benGRN 1.3.1

benchmarking gene regulatory networks

bengrn

Benchmark your gene regulatory networks inference algorithm (from scRNAseq or bulk RNAseq dataset) with BenGRN

The package is supposed to work with GRnnData and only uses biological ground truth datasets.

It can run Genie3 & pyscenic on your data as a comparison

It has 3 main different types of key ground truth data to compare your GRN to:

• Mc Calla et al.'s ChIP+Perturb ground truth
• omnipath's literature curated ground truth
• genome wide perturb seq 's dataset

You can find the documentation here

Install it from PyPI

pip install bengrn

Install it locally and run the notebooks:

git clone https://github.com/jkobject/benGRN.git
pip install -e benGRN

Usage

from bengrn import BenGRN
from bengrn import some_test_function

# a GRN in grnndata formart
grndata

BenGRN(grndata).do_tests()
#or
some_test_function(grndata)

see the notebooks in docs:

1. omnipath
2. genome wide perturb seq
3. Mc Calla

/!\ offline mode

If you want to run the notebooks offline, you need to download the data first.

from bengrn import download_perturb_gt, download_GT_db, download_sroy_gt

download_perturb_gt()

Development

Read the CONTRIBUTING.md file.

Awesome Benchmark of Gene Regulatory Networks created by @jkobject