Anticancer drug synergy prediction based on CatBoost

Abstract

Abstract Background The research of cancer treatments has always been a hot topic in medical field. Cancer monotherapy as a common therapy has been proven to have many disadvantages such as toxicity and drug resistance. With the development of network pharmacology, multi-targeted combination drugs have become an ideal option for cancer treatment. Since the number of potential drug combinations is very huge, it is not feasible to use clinical experience or high-throughput screening to identify the complete combinatorial space. Methods such as machine learning models offer the possibility to explore the combinatorial space effectively. Results In this work, we proposed a machine learning method based on CatBoost to predict the synergy scores of anticancer drug combinations on cancer cell lines, which utilized oblivious trees and Ordered Boosting technique to avoid overfitting and bias. The model was trained and tested using the data screened from NCI-ALMANAC dataset. The drugs were characterized with morgan fingerprints, drug target information, monotherapy information, and the cell lines were described with gene expression profiles. In the stratified five-fold cross-validation, our method obtained excellent results and performed significantly better than three other advanced models. Additionally, when using SHAP to interpret the biological significance of the prediction results, we found that those genes with some associations with cancer occurrence played an important role in the prediction effect. Conclusions The model based on CatBoost has good quality for predicting drug synergy and could be considered as an optional method for anticancer drug combination research.