Machine learning uncovers cell identity regulator by histone code

Bo Xia,Shu Meng,Yanqiang Li,Kaifu Chen,Lili Zhang,Jie Lv,John P Cooke,Alin S Tomoiaga,Dongyu Zhao,Xin Wang,Guangyu Wang,Qi Cao,Min Zhang

doi:10.1038/s41467-020-16539-4

Bo Xia, Shu Meng + Show 11 more

Open Access

https://doi.org/10.1038/s41467-020-16539-4

Copy DOI

Abstract

Conversion between cell types, e.g., by induced expression of master transcription factors, holds great promise for cellular therapy. Our ability to manipulate cell identity is constrained by incomplete information on cell identity genes (CIGs) and their expression regulation. Here, we develop CEFCIG, an artificial intelligent framework to uncover CIGs and further define their master regulators. On the basis of machine learning, CEFCIG reveals unique histone codes for transcriptional regulation of reported CIGs, and utilizes these codes to predict CIGs and their master regulators with high accuracy. Applying CEFCIG to 1,005 epigenetic profiles, our analysis uncovers the landscape of regulation network for identity genes in individual cell or tissue types. Together, this work provides insights into cell identity regulation, and delivers a powerful technique to facilitate regenerative medicine.

Highlights

Conversion between cell types, e.g., by induced expression of master transcription factors, holds great promise for cellular therapy
To build a solid foundation for study on cell identity genes (CIGs), we performed a thorough review of literature to develop Cell Identity Gene Data Base (CIGDB), the first database for manually curated known CIGs
Our meticulous review of literature motivated the use to define CIGs as belonging to at least one of four categories: (1) master transcription factors, which drive the differentiation towards a cell type when their expression is ectopically induced in another cell type; (2) required transcription factors, whose depletion impaired the differentiation towards a specific cell type; (3) genes required for key functions or phenotypes of a cell type; (4) genes that were widely used as markers for a cell type

Summary

Introduction

Conversion between cell types, e.g., by induced expression of master transcription factors, holds great promise for cellular therapy. Scientists define CIGs based on the expression specificity analysis[10], which requires comparison between a query cell type and most, if not all, other cell types It requires distinguishing between cell-type-specific and biologicalcondition-specific genes, e.g., heat shock genes. Super enhancers and a unique broad pattern of H3K4me[3] modification were found to regulate CIGs, whereas it is typical enhancer and sharp H3K4me[3] modification that regulate other expressed genes such as housekeeping genes[11,12,13,14] These discoveries suggest strong potential to systematically uncover CIGs on the basis of epigenetic signature analysis. We develop Computational Epigenetic Framework for Cell Identity Gene (CEFCIG), a computational framework to uncover CIGs and further define their master regulators. This work provides insights into cell identity regulation and delivers a powerful technique to facilitate regenerative medicine

Methods

Results

Conclusion