Accurate imputation of histone modifications using transcription release_tp3wwhbj5vcdzmukwlwv6vzt3u

by Zhong Wang, Alexandra G Chivu, Lauren A Choate, Edward J Rice, Donald C Miller, Tinyi Chu, Shao-Pei Chou, Nicole B Kingsley, Jessica L Peterson, Carrie J Finno, Rebecca R Bellone, Douglas F Antczak (+1 others)

Released as a post by Cold Spring Harbor Laboratory.

2020  

Abstract

Transcription is correlated with a variety of post-translational modifications to core histones, yet we still know relatively little about the precise and causal nature of their relationship. In particular, whether histone modifications serve a regulatory role, for example by marking functional elements for later activation, remains debated. Here we trained a support vector regression to infer the genomic distribution of 11 widely studied histone modifications using nascent transcription as input. Transcription captured the variation in active histone modifications, including punctate promoter and enhancer marks (e.g., H3K4me3, H3K4me1, and H3K27ac) and elongation marks (e.g., H3K36me3), down to single-nucleosome resolution at an accuracy that rivaled the correspondence between independent ChIP-seq experiments. The relationship between active histone modifications and transcription was conserved in all cell types examined. Surprisingly, transcription also accurately predicted the repressive chromatin mark H3K27me3. However, unlike active marks, H3K27me3 was found in two distinct patterns in different cell types: one pattern in fully differentiated cells covered broad regions with low levels of transcription, and a second pattern in stem cells was associated with transcription start sites of weakly transcribed genes. Transcription accurately identified bivalent promoters in embryonic stem cells, as well as other complex chromatin states identified by chromHMM. Although transcription accurately predicted nearly all histone modifications, we found a subset of DNase-I hypersensitive sites that were refractory to prediction, precluding models in which transcription initiates pervasively as a consequence of open chromatin. Our results demonstrate that histone modifications are tightly correlated with transcription, and provides a new tool for genome annotation using a single experiment.
In application/xml+jats format

Archived Files and Locations

application/pdf  3.3 MB
file_nceijhsnlfhpvpwf75ww7bm6am
www.biorxiv.org (repository)
web.archive.org (webarchive)
application/pdf  3.3 MB
file_s2iabavzcvfepnuw5loaxd5zfu
www.biorxiv.org (repository)
web.archive.org (webarchive)
application/pdf  3.3 MB
file_i52ax2vczzdnxe6tjqray7ncpu
www.biorxiv.org (repository)
web.archive.org (webarchive)
Read Archived PDF
Preserved and Accessible
Type  post
Stage   unknown
Date   2020-04-09
Work Entity
access all versions, variants, and formats of this works (eg, pre-prints)
Catalog Record
Revision: 124f356f-24b0-4f9a-8f77-42bdc7c0b8be
API URL: JSON