Last modified 2 years ago Last modified on 10/28/15 14:24:46

SIC-ChIP (Shape Index Clustering for ChIP-seq peaks)

In the last years, Chromatin immunoprecipitation followed by sequencing (ChIP-seq) experiments have been widely used to investigate protein-DNA interactions. At present, many algorithms and pipelines are available for downstream processing of ChIP-seq experiments. All these methods are usually based on the evaluation of signal intensities to detect local enrichment of uniquely aligned reads on the reference genome (we refer to them as ‘ChIP-seq peaks’). Other features of peak shape are almost always ignored by all the available analysis techniques.

Here, we propose a novel analysis pipeline able to distinguish different shapes in a set of ChIP-seq peaks. In our analysis method, we consider the peaks of a single ChIP-seq at a time and we use cluster analysis to evaluate whether they can be divided into groups, according to both the shape and the intensity of the coverage function that defines them. To achieve this goal we select five shape indices, embedding the problem into the framework of multivariate statistical analysis.

SIC-ChIP (Shape Index Clustering for ChIP-seq peaks) implements the core of our methodology. Given the ChIP-seq coverage function (BigWig file) and the list of peaks (BED file), SIC-ChIP defines the peak curves, compute the five indices of shape and uses k-mean algorithm to cluster the peaks, testing different numbers of clusters.


Detailed information about SIC-ChIP and the complete analysis pipeline can be found in our paper ​HERE

If you want to cite us, please use DOI: 10.1186/s12859-015-0787-6 Cremona et al.: Peak shape clustering reveals biological insights. BMC Bioinformatics.2015, 16:349

Software Download

You can download SIC-ChIP from HERE.


SIC-ChIP documentation can be found HERE.