Drug-induced cell viability prediction from LINCS-L1000 through WRFEN-XGBoost algorithm

Abstract Background Predicting the drug response of cancer diseases through cellular perturbation signatures under action specific compounds is very important in personalized medicine. In process testing responses to cancer, traditional experimental methods have been greatly hampered by cost and sample size. At present, public availability large amounts gene expression data makes it a challenging task use machine learning predict sensitivity. Results this study, we introduced WRFEN-XGBoost cell viability prediction algorithm based on LINCS-L1000 signatures. We integrated LINCS-L1000, CTRP Achilles datasets adopted weighted fusion random forest elastic net for key selection. Then FEBPSO was into XGBoost induced drugs. The proposed method compared with some new methods, found that our model achieved good results 0.83 Pearson correlation. same time, completed sensitivity validation NCI60 CCLE datasets, which further demonstrated effectiveness method. Conclusions showed conducive elucidation disease mechanisms exploration therapies, promoted progress clinical