Ви не можете вибрати більше 25 тем Теми мають розпочинатися з літери або цифри, можуть містити дефіси (-) і не повинні перевищувати 35 символів.

105 lines
6.5KB

  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. if [[ ${sample} =~ "gz" ]];then
  18. gunzip ${vcf} -c > unzip.vcf
  19. else
  20. cp ${vcf} unzip.vcf
  21. fi
  22. cat unzip.vcf | grep '#CHROM' > name
  23. cat unzip.vcf | grep -v '#' > body
  24. 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
  25. if grep -q PASS "body.filtered"; then
  26. cat body.filtered | grep -v '0/0' | grep 'PASS' > body.filtered.gt
  27. cat unzip.vcf | grep '##' | grep -v 'contig' | cat - ${contig} name body.filtered.gt > filtered.vcf
  28. else
  29. cat body.filtered | grep -v '0/0' > body.filtered.gt
  30. cat unzip.vcf | grep '##' | grep -v 'contig' | cat - ${contig} name body.filtered.gt > filtered.vcf
  31. fi
  32. echo -e "#CHROM\tPOS\tID\tREF\tALT\tQUAL\tFILTER\tINFO\tFORMAT\tLCL5" > LCL5_name
  33. echo -e "#CHROM\tPOS\tID\tREF\tALT\tQUAL\tFILTER\tINFO\tFORMAT\tLCL6" > LCL6_name
  34. echo -e "#CHROM\tPOS\tID\tREF\tALT\tQUAL\tFILTER\tINFO\tFORMAT\tLCL7" > LCL7_name
  35. echo -e "#CHROM\tPOS\tID\tREF\tALT\tQUAL\tFILTER\tINFO\tFORMAT\tLCL8" > LCL8_name
  36. /opt/rtg-tools/dist/rtg-tools-3.10.1-4d58ead/rtg bgzip filtered.vcf -c > ${project}.${sample}.rtg.vcf.gz
  37. /opt/rtg-tools/dist/rtg-tools-3.10.1-4d58ead/rtg index -f vcf ${project}.${sample}.rtg.vcf.gz
  38. /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
  39. /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
  40. if [[ ${sample} =~ "LCL5" ]];then
  41. /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
  42. /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
  43. cat filtered.vcf | grep '##' > header
  44. cat filtered.vcf | grep -v '#' > body
  45. cat header LCL5_name body > LCL5.vcf
  46. /opt/rtg-tools/dist/rtg-tools-3.10.1-4d58ead/rtg bgzip LCL5.vcf -c > ${project}.${sample}.reformed.vcf.gz
  47. /opt/rtg-tools/dist/rtg-tools-3.10.1-4d58ead/rtg index -f vcf ${project}.${sample}.reformed.vcf.gz
  48. elif [[ ${sample} =~ "LCL6" ]]; then
  49. /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
  50. /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
  51. cat filtered.vcf | grep '##' > header
  52. cat filtered.vcf | grep -v '#' > body
  53. cat header LCL6_name body > LCL6.vcf
  54. /opt/rtg-tools/dist/rtg-tools-3.10.1-4d58ead/rtg bgzip LCL6.vcf -c > ${project}.${sample}.reformed.vcf.gz
  55. /opt/rtg-tools/dist/rtg-tools-3.10.1-4d58ead/rtg index -f vcf ${project}.${sample}.reformed.vcf.gz
  56. elif [[ ${sample} =~ "LCL7" ]]; then
  57. /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
  58. /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
  59. cat filtered.vcf | grep '##' > header
  60. cat filtered.vcf | grep -v '#' > body
  61. cat header LCL7_name body > LCL7.vcf
  62. /opt/rtg-tools/dist/rtg-tools-3.10.1-4d58ead/rtg bgzip LCL7.vcf -c > ${project}.${sample}.reformed.vcf.gz
  63. /opt/rtg-tools/dist/rtg-tools-3.10.1-4d58ead/rtg index -f vcf ${project}.${sample}.reformed.vcf.gz
  64. elif [[ ${sample} =~ "LCL8" ]]; then
  65. /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
  66. /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
  67. cat filtered.vcf | grep '##' > header
  68. cat filtered.vcf | grep -v '#' > body
  69. cat header LCL8_name body > LCL8.vcf
  70. /opt/rtg-tools/dist/rtg-tools-3.10.1-4d58ead/rtg bgzip LCL8.vcf -c > ${project}.${sample}.reformed.vcf.gz
  71. /opt/rtg-tools/dist/rtg-tools-3.10.1-4d58ead/rtg index -f vcf ${project}.${sample}.reformed.vcf.gz
  72. else
  73. echo "only for quartet samples"
  74. fi
  75. cat ${sample}_SNV/summary.txt > ${project}.${sample}_SNV_precision_recall.txt
  76. cat ${sample}_INDEL/summary.txt > ${project}.${sample}_INDEL_precision_recall.txt
  77. >>>
  78. runtime {
  79. docker:docker
  80. cluster:cluster_config
  81. systemDisk:"cloud_ssd 40"
  82. dataDisk:"cloud_ssd " + disk_size + " /cromwell_root/"
  83. }
  84. output {
  85. File rtg_vcf = "${project}.${sample}.reformed.vcf.gz"
  86. File rtg_vcf_index = "${project}.${sample}.reformed.vcf.gz.tbi"
  87. File SNV_result = "${project}.${sample}_SNV_precision_recall.txt"
  88. File Indel_result = "${project}.${sample}_INDEL_precision_recall.txt"
  89. }
  90. }