hankenson, K. D. & Woolf, P. J. GAGE: generally applicable gene set enrichment for pathway analysis. BMC Bioinformatics 10, 161 , doi: 10.1 186/1471 -2105-10-161 (2009).

21 Kanehisa, M. & Goto, S. KEGG: lcyoto encyclopedia of genes and genomes. Nucleic acids research 28, 27-30 (2000).

22 Kanehisa, M., Sato, Y., Kawashima, M., Furumichi, M. & Tanabe, M. KEGG as a reference resource for gene and protein annotation. Nucleic acids research 44, D457-462, doi: 10.1093/nar/gkvl 070 (2016).

23 Kanehisa, M., Furumichi, M., Tanabe, M., Sato, Y. & Morishima, K. KEGG: new perspectives on genomes, pathways, diseases and drugs. Nucleic acids research 45, D353-D361 ,

doi: 10.1093/nar/gkw 1092 (2017).

24 Carter, S. L., Eklund, A. C., Kohane, I. S., Harris, L. N. & Szallasi, Z. A signature of chromosomal instability inferred from gene expression profdes predicts clinical outcome in multiple human cancers. Nat Genet 38, 1043-1048, doi: 10.1038/ngl 861 (2006).

25 Whitfield, M. L. et al. Identification of genes periodically expressed in the human cell cycle and their expression in tumors. Mol Biol Cell 13, 1977-2000, doi: 10.1091/mbc.02-02-0030. (2002).

26 Jones, D. et al. cgpCaVEManWrapper: Simple Execution of CaVEMan in Order to Detect Somatic Single Nucleotide Variants in NGS Data. Current protocols in bioinformatics 56, 15 10 1 1 -15 10 18, doi: 10.1002/cpbi.20 (2016).

27 Van Loo, P. et al. Allele-specific copy number analysis of tumors. Proc Natl Acad Sci U S A 107, 16910-16915, doi : 10.1073/pnas.1009843107 (2010).

28 Nik-Zainal, S. et al. The life history of 21 breast cancers. Cell 149, 994-1007,

doi: 10.1016/j. cell.2012.04.023 (2012).

29 Skinner, M. E., Uzilov, A. V., Stein, L. D., Mungall, C. J. & Holmes, I. H. JBrowse: a next-generation genome browser. Genome Res 19, 1630-1638, doi : 10.1 101 /gr.094607.109 (2009).

30 Raine, K. M. et al. cgpPindel: Identifying Somatically Acquired Insertion and Deletion Events from Paired End Sequencing. Curr Protoc Bioinformatics 52, 15 17 1 1 -12,

doi : 10.1002/0471250953.bi 1507s52 (2015).

31 Ye, K., Schulz, M. H., Long, Q., Apweiler, R. & Ning, Z. Pindel: a pattern growth approach to detect break points of large deletions and medium sized insertions from paired-end short reads.

Bioinformatics 25, 2865-2871 , doi: 10.1093/bioinformatics/btp394 (2009).

32 Papaemmanuil, E. et al. RAG-mediated recombination is the predominant driver of oncogenic rearrangement in ETV6-RUNX1 acute lymphoblastic leukemia. Nature genetics 46, 1 16-125, doi: 10.1038/ng.2874 (2014).

33 Robinson, J. T. et al. Integrative genomics viewer. Nat Biotechnol 29, 24-26,

doi: 10.1038/nbt.1754 (2011).

34 Thorvaldsdottir, H., Robinson, J. T. & Mesirov, J. P. Integrative Genomics Viewer (IGV): high-performance genomics data visualization and exploration. Brief Bioinform 14, 178-192,

doi : 10.1093/bib/bbs017 (2013).

35 Endesfelder, D. et al. Chromosomal instability selects gene copy-number variants encoding core regulators of proliferation in ER+ breast cancer. Cancer Res 74, 4853-4863, doi: 10.1 158/0008-5472.CAN-13-2664 (2014).

36 Forbes, S. A. et al. COSMIC: somatic cancer genetics at high-resolution. Nucleic acids research 45, D777-D783, doi: 10.1093/nar/gkwl 121 (2017).

37 McLaren, W. et al. The Ensembl Variant Effect Predictor. Genome Biol 17, 122,

doi: 10.1 186/s 13059-016-0974-4 (2016).

38 Martincorena, I. et al. Universal Patterns of Selection in Cancer and Somatic Tissues. Cell 171 , 1029-1041 e 10 1 , doi: 10.1016/j .cell.2017.09.042 (2017).

39 Miller, C. A. et al. SciClone: inferring clonal architecture and tracking the spatial and temporal patterns of tumor evolution. PLoS Comput Biol 10, el003665, doi: 10.1371/joumal.pcbi.1003665 (2014).

40 McGranahan, N. et al. Clonal neoantigens elicit T cell immunoreactivity and sensitivity to immune checkpoint blockade. Science 351 , 1463-1469, doi: 10.1126/science. aafl490 (2016).

41 Blokzijl, F., Janssen, R., van Boxtel, R. & Cuppen, E. MutationalPatterns: comprehensive genome-wide analysis of mutational processes. Genome Med 10, 33, doi: 10.1 186/s 13073 -018-0539-0 (2018).

42 Alexandrov, L. B. et al. Clock-like mutational processes in human somatic cells. Nat Genet 47, 1402-1407, doi: 10.1038/ng.3441 (2015).

43 Alexandrov, L. B. et al. Signatures of mutational processes in human cancer. Nature 500, 415-421 , doi:10.1038/naturel2477 (2013).

44 Martincorena, I. & Campbell, P. J. Somatic mutation in cancer and normal cells. Science 349,

1483- 1489, doi: 10.1 126/science.aab4082 (2015).

45 Farmery, J. H. R., Smith, M. L. & Lynch, A. G. Telomerecat: A Ploidy-Agnostic Method For Estimating Telomere Length From Whole Genome Sequencing Data. bioRxiv, doi: 10.1101/139972 (2017).

46 Olkhov-Mitsel, E and Bapat, B: Strategies for discovery and validation of methylated and hydroxymethylated DNA biomarkers. Cancer Medicine 2012, 1(2): 237-260.

47 Herman, J. G., Graff, J. R., Myohanen, S., Nelkin, B. D. & Baylin, S. B.: Methylation -specific PCR: a novel PCR assay for methylation status of CpG islands. Proc. Natl Acad. Sci. USA 1996, 93: 9821-9826.

48 Frommer, M. et al.: A genomic sequencing protocol that yields a positive display of 5-methylcytosine residues in individual DNA strands. Proc. Natl Acad. Sci. USA 1992, 89: 1827-1831.

49 Xiong, Z. & Laird, P. W.: COBRA: a sensitive and quantitative DNA methylation assay.

Nucleic Acids Res. 1997, 25: 2532-2534.

50 Gonzalgo, M. L. & Jones, P. A.: Rapid quantitation of methylation differences at specific sites using methylationsensitive single nucleotide primer extension (Ms-SNuPE). Nucleic Acids Res. 1997, 25: 2529-2531.

51 Paul DS, Guilhamon P, Karpathakis A, Butcher LM, Thirlwell C, Feber A, Beck S: Assessment of RainDrop BS-seq as a method for large-scale, targeted bisulfite sequencing. Epigenetics 2014, 9.