Upload tasks/clinical-wgs-interpretation/run_script.sh with huggingface_hub

tasks/clinical-wgs-interpretation/run_script.sh

@@ -0,0 +1,322 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# =============================================================================
+# Clinical WGS Interpretation Pipeline
+# =============================================================================
+#
+# DAG Structure (depth=12, convergence=5):
+#
+#  sample_R1.fq.gz    sample_R2.fq.gz
+#       |                  |
+#   [fastp QC] ------  [fastp QC]                              Level 1
+#       +--------+---------+
+#                |
+#        [bwa-mem2 align]                                       Level 2
+#        [samtools sort]                                        Level 3
+#        [picard MarkDuplicates]                                Level 4
+#        [gatk BQSR (BaseRecal + Apply)]                        Level 5
+#                |
+#        +-------+-------------------+------------------+
+#        |       |                   |                  |
+#   [bcftools  [Delly            [mosdepth           [samtools  Level 6
+#    mpileup    (SV call)]        (coverage +          flagstat]
+#    + call]                      per-gene)]
+#        |       |                   |                  |
+#   [bcftools  [bcftools             |                  |
+#    norm]      view SV VCF]         |                  |       Level 7
+#        |       |                   |                  |
+#   +----+----+  |                   |                  |
+#   |         |  |                   |                  |
+# [Select  [Select                   |                  |       Level 8
+#  SNP]     Indel]                   |                  |
+#   |         |                      |                  |
+# [Filter  [Filter                   |                  |       Level 9
+#  SNP]     Indel]                   |                  |
+#   +----+----+                      |                  |
+#        |                           |                  |
+#   [bcftools concat]                |                  |
+#        |       +-------------------+                  |
+#        |       |                                      |
+#   CONVERGENCE 1 (SNV/Indel + coverage)                        Level 9
+#        |       |
+#        |  CONVERGENCE 2 (Delly SVs)                           Level 10
+#        +-------+
+#                |
+#        CONVERGENCE 3 (all variants merged)
+#                |
+#        +-------+---------------+--------+
+#        |       |               |        |
+#   [SnpSift   [bcftools       [python    |                     Level 11
+#    annotate   annotate        score]    |
+#    ClinVar]   (dbSNP freq)]            |
+#        +-------+---------------+        |
+#                |                        |
+#        CONVERGENCE 4 (all annotations)  |
+#                |                        |
+#        +-------+-------+               |
+#        |       |       |               |                      Level 12
+#   [bcftools  [python  [python           |
+#    filter     gene    clinical           |
