Yejin Kim1, Shuyu Zheng2, Jing Tang2, Wenjin Jim Zheng1, Zhao Li1, Xiaoqian Jiang1. 1. Center for Safe Artificial Intelligence for Healthcare, School of Biomedical Informatics, University of Texas Health Science Center at Houston, Houston, Texas, USA. 2. Research Program in Systems Oncology, Faculty of Medicine, University of Helsinki, Helsinki, Finland.
Abstract
OBJECTIVE: Drug combination screening has advantages in identifying cancer treatment options with higher efficacy without degradation in terms of safety. A key challenge is that the accumulated number of observations in in-vitro drug responses varies greatly among different cancer types, where some tissues are more understudied than the others. Thus, we aim to develop a drug synergy prediction model for understudied tissues as a way of overcoming data scarcity problems. MATERIALS AND METHODS: We collected a comprehensive set of genetic, molecular, phenotypic features for cancer cell lines. We developed a drug synergy prediction model based on multitask deep neural networks to integrate multimodal input and multiple output. We also utilized transfer learning from data-rich tissues to data-poor tissues. RESULTS: We showed improved accuracy in predicting synergy in both data-rich tissues and understudied tissues. In data-rich tissue, the prediction model accuracy was 0.9577 AUROC for binarized classification task and 174.3 mean squared error for regression task. We observed that an adequate transfer learning strategy significantly increases accuracy in the understudied tissues. CONCLUSIONS: Our synergy prediction model can be used to rank synergistic drug combinations in understudied tissues and thus help to prioritize future in-vitro experiments. Code is available at https://github.com/yejinjkim/synergy-transfer.
OBJECTIVE: Drug combination screening has advantages in identifying cancer treatment options with higher efficacy without degradation in terms of safety. A key challenge is that the accumulated number of observations in in-vitro drug responses varies greatly among different cancer types, where some tissues are more understudied than the others. Thus, we aim to develop a drug synergy prediction model for understudied tissues as a way of overcoming data scarcity problems. MATERIALS AND METHODS: We collected a comprehensive set of genetic, molecular, phenotypic features for cancer cell lines. We developed a drug synergy prediction model based on multitask deep neural networks to integrate multimodal input and multiple output. We also utilized transfer learning from data-rich tissues to data-poor tissues. RESULTS: We showed improved accuracy in predicting synergy in both data-rich tissues and understudied tissues. In data-rich tissue, the prediction model accuracy was 0.9577 AUROC for binarized classification task and 174.3 mean squared error for regression task. We observed that an adequate transfer learning strategy significantly increases accuracy in the understudied tissues. CONCLUSIONS: Our synergy prediction model can be used to rank synergistic drug combinations in understudied tissues and thus help to prioritize future in-vitro experiments. Code is available at https://github.com/yejinjkim/synergy-transfer.
Authors: Marco P Licciardello; Anna Ringler; Patrick Markt; Freya Klepsch; Charles-Hugues Lardeau; Sara Sdelci; Erika Schirghuber; André C Müller; Michael Caldera; Anja Wagner; Rebecca Herzog; Thomas Penz; Michael Schuster; Bernd Boidol; Gerhard Dürnberger; Yasin Folkvaljon; Pär Stattin; Vladimir Ivanov; Jacques Colinge; Christoph Bock; Klaus Kratochwill; Jörg Menche; Keiryn L Bennett; Stefan Kubicek Journal: Nat Chem Biol Date: 2017-05-22 Impact factor: 15.040
Authors: James D Phelan; Ryan M Young; Daniel E Webster; Sandrine Roulland; George W Wright; Monica Kasbekar; Arthur L Shaffer; Michele Ceribelli; James Q Wang; Roland Schmitz; Masao Nakagawa; Emmanuel Bachy; Da Wei Huang; Yanlong Ji; Lu Chen; Yandan Yang; Hong Zhao; Xin Yu; Weihong Xu; Maryknoll M Palisoc; Racquel R Valadez; Theresa Davies-Hill; Wyndham H Wilson; Wing C Chan; Elaine S Jaffe; Randy D Gascoyne; Elias Campo; Andreas Rosenwald; German Ott; Jan Delabie; Lisa M Rimsza; Fausto J Rodriguez; Fayez Estephan; Matthias Holdhoff; Michael J Kruhlak; Stephen M Hewitt; Craig J Thomas; Stefania Pittaluga; Thomas Oellerich; Louis M Staudt Journal: Nature Date: 2018-06-20 Impact factor: 49.962
Authors: John G Tate; Sally Bamford; Harry C Jubb; Zbyslaw Sondka; David M Beare; Nidhi Bindal; Harry Boutselakis; Charlotte G Cole; Celestino Creatore; Elisabeth Dawson; Peter Fish; Bhavana Harsha; Charlie Hathaway; Steve C Jupe; Chai Yin Kok; Kate Noble; Laura Ponting; Christopher C Ramshaw; Claire E Rye; Helen E Speedy; Ray Stefancsik; Sam L Thompson; Shicai Wang; Sari Ward; Peter J Campbell; Simon A Forbes Journal: Nucleic Acids Res Date: 2019-01-08 Impact factor: 16.971