Abstract
Deep neural networks (DNNs) have achieved state-of-the-art performance in identifying gene regulatory sequences, but they have provided limited insight into the biology of regulatory elements due to the difficulty of interpreting the complex features they learn. Several models of how combinatorial binding of transcription factors, i.e. the regulatory grammar, drives enhancer activity have been proposed, ranging from the flexible TF billboard model to the stringent enhanceosome model. However, there is limited knowledge of the prevalence of these (or other) sequence architectures across enhancers. Here we perform several hypothesis-driven analyses to explore the ability of DNNs to learn the regulatory grammar of enhancers. We created synthetic datasets based on existing hypotheses about combinatorial transcription factor binding site (TFBS) patterns, including homotypic clusters, heterotypic clusters, and enhanceosomes, from real TF binding motifs from diverse TF families. We then trained deep residual neural networks (ResNets) to model the sequences under a range of scenarios that reflect real-world multi-label regulatory sequence prediction tasks. We developed a gradient-based unsupervised clustering method to extract the patterns learned by the ResNet models. We demonstrated that simulated regulatory grammars are best learned in the penultimate layer of the ResNets, and the proposed method can accurately retrieve the regulatory grammar even when there is heterogeneity in the enhancer categories and a large fraction of TFBS outside of the regulatory grammar. However, we also identify common scenarios where ResNets fail to learn simulated regulatory grammars. Finally, we applied the proposed method to mouse developmental enhancers and were able to identify the components of a known heterotypic TF cluster. Our results provide a framework for interpreting the regulatory rules learned by ResNets, and they demonstrate that the ability and efficiency of ResNets in learning the regulatory grammar depends on the nature of the prediction task.
Highlights
Enhancers are genomic regions distal to promoters that regulate the dynamic spatiotemporal patterns of gene expression required for the proper differentiation and development of multicellular organisms [1,2,3]
Gene regulatory sequences function through the combinatorial binding of transcription factors (TFs)
We simulated regulatory sequences based on existing hypotheses about the structure of possible regulatory grammars and trained Deep neural networks (DNNs) to model these sequences under a range of scenarios that reflect real-world regulatory sequence prediction tasks
Summary
Enhancers are genomic regions distal to promoters that regulate the dynamic spatiotemporal patterns of gene expression required for the proper differentiation and development of multicellular organisms [1,2,3]. Many additional features have been suggested to play a role in determining in vivo TF binding, such as heterogeneity of a TF’s binding motif [11], local DNA properties [12], broader sequence context and interposition dependence [13], cooperative binding of the TF with its partners [14,15,16,17], and condition-specific chromatin context [15, 18, 19] While both genomic and epigenomic features are important in determining the in vivo occupancy of a TF, recent studies have suggested that the epigenome can be accurately predicted from genomic context [12, 20,21,22], supporting the fundamental role of sequence in dictating the binding of TFs [23,24,25,26,27]. It is critical to understand the sequence patterns underlying enhancer regulatory functions and build sufficiently sophisticated models of enhancer sequence architecture
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
