3CPET: Finding Co-factor Complexes in Chia-PET experiment using a Hierarchical Dirichlet Process

3CPET is method based on a non-parametric hierarchical Dirichlet process, to infer the set of protein networks involved in maintaining ChIA-PET chromatin interactions (or any other targeted chromatin interaction data). 3CPET is based on the observation that a protein can be a backbone element in some chromatin loops and dispensable in others. By combining ChIA-PET data with transcription factors binding sites and protein interactions information, 3CPET tries to infer the set of the most probable cofactor complexes involved in maintaining the different spatial interaction contexts.

The rationale behind 3CPET is that transcription factors can use a distinct combination of coactivators, depending on the genomic and spatial context. We refer to each possible combination of cofactors and their interactions as Chromatin maintainer Network (CMN).

To infer these sets of chromatin maintainer networks (CMNs), first, we build, for each DNA-DNA interaction, a protein-protein interaction (PPI) network connecting the two interacting DNA regions with the help of TF binding site information. Then, we use this set of networks to infer the most enriched coactivator networks. We use a Hierarchical Dirichlet process algorithm to (i) find the number of CMNs and (ii) the proteins involved in each one of them.

R3CPET can be downloaded from bioconductor as follow:

## try http if https is not available
if (!requireNamespace("BiocManager", quietly=TRUE))
    install.packages("BiocManager")
BiocManager::install("R3CPET")

To install the latest development version of R3CPET:

install.packages("devtools")
library("devtools")
install_github("R3CPET","sirusb")

For issues, you can post them in this git page. For questions, it is better to post them on [Bioconductor support] (https://support.bioconductor.org/) using the R3CPET tag.

Mohamed Nadhir Djekidel : nde12 at mails dot tsinghua dot edu dot cn or djek.nad at gmail dot com