Generating and designing DNA with deep generative models

Abstract

We propose generative neural network methods to generate DNA sequences and tune them to have desired properties. We present three approaches: creating synthetic DNA sequences using a generative adversarial network; a DNA-based variant of the activation maximization ("deep dream") design method; and a joint procedure which combines these two approaches together. We show that these tools capture important structures of the data and, when applied to designing probes for protein binding microarrays, allow us to generate new sequences whose properties are estimated to be superior to those found in the training data. We believe that these results open the door for applying deep generative models to advance genomics research.

NIPS 2017 Computational Biology Workshop

References in corpus (11)

• Auto-Encoding Variational Bayes
• Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling
• Conditional Image Generation with PixelCNN Decoders
• Understanding Neural Networks Through Deep Visualization
• Objective-Reinforced Generative Adversarial Networks (ORGAN) for Sequence Generation Models
• Learning to Generate Reviews and Discovering Sentiment
• Neural Machine Translation in Linear Time
• Variational Lossy Autoencoder
• Synthesizing the preferred inputs for neurons in neural networks via deep generator networks
• Maximum-Likelihood Augmented Discrete Generative Adversarial Networks
• Boundary-Seeking Generative Adversarial Networks

Cited by in corpus (20)

• Quantum generative adversarial networks
• Population-Based Black-Box Optimization for Biological Sequence Design
• Fast differentiable DNA and protein sequence optimization for molecular design
• Feedback GAN (FBGAN) for DNA: a Novel Feedback-Loop Architecture for Optimizing Protein Functions
• Conformal Prediction Under Feedback Covariate Shift for Biomolecular Design
• AdaLead: A simple and robust adaptive greedy search algorithm for sequence design
• Design by adaptive sampling
• Plug & Play Directed Evolution of Proteins with Gradient-based Discrete MCMC
• Quantum Generative Adversarial Networks For Anomaly Detection In High Energy Physics
• Deep Learning in Protein Structural Modeling and Design
• Autofocused oracles for model-based design
• A primer on model-guided exploration of fitness landscapes for biological sequence design
• Deep Generative Modeling-based Data Augmentation with Demonstration using the BFBT Benchmark Void Fraction Datasets
• Offline Model-Based Optimization via Normalized Maximum Likelihood Estimation
• Likelihood Ratios for Out-of-Distribution Detection
• Differential-Critic GAN: Generating What You Want by a Cue of Preferences
• Amortized Synthesis of Constrained Configurations Using a Differentiable Surrogate
• DP-CGAN: Differentially Private Synthetic Data and Label Generation
• A Hierarchical Approach to Scaling Batch Active Search Over Structured Data
• Generating Multi-type Temporal Sequences to Mitigate Class-imbalanced Problem