Commit 5264ad15 authored by David Hoksza's avatar David Hoksza
Browse files

ability to turn off given resources in the pipeline

parent 82f3d173
......@@ -38,25 +38,36 @@ check_exit_code(){
log() {
echo `date +"%T"`: $1
}
# Get associations from DisGeNET
log "Querying DisGeNET..."
disgenet_out_path=${RES_DIR}/01-disgenet-n_${DISGENET_CNT_THRESHOLD}-s_${DISGENET_ASSOCIATION_SCORE_THRESHOLD_STR}.json
$PYTHON_BIN $ASSOCIATIONS_DIR/disgenet.py -o ${ORPHANET_IDS} -n ${DISGENET_CNT_THRESHOLD} -s ${DISGENET_ASSOCIATION_SCORE_THRESHOLD} > ${disgenet_out_path}
check_exit_code
log "DisGeNET gene and variant associations stored in ${disgenet_out_path}"
# Get associations from OpenTargets
log "Querying OpenTargets..."
opentargets_out_path=${RES_DIR}/01-opentargets-n_${DISGENET_CNT_THRESHOLD}-s_${OPENTARGETS_ASSOCIATION_SCORE_THRESHOLD_STR}.json
$PYTHON_BIN $ASSOCIATIONS_DIR/opentargets.py -o ${ORPHANET_IDS} -n ${OPENTARGETS_CNT_THRESHOLD} -s ${OPENTARGETS_ASSOCIATION_SCORE_THRESHOLD} > ${opentargets_out_path}
check_exit_code
log "Opentargets gene and variant associations stored in ${opentargets_out_path}"
out_paths=""
if [ ${USE_DISGENET} = 1 ]; then
log "Querying DisGeNET..."
disgenet_out_path=${RES_DIR}/01-disgenet-n_${DISGENET_CNT_THRESHOLD}-s_${DISGENET_ASSOCIATION_SCORE_THRESHOLD_STR}.json
$PYTHON_BIN $ASSOCIATIONS_DIR/disgenet.py -o ${ORPHANET_IDS} -n ${DISGENET_CNT_THRESHOLD} -s ${DISGENET_ASSOCIATION_SCORE_THRESHOLD} > ${disgenet_out_path}
check_exit_code
log "DisGeNET gene and variant associations stored in ${disgenet_out_path}"
out_paths="${out_paths},${disgenet_out_path}"
fi
if [ ${USE_OPENTARGETS} = 1 ]; then
# Get associations from OpenTargets
log "Querying OpenTargets..."
opentargets_out_path=${RES_DIR}/01-opentargets-n_${DISGENET_CNT_THRESHOLD}-s_${OPENTARGETS_ASSOCIATION_SCORE_THRESHOLD_STR}.json
$PYTHON_BIN $ASSOCIATIONS_DIR/opentargets.py -o ${ORPHANET_IDS} -n ${OPENTARGETS_CNT_THRESHOLD} -s ${OPENTARGETS_ASSOCIATION_SCORE_THRESHOLD} > ${opentargets_out_path}
check_exit_code
log "Opentargets gene and variant associations stored in ${opentargets_out_path}"
out_paths="${out_paths},${opentargets_out_path}"
fi
# Merge with ClinVar
log "Merging with ClinVar..."
genes_variants_out_path=${RES_DIR}/02-genes_variants.log
$PYTHON_BIN $ASSOCIATIONS_DIR/merge_with_clinvar.py -v $disgenet_out_path,$opentargets_out_path -c ${ASSOCIATIONS_DATA_DIR}/OrphaHPO_clinvar_variants_summary.tsv -oid ${ORPHANET_IDS} > ${genes_variants_out_path}
out_paths=${out_paths#,}
$PYTHON_BIN $ASSOCIATIONS_DIR/merge_with_clinvar.py -v $out_paths -c ${ASSOCIATIONS_DATA_DIR}/OrphaHPO_clinvar_variants_summary.tsv -oid ${ORPHANET_IDS} > ${genes_variants_out_path}
check_exit_code
log "Integration with ClinVar stored in ${genes_variants_out_path}"
......
......@@ -61,7 +61,7 @@ if (dbsource == "stringdb" | dbsource == "text_mining") {
filters = 'hgnc_symbol',
values = input,
mart = ensembl)
biomartCacheClear()
ensprotein_idest <- subset(ensprotein_idest, ensembl_peptide_id != "")
message(paste0("there are ", length(unique(ensprotein_idest$hgnc_symbol))), " genes in the input file with ensembl protein ids")
......@@ -132,7 +132,7 @@ if (dbsource == "stringdb" | dbsource == "text_mining") {
filters = 'ensembl_peptide_id',
values = protein_info,
mart = ensembl)
biomartCacheClear()
protein_info_2ensp <- subset(protein_info_2ensp, hgnc_symbol != "")
ppi_network <- merge(ppi_network, protein_info_2ensp, by.x = "source", by.y = "ensembl_peptide_id" )
......
#!/usr/bin/env bash
ORPHANET_IDS="130"
#ORPHANET_IDS="851,404493,46348,85295"
#ORPHANET_IDS="33,67046,79327,79321,86309"
USE_DISGENET=1
DISGENET_CNT_THRESHOLD=50
DISGENET_ASSOCIATION_SCORE_THRESHOLD="0"
USE_OPENTARGETS=1
OPENTARGETS_CNT_THRESHOLD=50
OPENTARGETS_ASSOCIATION_SCORE_THRESHOLD="0.3"
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment