You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

70 lines
4.5KB

  1. task benchmark {
  2. File vcf
  3. File benchmarking_dir
  4. File contig
  5. File sdf
  6. String project
  7. String sample = basename(vcf,".raw.vcf.gz")
  8. String docker
  9. String cluster_config
  10. String disk_size
  11. command <<<
  12. set -o pipefail
  13. set -e
  14. nt=$(nproc)
  15. mkdir -p /cromwell_root/tmp
  16. cp -r ${benchmarking_dir} /cromwell_root/tmp/
  17. gunzip ${vcf} -c > unzip.vcf
  18. cat unzip.vcf | grep '##' | grep -v 'contig' | cat - ${contig} > header
  19. cat unzip.vcf | grep '#CHROM' > name
  20. cat unzip.vcf | grep -v '#' > body
  21. cat body | grep -w '^chr1\|^chr2\|^chr3\|^chr4\|^chr5\|^chr6\|^chr7\|^chr8\|^chr9\|^chr10\|^chr11\|^chr12\|^chr13\|^chr14\|^chr15\|^chr16\|^chr17\|^chr18\|^chr19\|^chr20\|^chr21\|^chr22\|^chrX' > body.filtered
  22. cat header name body.filtered > filtered.vcf
  23. /opt/rtg-tools/dist/rtg-tools-3.10.1-4d58ead/rtg bgzip filtered.vcf -c > ${project}.${sample}.rtg.vcf.gz
  24. /opt/rtg-tools/dist/rtg-tools-3.10.1-4d58ead/rtg index -f vcf ${project}.${sample}.rtg.vcf.gz
  25. /opt/rtg-tools/dist/rtg-tools-3.10.1-4d58ead/rtg vcffilter -i ${project}.${sample}.rtg.vcf.gz -o ${sample}.rtg.SNV.vcf.gz --snps-only
  26. /opt/rtg-tools/dist/rtg-tools-3.10.1-4d58ead/rtg vcffilter -i ${project}.${sample}.rtg.vcf.gz -o ${sample}.rtg.INDEL.vcf.gz --non-snps-only
  27. if [[ ${sample} =~ "LCL5" ]];then
  28. /opt/rtg-tools/dist/rtg-tools-3.10.1-4d58ead/rtg vcfeval -b /cromwell_root/tmp/reference_datasets_v202103/LCL5.high.confidence.calls.SNV.vcf.gz -c ${sample}.rtg.SNV.vcf.gz -o ${sample}_SNV -t ${sdf} --evaluation-regions=/cromwell_root/tmp/reference_datasets_v202103/Quartet.high.confidence.region.v202103.bed
  29. /opt/rtg-tools/dist/rtg-tools-3.10.1-4d58ead/rtg vcfeval -b /cromwell_root/tmp/reference_datasets_v202103/LCL5.high.confidence.calls.INDEL.vcf.gz -c ${sample}.rtg.INDEL.vcf.gz -o ${sample}_INDEL -t ${sdf} --evaluation-regions=/cromwell_root/tmp/reference_datasets_v202103/Quartet.high.confidence.region.v202103.bed
  30. elif [[ ${sample} =~ "LCL6" ]]; then
  31. /opt/rtg-tools/dist/rtg-tools-3.10.1-4d58ead/rtg vcfeval -b /cromwell_root/tmp/reference_datasets_v202103/LCL6.high.confidence.calls.SNV.vcf.gz -c ${sample}.rtg.SNV.vcf.gz -o ${sample}_SNV -t ${sdf} --evaluation-regions=/cromwell_root/tmp/reference_datasets_v202103/Quartet.high.confidence.region.v202103.bed
  32. /opt/rtg-tools/dist/rtg-tools-3.10.1-4d58ead/rtg vcfeval -b /cromwell_root/tmp/reference_datasets_v202103/LCL6.high.confidence.calls.INDEL.vcf.gz -c ${sample}.rtg.INDEL.vcf.gz -o ${sample}_INDEL -t ${sdf} --evaluation-regions=/cromwell_root/tmp/reference_datasets_v202103/Quartet.high.confidence.region.v202103.bed
  33. elif [[ ${sample} =~ "LCL7" ]]; then
  34. /opt/rtg-tools/dist/rtg-tools-3.10.1-4d58ead/rtg vcfeval -b /cromwell_root/tmp/reference_datasets_v202103/LCL7.high.confidence.calls.SNV.vcf.gz -c ${sample}.rtg.SNV.vcf.gz -o ${sample}_SNV -t ${sdf} --evaluation-regions=/cromwell_root/tmp/reference_datasets_v202103/Quartet.high.confidence.region.v202103.bed
  35. /opt/rtg-tools/dist/rtg-tools-3.10.1-4d58ead/rtg vcfeval -b /cromwell_root/tmp/reference_datasets_v202103/LCL7.high.confidence.calls.INDEL.vcf.gz -c ${sample}.rtg.INDEL.vcf.gz -o ${sample}_INDEL -t ${sdf} --evaluation-regions=/cromwell_root/tmp/reference_datasets_v202103/Quartet.high.confidence.region.v202103.bed
  36. elif [[ ${sample} =~ "LCL8" ]]; then
  37. /opt/rtg-tools/dist/rtg-tools-3.10.1-4d58ead/rtg vcfeval -b /cromwell_root/tmp/reference_datasets_v202103/LCL8.high.confidence.calls.SNV.vcf.gz -c ${sample}.rtg.SNV.vcf.gz -o ${sample}_SNV -t ${sdf} --evaluation-regions=/cromwell_root/tmp/reference_datasets_v202103/Quartet.high.confidence.region.v202103.bed
  38. /opt/rtg-tools/dist/rtg-tools-3.10.1-4d58ead/rtg vcfeval -b /cromwell_root/tmp/reference_datasets_v202103/LCL8.high.confidence.calls.INDEL.vcf.gz -c ${sample}.rtg.INDEL.vcf.gz -o ${sample}_INDEL -t ${sdf} --evaluation-regions=/cromwell_root/tmp/reference_datasets_v202103/Quartet.high.confidence.region.v202103.bed
  39. else
  40. echo "only for quartet samples"
  41. fi
  42. cat ${sample}_SNV/summary.txt > ${project}.${sample}_SNV_precision_recall.txt
  43. cat ${sample}_INDEL/summary.txt > ${project}.${sample}_INDEL_precision_recall.txt
  44. >>>
  45. runtime {
  46. docker:docker
  47. cluster:cluster_config
  48. systemDisk:"cloud_ssd 40"
  49. dataDisk:"cloud_ssd " + disk_size + " /cromwell_root/"
  50. }
  51. output {
  52. File rtg_vcf = "${project}.${sample}.rtg.vcf.gz"
  53. File rtg_vcf_index = "${project}.${sample}.rtg.vcf.gz.tbi"
  54. File SNV_result = "${project}.${sample}_SNV_precision_recall.txt"
  55. File Indel_result = "${project}.${sample}_INDEL_precision_recall.txt"
  56. }
  57. }