CREWdb 1.0: Optimizing Chromatin Readers, Erasers, and Writers Database using Machine Learning-Based Approach

Abstract

Aberration in heterochromatin and euchromatin states contributes to various disease phenotypes. The transcriptional regulation between these two states is significantly governed by post-translational modifications made by three functional types of chromatin regulators: readers, writers, and erasers. Writers introduce a chemical modification to DNA and histone tails, readers bind the modification to histone tails using specialized domains, and erasers remove the modification introduced by writers. Altered regulation of these chromatin regulators results in complex diseases such as cancer, neurodevelopmental diseases, myocardial diseases, kidney diseases, and embryonic development. Due to the reversible nature of chromatin modifications, we can develop therapeutic approaches targeting these chromatin regulators. However, a limited number of chromatin regulators have been identified thus far, and a subset of them are ambiguously classified as multiple chromatin regulator functional types. Thus, we have developed machine learning-based approaches to predict and classify the functional roles of chromatin regulator proteins, thereby optimizing the accuracy of the first comprehensive database of chromatin regulators known as CREWdb. GitHub URLCREWdb source code is available at https://github.com/smollahlab/CREWdb Database URLCREWdb webtool is available at http://mollahlab.wustl.edu/crewdb