Update: buffer_size

tasks/VEP.wdl

@@ -23,14 +23,21 @@ task VEP {
     nt=$(nproc)
 
     awk -F'\t' '{if(($1~"^#")||($1!~"^#" && $7=="PASS")){print $0}}' ${vcf} > ${basename}.PASS.vcf
-    ncol=`awk -F'\t' '{if($1!~"^#"){print NF}}' ${basename}.PASS.vcf | uniq`
     
+    ncol=`awk -F'\t' '{if($1!~"^#"){print NF}}' ${basename}.PASS.vcf | uniq`
     if [ $ncol -lt 11 ]; then
       vcf2maf_ID="--tumor-id ${tumor_id} --normal-id ${normal_id}"
     else
       vcf2maf_ID="--tumor-id ${tumor_id}"
     fi
 
+    nrow=`awk -F'\t' '{if($1~"^chr"){print $0}}' ${basename}.PASS.vcf | wc -l`
+    if [ $nrow -lt 5000 ]; then
+      buffer_size="--buffer_size 5000"
+    else
+      buffer_size="--buffer_size 500"
+    fi
+
     # Extract the BND variants from VCF
     # awk -F'\t' '{if(($1~"^#")||($8!~".*SVTYPE=BND.*")){print $0}}' ${sample_id}.PASS.vcf > ${sample_id}.PASS.vcf2maf.vcf
     # awk -F'\t' '{if(($1~"^#")||($8~".*SVTYPE=BND.*")){print $0}}' ${sample_id}.PASS.vcf > ${sample_id}.INPUT.VEP.vcf
@@ -57,7 +64,7 @@ task VEP {
     --format vcf --vcf \
     --no_progress \
     --no_stats \
-    --buffer_size 5000 \
+    $buffer_size \
     --sift b \
     --ccds --uniprot --hgvs --symbol --numbers --domains --gene_phenotype --canonical --protein --biotype --uniprot --tsl --variant_class --shift_hgvs 1 --check_existing --total_length --allele_number --no_escape --xref_refseq --failed 1 --flag_pick_allele --pick_order canonical,tsl,biotype,rank,ccds,length --force_overwrite --offline --pubmed --regulatory