+#    PASS]      panel]  report]            |
+#        +-------+-------+               |
+#                |                        |
+#        CONVERGENCE 5 (coverage + QC + final)
+#        [final clinical report]
+#
+# =============================================================================
+
+THREADS=$(( $(nproc) > 8 ? 8 : $(nproc) ))
+WORKDIR="$(cd "$(dirname "$0")" && pwd)"
+DATA="${WORKDIR}/data"
+REF="${WORKDIR}/reference"
+OUT="${WORKDIR}/outputs"
+RESULTS="${WORKDIR}/results"
+
+mkdir -p "${OUT}"/{qc,aligned,processed,snv,sv,filtered,annotation,stats,coverage} "${RESULTS}"
+
+# ---- Build indexes if needed ----
+[ ! -f "${REF}/genome.fa.fai" ] && samtools faidx "${REF}/genome.fa"
+[ ! -f "${REF}/genome.dict" ] && picard CreateSequenceDictionary R="${REF}/genome.fa" O="${REF}/genome.dict" 2>&1
+[ ! -f "${REF}/genome.fa.bwt.2bit.64" ] && bwa-mem2 index "${REF}/genome.fa"
+[ ! -f "${REF}/dbsnp.vcf.gz.tbi" ] && tabix -p vcf "${REF}/dbsnp.vcf.gz"
+[ ! -f "${REF}/clinvar.vcf.gz.tbi" ] && tabix -p vcf "${REF}/clinvar.vcf.gz"
+
+# ---- Level 1: fastp QC ----
+if [ ! -f "${OUT}/qc/trimmed_R1.fastq.gz" ]; then
+  echo ">>> Level 1: fastp QC"
+  fastp -i "${DATA}/sample_R1.fastq.gz" -I "${DATA}/sample_R2.fastq.gz" \
+    -o "${OUT}/qc/trimmed_R1.fastq.gz" -O "${OUT}/qc/trimmed_R2.fastq.gz" \
+    --json "${OUT}/qc/fastp.json" --html "${OUT}/qc/fastp.html" \
+    --thread ${THREADS} --detect_adapter_for_pe --cut_front --cut_tail
+fi
+
+# ---- Level 2: bwa-mem2 align ----
+if [ ! -f "${OUT}/aligned/raw.bam" ]; then
+  echo ">>> Level 2: bwa-mem2 alignment"
+  bwa-mem2 mem -t ${THREADS} \
+    -R "@RG\tID:clinical\tSM:patient1\tPL:ILLUMINA\tLB:WGS\tPU:unit1" \
+    "${REF}/genome.fa" \
+    "${OUT}/qc/trimmed_R1.fastq.gz" "${OUT}/qc/trimmed_R2.fastq.gz" | \
+    samtools view -bS - > "${OUT}/aligned/raw.bam"
+fi
+
+# ---- Level 3: samtools sort ----
+if [ ! -f "${OUT}/aligned/sorted.bam" ]; then
+  echo ">>> Level 3: samtools sort"
+  samtools sort -@ ${THREADS} "${OUT}/aligned/raw.bam" -o "${OUT}/aligned/sorted.bam"
+  samtools index "${OUT}/aligned/sorted.bam"
+fi
+
+# ---- Level 4: picard MarkDuplicates ----
+if [ ! -f "${OUT}/processed/markdup.bam" ]; then
+  echo ">>> Level 4: picard MarkDuplicates"
+  picard MarkDuplicates \
+    I="${OUT}/aligned/sorted.bam" O="${OUT}/processed/markdup.bam" \
+    M="${OUT}/processed/markdup_metrics.txt" \
+    REMOVE_DUPLICATES=false CREATE_INDEX=true
+fi
+
+# ---- Level 5: GATK BQSR ----
+if [ ! -f "${OUT}/processed/recal.bam" ]; then
+  echo ">>> Level 5: GATK BQSR"
+  gatk BaseRecalibrator -R "${REF}/genome.fa" -I "${OUT}/processed/markdup.bam" \
+    --known-sites "${REF}/dbsnp.vcf.gz" -O "${OUT}/processed/recal_table.txt" 2>&1
+  gatk ApplyBQSR -R "${REF}/genome.fa" -I "${OUT}/processed/markdup.bam" \
+    --bqsr-recal-file "${OUT}/processed/recal_table.txt" \
+    -O "${OUT}/processed/recal.bam" 2>&1
+fi
+
+# ============================
+# Level 6: Three parallel tracks from recal BAM
+# ============================
+
+# ---- Track 1: SNV/Indel calling (bcftools) ----
+if [ ! -f "${OUT}/snv/bcf_raw.vcf.gz" ]; then
