Dataset Information

Discovering epistatic feature interactions from neural network models of regulatory DNA sequences.

ABSTRACT: Motivation:Transcription factors bind regulatory DNA sequences in a combinatorial manner to modulate gene expression. Deep neural networks (DNNs) can learn the cis-regulatory grammars encoded in regulatory DNA sequences associated with transcription factor binding and chromatin accessibility. Several feature attribution methods have been developed for estimating the predictive importance of individual features (nucleotides or motifs) in any input DNA sequence to its associated output prediction from a DNN model. However, these methods do not reveal higher-order feature interactions encoded by the models. Results:We present a new method called Deep Feature Interaction Maps (DFIM) to efficiently estimate interactions between all pairs of features in any input DNA sequence. DFIM accurately identifies ground truth motif interactions embedded in simulated regulatory DNA sequences. DFIM identifies synergistic interactions between GATA1 and TAL1 motifs from in vivo TF binding models. DFIM reveals epistatic interactions involving nucleotides flanking the core motif of the Cbf1 TF in yeast from in vitro TF binding models. We also apply DFIM to regulatory sequence models of in vivo chromatin accessibility to reveal interactions between regulatory genetic variants and proximal motifs of target TFs as validated by TF binding quantitative trait loci. Our approach makes significant strides in improving the interpretability of deep learning models for genomics. Availability and implementation:Code is available at: https://github.com/kundajelab/dfim. Supplementary information:Supplementary data are available at Bioinformatics online.

SUBMITTER: Greenside P

PROVIDER: S-EPMC6129272 | biostudies-literature | 2018 Sep

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Discovering epistatic feature interactions from neural network models of regulatory DNA sequences.

Greenside Peyton P Shimko Tyler T Fordyce Polly P Kundaje Anshul A

Bioinformatics (Oxford, England) 20180901 17

<h4>Motivation</h4>Transcription factors bind regulatory DNA sequences in a combinatorial manner to modulate gene expression. Deep neural networks (DNNs) can learn the cis-regulatory grammars encoded in regulatory DNA sequences associated with transcription factor binding and chromatin accessibility. Several feature attribution methods have been developed for estimating the predictive importance of individual features (nucleotides or motifs) in any input DNA sequence to its associated output pre ...[more]

PMID: 30423062

Similar Datasets

Project description:Computational methods for discovery of sequence elements that are enriched in a target set compared with a background set are fundamental in molecular biology research. One example is the discovery of transcription factor binding motifs that are inferred from ChIP-chip (chromatin immuno-precipitation on a microarray) measurements. Several major challenges in sequence motif discovery still require consideration: (i) the need for a principled approach to partitioning the data into target and background sets; (ii) the lack of rigorous models and of an exact p-value for measuring motif enrichment; (iii) the need for an appropriate framework for accounting for motif multiplicity; (iv) the tendency, in many of the existing methods, to report presumably significant motifs even when applied to randomly generated data. In this paper we present a statistical framework for discovering enriched sequence elements in ranked lists that resolves these four issues. We demonstrate the implementation of this framework in a software application, termed DRIM (discovery of rank imbalanced motifs), which identifies sequence motifs in lists of ranked DNA sequences. We applied DRIM to ChIP-chip and CpG methylation data and obtained the following results. (i) Identification of 50 novel putative transcription factor (TF) binding sites in yeast ChIP-chip data. The biological function of some of them was further investigated to gain new insights on transcription regulation networks in yeast. For example, our discoveries enable the elucidation of the network of the TF ARO80. Another finding concerns a systematic TF binding enhancement to sequences containing CA repeats. (ii) Discovery of novel motifs in human cancer CpG methylation data. Remarkably, most of these motifs are similar to DNA sequence elements bound by the Polycomb complex that promotes histone methylation. Our findings thus support a model in which histone methylation and CpG methylation are mechanistically linked. Overall, we demonstrate that the statistical framework embodied in the DRIM software tool is highly effective for identifying regulatory sequence elements in a variety of applications ranging from expression and ChIP-chip to CpG methylation data. DRIM is publicly available at http://bioinfo.cs.technion.ac.il/drim.

Project description:BackgroundVisualization tools for deep learning models typically focus on discovering key input features without considering how such low level features are combined in intermediate layers to make decisions. Moreover, many of these methods examine a network's response to specific input examples that may be insufficient to reveal the complexity of model decision making.ResultsWe present DeepResolve, an analysis framework for deep convolutional models of genome function that visualizes how input features contribute individually and combinatorially to network decisions. Unlike other methods, DeepResolve does not depend upon the analysis of a predefined set of inputs. Rather, it uses gradient ascent to stochastically explore intermediate feature maps to 1) discover important features, 2) visualize their contribution and interaction patterns, and 3) analyze feature sharing across tasks that suggests shared biological mechanism. We demonstrate the visualization of decision making using our proposed method on deep neural networks trained on both experimental and synthetic data. DeepResolve is competitive with existing visualization tools in discovering key sequence features, and identifies certain negative features and non-additive feature interactions that are not easily observed with existing tools. It also recovers similarities between poorly correlated classes which are not observed by traditional methods. DeepResolve reveals that DeepSEA's learned decision structure is shared across genome annotations including histone marks, DNase hypersensitivity, and transcription factor binding. We identify groups of TFs that suggest known shared biological mechanism, and recover correlation between DNA hypersensitivities and TF/Chromatin marks.ConclusionsDeepResolve is capable of visualizing complex feature contribution patterns and feature interactions that contribute to decision making in genomic deep convolutional networks. It also recovers feature sharing and class similarities which suggest interesting biological mechanisms. DeepResolve is compatible with existing visualization tools and provides complementary insights.

Dataset Information

Discovering epistatic feature interactions from neural network models of regulatory DNA sequences.

Publications

Discovering epistatic feature interactions from neural network models of regulatory DNA sequences.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets