Usage

1. Put data into input_data/.

This repository already has a few files in input_data/. You need to add data in input_data/pathways/ and input_data/genes/.

The files in input_data/pathways/ contain the associations between each phenotype and pathway. Their format is:

• filenames: PheCode_*_{GO,Curated}.wConditional.txt.gz
  • eg, PheCode_986_Curated.wConditional.txt.gz contains data about the Curated pathways for phecode 986.
• format: tab-delimited with columns pathway_name, pathway_url, association_pvalue, pathway_genes_that_the_method_selected_as_meaningful.
  • eg, KEGG_GLYCOLYSIS_GLUCONEOGENESIS http://www.broadinstitute.org/gsea/msigdb/cards/KEGG_GLYCOLYSIS_GLUCONEOGENESIS 0.993 ALDH3A2,ADH1B,HK3,ALDH1B1,FBP1,TPI1,PFKP
  • optionally, you may include unused columns before and after pathway_genes_that_the_method_selected_as_meaningful

The files in input_data/genes/ contain the associations between each phenotype and gene. Their format is:

• filenames: OUTF_PheCode_*.txt.gz
  • eg, OUTF_PheCode_803.2.txt.gz contains data about phenotype 803.2
• format: tab-delimited or space-delimited with columns gene_name, pvalue_or_NA
  • eg, 7SK NA or A1BG 8.35e-01
  • optionally, you may add 3 columns before and after pvalue_or_NA.

These files can be collected from Flux by running:

mkdir -p ~/pathweb-data/pathways/ && cd /scratch/leeshawn_fluxod/diptavo/1000G/RESULTS/ && find . -name 'PheCode_*[dO].wConditional.txt'|sort -n|while read f; do echo $f; cat $f|cut -d' ' -f1-3,5|tr " " "\t"|gzip - > ~/pathweb-data/pathways/$(basename $f).gz; done
mkdir -p ~/pathweb-data/genes/ && cd /scratch/leeshawn_fluxod/diptavo/1000G/RESULTS/ && find . -name 'OUTF_PheCode_*.txt'|sort -n|while read f; do echo $f; cat $f|perl -nae 'print "$F[0]\t";if ($F[4] eq "NA") {print("NA\n")} else {printf("%.2e\n",$F[4])}' | gzip - > ~/pathweb-data/genes/$(basename $f).gz; done

2. Populate the databases and run the server.

If you are on an Ubuntu server you can simply run ./setup-server.sh which should install the required tools, process the data, and configure and start the Flask app and Nginx using Systemd.

If you are on a laptop or otherwise don't want to use that script, then:

1. run pip3 install -r requirments.txt (which may require you to set up and activate a virtualenv or miniconda or use sudo)

2. run python3 pathweb/make_sqlite3_db.py to produce pathweb/pheno_pathway_assoc.db.

3. run python3 pathweb/make_gene_sqlite3_db.py to produce pathweb/gene.db.

4. run python3 pathweb/make_tables.py to produce pathweb/static/phenotypes.json and pathweb/static/pathways.json.

5. run the server with one of these:

   • python3 pathweb/serve.py (insecure and slow for development/debugging)
   • cd pathweb && gunicorn serve:app -k gevent -w4 --bind 0.0.0.0:8000 (fast for production)
   • sudo systemctl restart gunicorn-pathweb (if the server is run by /etc/systemd/system/gunicorn-pathweb.service)