[vc_row el_class=”inner-body-content” css=”.vc_custom_1668061214649{padding-top: 30px !important;padding-bottom: 20px !important;}”][vc_column][vc_column_text css=”.vc_custom_1669962371199{margin-bottom: 0px !important;}”]
Title
Genome Wide Mapping of Chromatin States Based on Histone Combinatorics for Determination of Epigenetic Expression
Abstract
Histone proteins wrap DNA around in small globular entities commonly known as
nucleosomes. The post translational modifications to the histone tails are referred as
histone modifications (HMs). The regulation of DNA in order to access the transcription machinery is epigenetically programmed by specific DNA and chromatin covalent modifications. HMs could either be present or absent at particular genomic loci and the combinatorial patterns of the specific modifications being addressed as ‘histone codes’, are believed to co-regulate significant biological processes. Regions defined by combinatorial patterns of marks can be referred to as chromatin states. Chromatin states associated with genomic locations correlate with specific functional elements as enhancers, transcription start sites, which can be exclusively inferred from successive combinations of chromatin marks in their contiguous locations. Biologically significant combinatorics of epigenetic modifications and their subsequent functional interplay are still mostly unrevealed. We aimed to use ChIP-Seq data of Histone modifications at different genomic loci to highlight the unbiased genomic grouping and to decode the complex biological network of HMs in association with other chromatin players in defining various chromatin states
We used different tools and techniques to accomplish our task. Complex biological networks underlying the hidden chromatin states were revealed via merging machine learning and graph theory existing approaches. Histone modification efficient and simple combinatorics was studied at a global and local scale by developing and implementing a clustering and biclustering tool. Results have been compared with the existing approaches. Meanwhile a simple and efficient computational methodology for efficient chromatin states identification for ChromHMM (HMM based chromatin segmentation) has also been developed by utilizing Hidden Markov Models components.
As a result of above study we revealed the role of various factors in maintaining the chromatin state connectivity via focusing chromatin state networks. Our studies highlighted the minimum dominating nodes set and various hubs in chromatin state networks focusing their interaction patterns. Along with we developed and tested a clustering tool ChromClust and a biclustering tool ChromBiSim to highlight histone combinatorics in binarized signal data in an efficient and interactive way. ChromClust operates at global level while ChromBiSim mines local patterns of histone modifications associations.
We conclude that epigenomic landscape is portrayed as interplay of various factors including histone combinations, transcription factors, chromatin modifiers and most importantly the underlying DNA motifs. Each chromatin state has a specific set of these factors which interact with each other to mark that state hence creating the whole chromatin states network.
[/vc_column_text][/vc_column][/vc_row]