Basset: learning the regulatory code of the accessible genome with deep convolutional neural networks

Abstract

The complex language of eukaryotic gene expression remains incompletely understood. Despite the importance suggested by many noncoding variants statistically associated with human disease, nearly all such variants have unknown mechanisms. Here, we address this challenge using an approach based on a recent machine learning advance—deep convolutional neural networks (CNNs). We introduce the open source package Basset to apply CNNs to learn the functional activity of DNA sequences from genomics data. We trained Basset on a compendium of accessible genomic sites mapped in 164 cell types by DNase-seq, and demonstrate greater predictive accuracy than previous methods. Basset predictions for the change in accessibility between variant alleles were far greater for Genome-wide association study (GWAS) SNPs that are likely to be causal relative to nearby SNPs in linkage disequilibrium with them. With Basset, a researcher can perform a single sequencing assay in their cell type of interest and simultaneously learn that cell's chromatin accessibility code and annotate every mutation in the genome with its influence on present accessibility and latent potential for accessibility. Thus, Basset offers a powerful computational approach to annotate and interpret the noncoding genome.

The functions and unique features of long intergenic non-coding RNA

“…Orthogonal approaches to predict effects of non-coding variants using deep convolutional neural networks have been proposed 166 . Multiple methods to characterize lincRNA interactomes have been described; several are highlighted here.…”

Inference of cell type specific regulatory networks on mammalian lineages

“…SCM’s regression approach bypasses the need to call peaks and models synergistic relationships among nearby k -mers to directly predict the accessibility of a region. More recently, several deep learning methods [74, 75] aim to predict chromatin features including accessibility [76–78] (Table S1). A recent approach is Basset [78], which uses convolutional neural networks to learn context-specific sequence predictors of DNA accessibility.…”

Fast, scalable prediction of deleterious noncoding variants from functional and population genomic data

“…The development of statistical and machine-learning methods that attempt to address this integrative prediction challenge has emerged as an active, fast-moving area of research. Recently published methods in this area can be roughly divided intothree categories: (1) machine-learning classifiers that attempt to separate known disease variants from putatively benign variants using a variety of genomic features (e.g., GWAVA 13 and FATHMM-MKL 14 ); (2) sequence- and motif-based predictors for the impact of noncoding variants on cell-type-specific molecular phenotypes, such as chromatin accessibility or histone modifications (e.g., DeepBind 15 , DeepSEA 16 and Basset 17 ); and (3) evolutionary methods that consider data on genetic variation together with functional genomic data and aim to predict the effects of noncoding variants on fitness (e.g., CADD 18 , DANN 19 , FunSeq2 20 , and fitCons 3 ). A limitation of methods of the first class is that they depend strongly on the available training data, which may be limited and may not be representative of the broader class of regulatory sequences of interest.…”