Surabhi Jagtap, Abdulkadir Çelikkanat, Aurélie Pirayre, Frédérique Bidard, Laurent Duval, Fragkiskos D Malliaros, BraneMF: integration of biological networks for functional analysis of proteins, Bioinformatics, Volume 38, Issue 24, 15 December 2022, Pages 5383–5389, https://doi.org/10.1093/bioinformatics/btac691
1. Create a virtual environment with name 'branemf'
conda create -n branemf python=3.6
2. Activate the virtual environment
conda activate branemf
3. Install the required packages
pip install -r requirements.txt
4. Download the precomputed PPMI matrix files
Google drive: https://drive.google.com/drive/folders/1X5Gj5udIPKiLEvzeKWuUucnWw5AZOsnY?usp=sharing
To compute PPMI matrices from String networks, run the following command by providing taxanomy ID of the organism (e.g 4932: yeast, 9606: human) other organisms: https://stringdb-static.org/download/species.v11.5.txt
python BraneMF_buildPPMI.py --O 4932
5. Computation of embeddings
run file: branemf.m
To reproduce the results, use precomuted embeddings and following commands
6. Perform clustering and GO enrichment
python cluster_enrichment.py --emb ./data/emb/yeast_branemf_w1_alpha_1.emb --k 40 --sim 20 --genes ./data/yeast_string_genes.txt
7. Perform protein function prediction
python Protein_function_prediction.py --emb ./emb/Integrated_emb/BraneMF_emb_sel/w10/Org4932BraneMF_g1_d128_w10_23May2022.emb --anno ./data/org_4932_BPl1.txt --trials 10
8. Perform protein Interaction prediction
a. Prepare the training and test sets.
python ppi_pre_preprocess_files.py --new ./data/yeast_string_refnet_2021.txt --old ./data/old_ppis.txt --genes ./data/yeast_string_genes.txt
b. compute the scores
python predict_ppi.py --emb ./data/emb/yeast_branemf_w3_alpha_1.emb --train ./data/train.pkl --test ./data/test.pkl
9. Perform Network reconstruction
python net_reconst.py --emb ./data/emb/yeast_branemf_w1_alpha_1.emb --refnet ./data/yeast_string_refnet_2021.txt --genes ./data/yeast_string_genes.txt