|
|
@@ -23,19 +23,15 @@ task VEP { |
|
|
|
set -e |
|
|
|
nt=$(nproc) |
|
|
|
|
|
|
|
if [ only_pass ]; then |
|
|
|
awk -F'\t' '{if(($1~"^#")||($1!~"^#" && $7=="PASS")){print $0}}' ${vcf} > ${sample_id}.INPUT.vcf |
|
|
|
else |
|
|
|
cp ${vcf} ${sample_id}.INPUT.vcf |
|
|
|
fi |
|
|
|
|
|
|
|
awk -F'\t' '{if(($1~"^#")||($1!~"^#" && $7=="PASS")){print $0}}' ${vcf} > ${sample_id}.PASS.vcf |
|
|
|
|
|
|
|
# Extract the BND variants from VCF |
|
|
|
# awk -F'\t' '{if(($1~"^#")||($8!~".*SVTYPE=BND.*")){print $0}}' ${sample_id}.INPUT.vcf > ${sample_id}.INPUT.vcf2maf.vcf |
|
|
|
# awk -F'\t' '{if(($1~"^#")||($8~".*SVTYPE=BND.*")){print $0}}' ${sample_id}.INPUT.vcf > ${sample_id}.INPUT.VEP.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 |
|
|
|
|
|
|
|
# vcf2maf |
|
|
|
# perl ${vcf2maf_path}/vcf2maf.pl \ |
|
|
|
# --input-vcf ${sample_id}.INPUT.vcf2maf.vcf --output-maf ${basename}.maf \ |
|
|
|
# --input-vcf ${sample_id}.PASS.vcf2maf.vcf --output-maf ${basename}.maf \ |
|
|
|
# --tumor-id ${tumor_id} --normal-id ${normal_id} \ |
|
|
|
# --ref-fasta ${ref_dir}/${fasta} \ |
|
|
|
# --vep-path ${vep_path} \ |
|
|
@@ -46,13 +42,14 @@ task VEP { |
|
|
|
|
|
|
|
# vep |
|
|
|
perl ${vep_path}/vep \ |
|
|
|
--input_file ${sample_id}.INPUT.vcf --output_file ${basename}.vep.vcf \ |
|
|
|
--input_file ${sample_id}.PASS.vcf --output_file ${basename}.vep.vcf \ |
|
|
|
--fasta ${ref_dir}/${fasta} \ |
|
|
|
--dir ${cache} \ |
|
|
|
--assembly ${ncbi_build} \ |
|
|
|
--species ${species} \ |
|
|
|
--fork $nt \ |
|
|
|
--no_progress --no_stats --buffer_size 5000 --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 --tab --flag_pick_allele --pick_order canonical,tsl,biotype,rank,ccds,length --format vcf --force_overwrite --offline --pubmed --regulatory |
|
|
|
--format vcf --vcf \ |
|
|
|
--no_progress --no_stats --buffer_size 5000 --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 |
|
|
|
|
|
|
|
# vcf2maf |
|
|
|
perl ${vcf2maf_path}/vcf2maf.pl \ |
|
|
@@ -60,10 +57,6 @@ task VEP { |
|
|
|
--input-vcf ${basename}.vep.vcf --output-maf ${basename}.maf \ |
|
|
|
--tumor-id ${tumor_id} --normal-id ${normal_id} \ |
|
|
|
--ref-fasta ${ref_dir}/${fasta} \ |
|
|
|
--vep-path ${vep_path} \ |
|
|
|
--vep-data ${cache} \ |
|
|
|
--ncbi-build ${ncbi_build} \ |
|
|
|
--species ${species} \ |
|
|
|
--vep-fork $nt |
|
|
|
>>> |
|
|
|
|
|
|
@@ -75,7 +68,7 @@ task VEP { |
|
|
|
} |
|
|
|
|
|
|
|
output { |
|
|
|
File input_vcf = "${sample_id}.INPUT.vcf" |
|
|
|
File input_vcf = "${sample_id}.PASS.vcf" |
|
|
|
File vep_vcf = "${basename}.vep.vcf" |
|
|
|
File maf = "${basename}.maf" |
|
|
|
} |