Tahmid Mehdi1,2, Swneke D Bailey1, Paul Guilhamon1, Mathieu Lupien1,3,4. 1. Princess Margaret Cancer Centre, University Health Network, Toronto, ON, Canada. 2. Department of Statistics and Actuarial Science, University of Waterloo, Waterloo, ON, Canada. 3. Department of Medical Biophysics, University of Toronto, Toronto, ON, Canada. 4. Ontario Institute for Cancer Research, Toronto, ON, Canada.
Abstract
MOTIVATION: The 3D genome architecture influences the regulation of genes by facilitating chromatin interactions between distal cis-regulatory elements and gene promoters. We implement Cross Cell-type Correlation based on DNA accessibility (C3D), a customizable computational tool that predicts chromatin interactions using an unsupervised algorithm that utilizes correlations in chromatin measurements, such as DNaseI hypersensitivity signals. RESULTS: C3D accurately predicts 32.7%, 18.3% and 24.1% of interactions, validated by ChIA-PET assays, between promoters and distal regions that overlie DNaseI hypersensitive sites in K562, MCF-7 and GM12878 cells, respectively. AVAILABILITY AND IMPLEMENTATION: Source code is open-source and freely available on GitHub (https://github.com/LupienLabOrganization/C3D) under the GNU GPLv3 license. C3D is implemented in Bash and R; it runs on any platform with Bash (≥4.0), R (≥3.1.1) and BEDTools (≥2.19.0). It requires the following R packages: GenomicRanges, Sushi, data.table, preprocessCore and dynamicTreeCut. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.
MOTIVATION: The 3D genome architecture influences the regulation of genes by facilitating chromatin interactions between distal cis-regulatory elements and gene promoters. We implement Cross Cell-type Correlation based on DNA accessibility (C3D), a customizable computational tool that predicts chromatin interactions using an unsupervised algorithm that utilizes correlations in chromatin measurements, such as DNaseI hypersensitivity signals. RESULTS:C3D accurately predicts 32.7%, 18.3% and 24.1% of interactions, validated by ChIA-PET assays, between promoters and distal regions that overlie DNaseI hypersensitive sites in K562, MCF-7 and GM12878 cells, respectively. AVAILABILITY AND IMPLEMENTATION: Source code is open-source and freely available on GitHub (https://github.com/LupienLabOrganization/C3D) under the GNU GPLv3 license. C3D is implemented in Bash and R; it runs on any platform with Bash (≥4.0), R (≥3.1.1) and BEDTools (≥2.19.0). It requires the following R packages: GenomicRanges, Sushi, data.table, preprocessCore and dynamicTreeCut. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.
Authors: Michael J Johnston; Ana Nikolic; Nicoletta Ninkovic; Paul Guilhamon; Florence M G Cavalli; Steven Seaman; Franz J Zemp; John Lee; Aly Abdelkareem; Katrina Ellestad; Alex Murison; Michelle M Kushida; Fiona J Coutinho; Yussanne Ma; Andrew J Mungall; Richard Moore; Marco A Marra; Michael D Taylor; Peter B Dirks; Trevor J Pugh; Sorana Morrissy; Bradley St Croix; Douglas J Mahoney; Mathieu Lupien; Marco Gallo Journal: Genome Res Date: 2019-06-27 Impact factor: 9.043