Predicting tumor cell line response to drug pairs with deep learning

Fangfang Xia, Maulik Shukla, Thomas Brettin, Cristina Garcia-Cardona, Judith Cohn, Jonathan E. Allen, Sergei Maslov, Susan L. Holbeck, James H. Doroshow, Yvonne A. Evrard, Eric A. Stahlberg, Rick L. Stevens
2018 BMC Bioinformatics  
The National Cancer Institute drug pair screening effort against 60 well-characterized human tumor cell lines (NCI-60) presents an unprecedented resource for modeling combinational drug activity. Results: We present a computational model for predicting cell line response to a subset of drug pairs in the NCI-ALMANAC database. Based on residual neural networks for encoding features as well as predicting tumor growth, our model explains 94% of the response variance. While our best result is
more » ... d with a combination of molecular feature types (gene expression, microRNA and proteome), we show that most of the predictive power comes from drug descriptors. To further demonstrate value in detecting anticancer therapy, we rank the drug pairs for each cell line based on model predicted combination effect and recover 80% of the top pairs with enhanced activity. Conclusions: We present promising results in applying deep learning to predicting combinational drug response. Our feature analysis indicates screening data involving more cell lines are needed for the models to make better use of molecular features. Background Deep learning has already revolutionized the fields of computer vision, robotics, gaming, and natural language processing. It is rapidly making strides in genomics, medical diagnosis, and computational chemistry. At the heart of deep learning is a set of generalizable techniques that thrive with large-scale data in inferring complex relationships. While emerging neural networks are outperforming state-of-the-art models in predicting a wide range of molecular properties, to date, their use has been limited to single molecules due to the lack of training data. The recently published NCI-ALMANAC resource [1] will change that. It offers a promising look at the combined
doi:10.1186/s12859-018-2509-3 fatcat:inq24w5lq5er7g4hlxd5o7njru