LAUSR

Specific combinations of Histone Modifications (HMs) contributing towards histone code hypothesis lead to various biological functions. HMs combinations have been utilized by various studies to divide the genome into different regions. These study regions have been classified as chromatin states. Mostly Hidden Markov Model (HMM) based techniques have been utilized for this purpose. In case of chromatin studies, data from Next Generation Sequencing (NGS) platforms is being used. Chromatin states based on histone modification combinatorics are annotated by mapping them to functional regions of the genome. The number of states being predicted so far by the HMM tools have been justified biologically till now. The present study aimed at providing a computational scheme to identify the underlying hidden states in the data under consideration. Methods: We proposed a computational scheme HCVS based on hierarchical clustering and visualization strategy in order to achieve the objective of study. We tested our proposed scheme on a real data set of nine cell types comprising of nine chromatin marks. The approach successfully identified the state numbers for various possibilities. The results have been compared with one of the existing models as well which showed quite good correlation. The HCVS model not only helps in deciding the optimal state numbers for a particular data but it also justifies the results biologically thereby correlating the computational and biological aspects.