Motivation: Convolutional neural networks (CNN) have outperformed conventional methods in modeling the sequence specificity of DNA–protein binding. Yet inappropriate CNN architectures can yield poorer performance than simpler models. Thus an in-depth understanding of how to match CNN architecture to a given task is needed to fully harness the power of CNNs for computational biology applications.Results: We present a systematic exploration of CNN architectures for predicting DNA sequence binding using a large compendium of transcription factor datasets. We identify the best-performing architectures by varying CNN width, depth and pooling designs. We find that adding convolutional kernels to a network is important for motif-based tasks. We show the benefits of CNNs in learning rich higher-order sequence features, such as secondary motifs and local sequence context, by comparing network performance on multiple modeling tasks ranging in difficulty. We also demonstrate how careful construction of sequence benchmark datasets, using approaches that control potentially confounding effects like positional or motif strength bias, is critical in making fair comparisons between competing methods. We explore how to establish the sufficiency of training data for these learning tasks, and we have created a flexible cloud-based framework that permits the rapid exploration of alternative neural network architectures for problems in computational biology.Availability and Implementation: All the models analyzed are available at http://cnn.csail.mit.edu.Contact: gifford@mit.eduSupplementary information: Supplementary data are available at Bioinformatics online.

Human immunodeficiency virus-1 (HIV-1) is a retrovirus with a 10-kb single-stranded RNA genome. HIV-1 must express all of its gene products from the same primary transcript, which undergoes alternative splicing to produce diverse protein products, including structural proteins and regulatory factors 1 , 2 . Despite the critical role of alternative splicing, the mechanisms driving splice-site choice are poorly understood. Synonymous RNA mutations that lead to severe defects in splicing and viral replication indicate the presence of unknown cis-regulatory elements 3 . We use DMS-MaPseq to probe the structure of HIV-1 RNA in cells and develop an algorithm called D etection of R NA folding E nsembles using E xpectation- M aximization (DREEM), which reveals alternative conformations assumed by the same RNA sequence. Contrary to previous models, which analyzed population averages 4 , our results reveal the widespread heterogeneous nature of HIV-1 RNA structure. In addition to confirming that in vitro characterized alternative structures for the HIV-1 Rev Responsive Element (RRE) exist in cells, we discover alternative conformations at critical splice sites that influence the ratio of transcript isoforms. Our simultaneous measurement of splicing and intracellular RNA structure provides evidence for the long-standing hypothesis 5 – 7 that RNA conformation heterogeneity regulates splice site usage and viral gene expression.

Quantifying the effects of cis-regulatory DNA on gene expression is a major challenge. Here, we present the multiplexed editing regulatory assay (MERA); a high-throughput CRISPR/Cas9-based approach that analyzes the functional impact of the regulatory genome in its native context. MERA tiles thousands of mutations across ~40 kb of cis-regulatory genomic space and uses knock-in green fluorescent protein (GFP) reporters to read out gene activity. Using this approach, we obtain quantitative information on the contribution of cis-regulatory regions to gene expression. We identify proximal and distal regulatory elements necessary for expression of four embryonic stem cell-specific genes. We show a consistent contribution of neighboring gene promoters to gene expression and identify unmarked regulatory elements (UREs) that control gene expression but do not have typical enhancer epigenetic or chromatin features. We compare thousands of functional and non-functional genotypes at a genomic location and identify the basepair-resolution functional motifs of regulatory elements.

BackgroundModern genetics has been transformed by high-throughput sequencing. New experimental designs in model organisms involve analyzing many individuals, pooled and sequenced in groups for increased efficiency. However, the uncertainty from pooling and the challenge of noisy sequencing data demand advanced computational methods.ResultsWe present MULTIPOOL, a computational method for genetic mapping in model organism crosses that are analyzed by pooled genotyping. Unlike other methods for the analysis of pooled sequence data, we simultaneously consider information from all linked chromosomal markers when estimating the location of a causal variant. Our use of informative sequencing reads is formulated as a discrete dynamic Bayesian network, which we extend with a continuous approximation that allows for rapid inference without a dependence on the pool size. MULTIPOOL generalizes to include biological replicates and case-only or case-control designs for binary and quantitative traits.ConclusionsOur increased information sharing and principled inclusion of relevant error sources improve resolution and accuracy when compared to existing methods, localizing associations to single genes in several cases. MULTIPOOL is freely available at http://cgs.csail.mit.edu/multipool/.

ENCODE 3 (2012-2017) expanded production and added new types of assays 8 (Fig. 1, Extended Data Fig. 1), which revealed landscapes of RNA binding and the 3D organization of chromatin via methods such as chromatin interaction analysis by paired-end tagging (ChIA-PET) and Hi-C chromosome conformation capture. Phases 2 and 3 delivered 9,239 experiments (7,495 in human and 1,744 in mouse) in more than 500 cell types and tissues, including mapping of transcribed regions and transcript isoforms, regions of transcripts recognized by RNA-binding proteins, transcription factor binding regions, and regions that harbour specific histone modifications, open chromatin, and 3D chromatin interactions. The results of all of these experiments are available at the ENCODE portal (http://www.encodeproject.org). These efforts, combined with those of related projects and many other laboratories, have produced a greatly enhanced view of the human genome (Fig. 2), identifying 20,225 protein-coding and 37,595 noncoding genes