Title: Integrating regulatory DNA sequence and gene expression to predict genome-wide chromatin accessibility across cellular contexts
Abstract Motivation Genome-wide profiles of chromatin accessibility and gene expression in diverse cellular contexts are critical to decipher the dynamics of transcriptional regulation. Recently, convolutional neural networks have been used to learn predictive cis-regulatory DNA sequence models of context-specific chromatin accessibility landscapes. However, these context-specific regulatory sequence models cannot generalize predictions across cell types. Results We introduce multi-modal, residual neural network architectures that integrate cis-regulatory sequence and context-specific expression of trans-regulators to predict genome-wide chromatin accessibility profiles across cellular contexts. We show that the average accessibility of a genomic region across training contexts can be a surprisingly powerful predictor. We leverage this feature and employ novel strategies for training models to enhance genome-wide prediction of shared and context-specific chromatin accessible sites across cell types. We interpret the models to reveal insights into cis- and trans-regulation of chromatin dynamics across 123 diverse cellular contexts. Availability and implementation The code is available at https://github.com/kundajelab/ChromDragoNN. Supplementary information Supplementary data are available at Bioinformatics online.  more » « less
Award ID(s):
1730158
PAR ID:
10171052
Author(s) / Creator(s):
; ; ;
Date Published:
Journal Name:
Bioinformatics
Volume:
35
Issue:
14
ISSN:
1367-4803
Page Range / eLocation ID:
i108 to i116
Format(s):
Medium: X
Sponsoring Org:
National Science Foundation
More Like this
  1. Gene expression and complex phenotypes are determined by the activity of cis-regulatory elements. However, an understanding of how extant genetic variants affect cis regulation remains limited. Here, we investigated the consequences of cis-regulatory diversity using single-cell genomics of more than 0.7 million nuclei across 172Zea mays(maize) inbreds. Our analyses pinpointed cis-regulatory elements distinct to domesticated maize and revealed how historical transposon activity has shaped the cis-regulatory landscape. Leveraging population genetics principles, we fine-mapped about 22,000 chromatin accessibility–associated genetic variants with widespread cell type–specific effects. Variants in TEOSINTE BRANCHED1/CYCLOIDEA/PROLIFERATING CELL FACTOR–binding sites were the most prevalent determinants of chromatin accessibility. Finally, integrating chromatin accessibility–associated variants, organismal trait variation, and population differentiation revealed how local adaptation has rewired regulatory networks in unique cellular contexts to alter maize flowering. 
    more » « less
  2. Abstract Gene duplication is a major source of evolutionary innovation, enabling the emergence of novel expression patterns and functions. Leveraging single-cell genomics, we investigated the transcriptional and cis-regulatory landscapes of duplicated genes in cultivated soybean (Glycine max), which has undergone 2 rounds of whole-genome duplication. Our analysis revealed extensive diversity of transcriptional profiles within and across tissues among duplicated gene pairs. Within-tissue divergence was largely attributable to genetic variation in their associated accessible chromatin regions (ACRs), where cis-regulatory elements reside, whereas cross-tissue divergence was more likely shaped by dynamics in ACR chromatin accessibility profiles across tissues. Distinct duplication mechanisms also likely give rise to different types of cis-regulatory variants, contributing variably to transcriptional divergence. By comparing ACRs associated with gene sets derived from 2 rounds of whole-genome duplication and sharing a common ancestral gene, we found that most ACRs retained one or multiple corresponding duplicated sequences in which mutations gradually accumulated over time, while a subset likely arose de novo. Finally, we traced the evolution of cell-type-specific expression and cell-type-specific ACRs within duplicated gene sets, illustrating a powerful framework for identifying candidate regulatory regions driving cell-type-specific expression. Collectively, our findings highlight the important role of cis-regulatory evolution in shaping transcriptional divergence in a spatiotemporal manner, uncovered with the resolution of single-cell genomics. 
