A Novel Method for Peak Calling of Chip-SEQ Data Based on Scale Space Filtering to Combat Against Random Variations
This paper proposes a novel algorithm for peak calling in chromatin immune precipitation sequencing (ChIP-Seq) data using scale-space-filtering. In this method, Gaussian convolution is applied to expand the given signal over a continuum of sizes. The generated scale space image is explored just like a tree by using its qualitative structure having concise but complete description of observation at all possible scales. It is of great practical importance and quite challenging to understand genome-wide DNA sequencing for the identification of transcription factor binding sites in large-scale ChIP-Seq data. This method explores not only enriched peak regions effectively but also low-density background regions as well. Computer simulations reveal that the proposed methodology has a great potential to produce good results by accurately estimating the average fragment length and threshold level. Keywords- Chromatin Immune Precipitation Sequencing, Gaussian Convolution, Scale-Space-Filtering, Transcription Factor Binding Sites.