HiCloud: A Pipeline for a More Inclusive Way to Analyze Hi-C Data

Attention for studying chromatin structure is rising due to its implications in gene regulation. Erez Lieberman-Aiden, et al. (2009) described Hi-C, a genome-wide approach that allows to identify any contact between a pair of loci by coupling proximity-based ligation (cross-linking with formaldehyde and chromatin digestion) with sequencing. Bioinformatics approaches to analyze Hi-C data consists of: 1.1) removing chimeric parts of reads followed by mapping or 1.2) iterative mapping where the unmapped reads, whose ends are trimmed 5-10nt each round, are realigned; 2) filtering read-pairs that are not close to any cleavage site from the restriction enzyme used at the chromatin digestion step; 3) filtering out possible artifacts and duplicates; 4) contact map creation, which is a square matrix of non-overlapping bins across the genome having each cell the number of pairs connecting 2 bins; 5) contact map normalization; 6) identification of Topologically Associating Domains (TADs); 7) visual representation. Current tools for analysing Hi-C data need to be easier-to-use for researchers to explore the data. Besides, they fail at the time to perform visual or statistical analysis of two or more contact maps. We are designing a fast and easy-to-use tool that will allow -using web technologies- a more interactive way to explore Hi-C results. Compared with HiCUP, which is one of the most used pipelines to analyze Hi-C data, our pipeline uses iterative mapping with STAR instead of reads preprocessing yielding faster mapping and higher efficiency.