    more » « less
  3. Purugganan, Michael (Ed.)
    Polyploidy is a prominent mechanism of plant speciation and adaptation, yet the mechanistic understandings of duplicated gene regulation remain elusive. Chromatin structure dynamics are suggested to govern gene regulatory control. Here, we characterized genome-wide nucleosome organization and chromatin accessibility in allotetraploid cotton, Gossypium hirsutum (AADD, 2n = 4X = 52), relative to its two diploid parents (AA or DD genome) and their synthetic diploid hybrid (AD), using DNS-seq. The larger A-genome exhibited wider average nucleosome spacing in diploids, and this intergenomic difference diminished in the allopolyploid but not hybrid. Allopolyploidization also exhibited increased accessibility at promoters genome-wide and synchronized cis-regulatory motifs between subgenomes. A prominent cis-acting control was inferred for chromatin dynamics and demonstrated by transposable element removal from promoters. Linking accessibility to gene expression patterns, we found distinct regulatory effects for hybridization and later allopolyploid stages, including nuanced establishment of homoeolog expression bias and expression level dominance. Histone gene expression and nucleosome organization are coordinated through chromatin accessibility. Our study demonstrates the capability to track high-resolution chromatin structure dynamics and reveals their role in the evolution of cis-regulatory landscapes and duplicate gene expression in polyploids, illuminating regulatory ties to subgenomic asymmetry and dominance. 
    more » « less
  4. Abstract Cis-regulatory elements (CREs) are essential for regulating gene expression, yet their evolutionary dynamics in plants remain elusive. Here we constructed a single-cell chromatin accessibility atlas forOryza sativafrom 103,911 nuclei representing 126 cell states across nine organs. Comparative genomics betweenO. sativaand 57,552 nuclei from four additional grass species (Zea mays,Sorghum bicolor,Panicum miliaceumandUrochloa fusca) revealed that chromatin accessibility conservation varies with cell-type specificity. Epidermal accessible chromatin regions in the leaf were less conserved compared to other cell types, indicating accelerated regulatory evolution in the L1-derived epidermal layer ofO. sativarelative to other species. Conserved accessible chromatin regions overlapping the repressive histone modification H3K27me3 were identified as potentially silencer-like CREs, as deleting these regions led to up-regulation of gene expression. This study provides a comprehensive epigenomic resource for the rice community, demonstrating the utility of a comparative genomics approach that highlights the dynamics of plant cell-type-specific CRE evolution. 
    more » « less
  5. Wittkopp, Patricia (Ed.)
    Abstract Chromatin accessibility plays an important role in shaping gene expression, yet little is known about the genetic and molecular mechanisms that influence the evolution of chromatin configuration. Both local (cis) and distant (trans) genetic influences can in principle influence chromatin accessibility and are based on distinct molecular mechanisms. We, therefore, sought to characterize the role that each of these plays in altering chromatin accessibility in 2 closely related sea urchin species. Using hybrids of Heliocidaris erythrogramma and Heliocidaris tuberculata, and adapting a statistical framework previously developed for the analysis of cis and trans influences on the transcriptome, we examined how these mechanisms shape the regulatory landscape at 3 important developmental stages, and compared our results to similar analyses of the transcriptome. We found extensive cis- and trans-based influences on evolutionary changes in chromatin, with cis effects generally larger in effect. Evolutionary changes in accessibility and gene expression are correlated, especially when expression has a local genetic basis. Maternal influences appear to have more of an effect on chromatin accessibility than on gene expression, persisting well past the maternal-to-zygotic transition. Chromatin accessibility near gene regulatory network genes appears to be distinctly regulated, with trans factors appearing to play an outsized role in the configuration of chromatin near these genes. Together, our results represent the first attempt to quantify cis and trans influences on evolutionary divergence in chromatin configuration in an outbred natural study system and suggest that chromatin regulation is more genetically complex than was previously appreciated. 
    more » « less