Search International and National Patent Collections

1. (WO2018103684) METHOD FOR ESTABLISHING MACHINE LEARNING MODEL FOR PREDICTING TOXICITY OF SIRNA AGAINST CERTAIN TYPE OF CELL AND USE THEREOF

Pub. No.:    WO/2018/103684    International Application No.:    PCT/CN2017/114938
Publication Date: Fri Jun 15 01:59:59 CEST 2018 International Filing Date: Fri Dec 08 00:59:59 CET 2017
IPC: G06F 19/00
Applicants: HANGZHOU CONVERD CO., LTD.
杭州康万达医药科技有限公司
Inventors: CAI, Jinlu
蔡金露
ZHONG, Nan
钟南
ZHANG, Qingyong
张庆勇
JIN, Ying
金盈
ZHANG, Xiuqin
张秀琴
Title: METHOD FOR ESTABLISHING MACHINE LEARNING MODEL FOR PREDICTING TOXICITY OF SIRNA AGAINST CERTAIN TYPE OF CELL AND USE THEREOF
Abstract:
Provided are a method for establishing a machine learning model for predicting the toxicity of an siRNA against a certain type of cell and the use. The method comprises A) providing n siRNAs of 19-29 bp, n ≥ 2; B) obtaining the input and output values for each siRNA to establish the model, wherein the input values are obtained as follows: i) aligning the siRNAs to the genomic mRNAs, and selecting off-target genes that are complementary and have mismatched bases less than or equal to 7; ii) obtaining the off-target weight according to characteristics of the mismatched bases in the complementary region and the secondary structures of mRNAs; iii) obtaining omics weight for off-target genes using a database; iv) based on the omics weight and off-target weight for all off-target genes, calculating omics characteristic values as the input values; and the output values are obtained as follows: performing cell experiments with siRNAs to obtain cell survival indexes under siRNAs as the output values; and C) calculating the input and output values for n siRNAs by means of a machine learning algorithm.