Dr.VAE: Drug Response Variational Autoencoder

Abstract

We present two deep generative models based on Variational Autoencoders to improve the accuracy of drug response prediction. Our models, Perturbation Variational Autoencoder and its semi-supervised extension, Drug Response Variational Autoencoder (Dr.VAE), learn latent representation of the underlying gene states before and after drug application that depend on: (i) drug-induced biological change of each gene and (ii) overall treatment response outcome. Our VAE-based models outperform the current published benchmarks in the field by anywhere from 3 to 11% AUROC and 2 to 30% AUPR. In addition, we found that better reconstruction accuracy does not necessarily lead to improvement in classification accuracy and that jointly trained models perform better than models that minimize reconstruction error independently.

submitted to NIPS 2017

References in corpus (4)

• Adam: A Method for Stochastic Optimization
• Stochastic Backpropagation and Approximate Inference in Deep Generative Models
• Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs)
• Improving Variational Inference with Inverse Autoregressive Flow

Cited by in corpus (9)

• Deep Learning for Genomics: A Concise Overview
• Dropout Feature Ranking for Deep Learning Models
• Evaluating deep variational autoencoders trained on pan-cancer gene expression
• Variational Autoencoder for Anti-Cancer Drug Response Prediction
• Accelerated Discovery of Sustainable Building Materials
• Deep learning in bioinformatics: introduction, application, and perspective in big data era
• retina-VAE: Variationally Decoding the Spectrum of Macular Disease
• Segment-Based Credit Scoring Using Latent Clusters in the Variational Autoencoder
• Learning a Generative Model of Cancer Metastasis