The main focus of the group is the role of chromatin structure and epigenetic regulation during normal development and in disease. We are also interested in evolutionary aspects of chromatin organization. We would like to be able to translate our knowledge to clinical applications and we work towards creating novel computational methods for handling, mining, and interpretation of 'big' epigenetic data.

Chromatin has two major biological roles: (i) it compacts genomic DNA to allow this 2-meter long molecule to fit into a cell nucleus and (ii) it facilitates regulation of the genome during DNA transcription, replication, and repair. We study both these aspects of chromatin structure. Our major research directions are described in more details below.

Genomic DNA accessibility: a versatile metric of chromatin state for biomedical research

In eukaryotes, including humans, read-out of genetic information occurs in the context of chromatin, which regulates accessibility of DNA for transcriptional machinery. However, the mechanisms of this regulation remain poorly understood, partly because of the limitations of the existing approaches measuring chromatin accessibility. Recently, in collaboration with the Kingston lab (MGH), we have developed a novel MNase-based methodology -- MACC -- that overcomes many of these limitations by combining information from several chromatin digestions of increasing depths. This methodology allows interrogating the regions of both open and closed chromatin in a balanced way, on the whole genome-scale. Furthermore, MACC profiles both chromatin accessibility and nucleosome occupancy in the same assay, and, when supplemented with a histone enrichment step, it can distinguish DNA protection by nucleosomes from that associated with binding of non-histone factors. This approach greatly facilitates sample-to-sample comparison and it was already successfully applied for chromatin accessibility profiling in different species from fly to human.

The pipeline for processing MACC data was implemented as a freely available R package.
MNase accessibility (MACC) workflow

Currently we pursue several collaborative projects where this approach is used to investigate chromatin accessibility regulation during cell fate determination and reprogramming.
We continue developing this methodology further to integrate it with other epigenetic data types as well as with the data on 3D genome organization. We also use machine-learning algorithms to generate predictive models of chromatin structure based on the accessibility data, to facilitate efficient epigenome profiling in large-scale projects.

Chromatin-associated proteins: regulators of genome and emerging targets in cancer research

Chromatin structure remodeling, required for regulation of genome function, involves changes in the composition and positioning of the DNA-associated proteins. We aim to understand molecular mechanisms and pathways of such remodeling. Furthermore, chromatin-associated proteins are often mutated in cancers, and despite their biological and biomedical significance the regulation and evolution of the genes encoding these proteins are still poorly understood. We use systems biology approach to integrate the knowledge on gene expression, chromatin structure, and DNA mutability to fill this gap.
   For instance, a special class of chromatin-associated proteins, histones, organize the genome through nucleosome placement. They are subject to covalent modifications during chromatin remodeling and are represented as ‘canonical’ histones and ‘non-canonical’, replication independent variants in the human genome. Non-canonical histone variants are essential for normal development in mammals, and one of them H3.3 has recently been reported to be mutated in pediatric brain and bone cancers. Our integrative analysis helped to discover differential regulation of the genes encoding this histone variant and shed new light on the possible role of the mutations in each of the H3.3 genes in carcinogenesis.

SWI/SNF inactivation results in altered transcription profile with net up-regulation effect, promoting cell proliferation

Differential regulation of H3.3-encoding genes

'transcriptional dosage' of the individual H3.3 genes
varies in different cell types
We now expanded this analysis on other histones and chromatin remodeling enzymes, focusing on those that have biomedical significance.