Improving Chromatin-Interaction Prediction Using Single-Cell Open-Chromatin Profiles and Making Insight Into the Cis-Regulatory Landscape of the Human Brain.

Neetesh Pandey,Omkar Chandra Omkar Chandra,Shreya Mishra,Vibhor Kumar

doi:10.3389/fgene.2021.738194

Neetesh Pandey, Omkar Chandra Omkar Chandra + Show 2 more

Open Access

https://doi.org/10.3389/fgene.2021.738194

Copy DOI

Abstract

Single-cell open-chromatin profiles have the potential to reveal the pattern of chromatin-interaction in a cell type. However, currently available cis-regulatory network prediction methods using single-cell open-chromatin profiles focus more on local chromatin interactions despite the fact that long-range interactions among genomic sites play a significant role in gene regulation. Here, we propose a method that predicts both short and long-range interactions among genomic sites using single-cell open chromatin profiles. Our method, termed as single-cell epigenome based chromatin-interaction analysis (scEChIA) exploits signal imputation and refined L1 regularization. For a few single-cell open-chromatin profiles, scEChIA outperformed other tools even in terms of accuracy of prediction. Using scEChIA, we predicted almost 0.7 million interactions among genomic sites across seven cell types in the human brain. Further analysis revealed cell type for connection between genes and expression quantitative trait locus (eQTL) in the human brain and making insight about target genes of human-accelerated-elements and disease-associated mutations. Our analysis enabled by scEChIA also hints about the possible action of a few transcription factors (TFs), especially through long-range interaction in brain endothelial cells.

Highlights

Spatial interactions between different genomic loci are required for multiple regulatory functions
It is known that CCCTCbinding factor (CTCF) mediated chromatin-interaction and looping are mostly conserved and have a major impact on chromatin architecture
In addition to using sensitive L1 normalization, single-cell epigenome based chromatin-interaction analysis (scEChIA) uses its inbuilt function for matrix factorization to reduce noise in the read-count matrix to further improve the accuracy of prediction of chromatin-interaction

Summary

Introduction

Spatial interactions between different genomic loci are required for multiple regulatory functions (de Wit and de Laat, 2012). Many groups have profiled chromatin-interaction in multiple cell types using different experimental high-throughput methods to study such complex patterns in chromatin architecture and gene regulation. The experimental methods based on chromosome conformation capture (3C) are more focused on local genomic loci (Dekker et al, 2002). The Chromatin-interaction Analysis by Paired-End Tag Sequencing (ChIA-PET) method captures distal interactions, but it is limited to only binding sites of the protein of interest (Tang et al, 2015). The high-throughput chromosome conformation capture (HiC) assay provides a genomewide chromatin-interaction profile but requires deep-sequencing to achieve high resolution (de Wit and de Laat, 2012).

Methods

Results

Conclusion