GOntact

GOntact is a tool for functional enrichment analysis of cis-regulatory elements using chromatin contact data. Unlike traditional proximity-based approaches, GOntact leverages promoter-capture Hi-C (PCHi-C) data to infer gene-enhancer relationships and derive Gene Ontology enrichments.

It implements the method described in Laverré et al., 2022 (bioRxiv), and uses PCHi-C data from Laverré et al., Genome Research, 2022 to provide biologically coherent functional annotations and novel insights into gene regulation across human and mouse genomes.

GOntact installation

GOntact usage

Installation

Installing GOntact from source

To install GOntact from source, first clone the GitLab repository using the following command:

git clone git@gitlab.in2p3.fr:anamaria.necsulea/GOntact.git

To compile GOntact, you will need the opam package manager for OCaml. On Debian-based Linux distributions, you can install it and initialize it using the following commands:

sudo apt install opam
opam init

You will also need the dune build system for OCaml.

opam install dune

Once opam and dune are installed, in the GOntact directory created before, install the required dependencies using the following commands:

opam pin add -n gontact .
opam install gontact --deps-only

opam pin add -n gontact-server .
opam install gontact-server --deps-only

You can now compile GOntact with the following command, run from the GOntact directory:

dune build

To use this version of GOntact, you can use the following commands (note the extra "--", which signifies that the command-line arguments that follow are for the gontact executable rather than for `dune):

dune exec gontact -- --help

All GOntact commands thereafter will start with dune exec gontact -- instead of gontact.

Installing the GOntact package directly using opam

GOntact can be installed via opam, using the following command:

opam pin add -y gontact https://gitlab.in2p3.fr/anamaria.necsulea/GOntact.git

This will install an executable named gontact in your .opam directory. This executable should be accessible from any directory, if your PATH variable was correctly configured when installing opam.

Usage

For the following usage commands, the GOntact executable is named gontact. Please refer to the above instructions if you installed GOntact directly from source.

To test GOntact, you can use this example dataset (make sure you don't have a data directory where you run this!):

wget http://pbil.univ-lyon1.fr/members/necsulea/GOntact/data.tar.gz
test ! -d data && tar -xzvf data.tar.gz

These commands will create a directory named data, containing several data types.

Data types

Chromatin contact data

Gene annotations

Gene Ontology annotations

Enhancer coordinates

Chromatin contact data

This dataset includes chromatin contacts detected with the PCHi-C approach, for human and mouse. This data was described in Laverré et al., Genome Research, 2022. Chromatin interactions were scored for several cell types for each species and are provided in the ibed format, in the data/PCHi-C subfolder.

Note that these chromatin contacts are provided with respect to the GRCh38 (hg38) genome assembly for human and to the GRCm38 (mm10) genome assembly for mouse.

Gene annotations

GOntact also needs a set of genomic annotations as an input. These should be provided in the GTF format. Example files for human and mouse are provided in the subfolder data/ensembl_annotations. They were downloaded from the Ensembl database and correspond to the GRCh38 (hg38) genome assembly for human and to the GRCm38 (mm10) genome assembly for mouse.

Gene Ontology annotation

Gene Ontology annotations are provided for human and mouse in the data/GeneOntology subfolder (goa_human.gaf for human and mgi.gaf for mouse). These files correspond to the 2025-07-22 release of the Gene Ontology database. They were downloaded from geneontology.org.

Enhancer coordinates

Example enhancer datasets are also provided in the subfolder data/enhancers.

Basic usage

Here is an example of a command line that runs GOntact in "GREAT" mode, using Vista midbrain enhancers as a foreground set and the full set of ENCODE enhancers as a background set:

mkdir GREAT_results_bp

gontact enrich \
--mode=GREAT \
--gene-annot=data/genomic_annotations/Homo_sapiens.GRCh38.115.gtf \
--functional-annot=data/GeneOntology/goa_human.gaf \
--ontology=data/GeneOntology/go-basic.obo \
--foreground=data/enhancers/human/VistaEnhancers_midbrain_hg38.bed \
--background=data/enhancers/human/ENCODE.Laverre2022.bed \
--upstream=5000 \
--downstream=1000 \
--extend=1000000 \
--output-dir=GREAT_results_bp \
--output-prefix=VistaEnhancers_midbrain \
--domain=biological_process 

Here is an example of a command line that runs GOntact in "contacts" mode, using Vista midbrain enhancers as a foreground set, the full set of ENCODE enhancers as a background set, and the set of PCHi-C interactions that are shared in at least two samples.

mkdir contacts_results_bp

gontact enrich \
--mode=contacts \
--gene-annot=data/ensembl_annotations/Homo_sapiens.GRCh38.115.gtf \
--functional-annot=data/GeneOntology/goa_human.gaf \
--ontology=data/GeneOntology/go-basic.obo \
--foreground=data/enhancers/human/VistaEnhancers_midbrain_hg38.bed \
--background=data/enhancers/human/ENCODE.Laverre2022.bed \
--ibed-path=data/PCHi-C/human/ibed_files/shared_contacts_min2samples.ibed \
--min-score=0 \
--bait-coords=data/PCHi-C/human/hg38.baitmap \
--output-dir=contacts_results_bp \
--output-prefix=VistaEnhancers_midbrain \
--domain=biological_process