+  echo ">>> Level 6a: bcftools SNV/Indel calling"
+  bcftools mpileup -f "${REF}/genome.fa" -q 20 -Q 20 --max-depth 1000 \
+    "${OUT}/processed/recal.bam" 2>/dev/null | \
+    bcftools call -mv --ploidy GRCh38 -Oz -o "${OUT}/snv/bcf_raw.vcf.gz" 2>/dev/null
+  bcftools index "${OUT}/snv/bcf_raw.vcf.gz"
+fi
+
+# Level 7: Normalize
+if [ ! -f "${OUT}/snv/bcf_norm.vcf.gz" ]; then
+  echo ">>> Level 7: Normalize variants"
+  bcftools norm -f "${REF}/genome.fa" "${OUT}/snv/bcf_raw.vcf.gz" -Oz \
+    -o "${OUT}/snv/bcf_norm.vcf.gz" 2>/dev/null
+  bcftools index "${OUT}/snv/bcf_norm.vcf.gz"
+fi
+
+# Level 8-9: Select + Filter SNPs and Indels
+if [ ! -f "${OUT}/filtered/snps_filtered.vcf.gz" ]; then
+  echo ">>> Level 8-9: SNP filtering"
+  bcftools view -v snps "${OUT}/snv/bcf_norm.vcf.gz" -Oz -o "${OUT}/filtered/snps_raw.vcf.gz" 2>/dev/null
+  bcftools index "${OUT}/filtered/snps_raw.vcf.gz"
+  bcftools filter -i 'QUAL>=20 && INFO/DP>=2' "${OUT}/filtered/snps_raw.vcf.gz" -Oz \
+    -o "${OUT}/filtered/snps_filtered.vcf.gz" 2>/dev/null
+  bcftools index "${OUT}/filtered/snps_filtered.vcf.gz"
+fi
+
+if [ ! -f "${OUT}/filtered/indels_filtered.vcf.gz" ]; then
+  echo ">>> Level 8-9: Indel filtering"
+  bcftools view -v indels "${OUT}/snv/bcf_norm.vcf.gz" -Oz -o "${OUT}/filtered/indels_raw.vcf.gz" 2>/dev/null
+  bcftools index "${OUT}/filtered/indels_raw.vcf.gz"
+  bcftools filter -i 'QUAL>=20 && INFO/DP>=2' "${OUT}/filtered/indels_raw.vcf.gz" -Oz \
+    -o "${OUT}/filtered/indels_filtered.vcf.gz" 2>/dev/null
+  bcftools index "${OUT}/filtered/indels_filtered.vcf.gz"
+fi
+
+# Merge filtered SNPs + Indels
+if [ ! -f "${OUT}/filtered/snv_merged.vcf.gz" ]; then
+  bcftools concat -a "${OUT}/filtered/snps_filtered.vcf.gz" "${OUT}/filtered/indels_filtered.vcf.gz" -Oz \
+    -o "${OUT}/filtered/snv_merged_unsorted.vcf.gz" 2>/dev/null
+  bcftools sort "${OUT}/filtered/snv_merged_unsorted.vcf.gz" -Oz \
+    -o "${OUT}/filtered/snv_merged.vcf.gz" 2>/dev/null
+  bcftools index "${OUT}/filtered/snv_merged.vcf.gz"
+  rm -f "${OUT}/filtered/snv_merged_unsorted.vcf.gz"
+fi
+
+# ---- Track 2: SV detection (Delly) ----
+if [ ! -f "${OUT}/sv/delly.vcf.gz" ]; then
+  echo ">>> Level 6b: Delly SV calling"
+  delly call -g "${REF}/genome.fa" "${OUT}/processed/recal.bam" \
+    -o "${OUT}/sv/delly.bcf" 2>&1 || true
+  if [ -f "${OUT}/sv/delly.bcf" ]; then
+    bcftools view "${OUT}/sv/delly.bcf" -Oz -o "${OUT}/sv/delly.vcf.gz" 2>/dev/null || true
+    bcftools index "${OUT}/sv/delly.vcf.gz" 2>/dev/null || true
+  fi
+fi
+
+# ---- Track 3: Coverage analysis (mosdepth) ----
+if [ ! -f "${OUT}/coverage/coverage.mosdepth.summary.txt" ]; then
+  echo ">>> Level 6c: mosdepth coverage"
+  mosdepth --threads ${THREADS} "${OUT}/coverage/coverage" "${OUT}/processed/recal.bam"
+fi
+
+# ---- Flagstat ----
+if [ ! -f "${OUT}/stats/flagstat.txt" ]; then
+  echo ">>> Level 6d: samtools flagstat"
+  samtools flagstat "${OUT}/processed/recal.bam" > "${OUT}/stats/flagstat.txt"
+fi
+
+# ============================
+# CONVERGENCE 1-3: Merge all variant tracks
+# ============================
+echo ">>> CONVERGENCE 1-3: Merging variant tracks"
+
+SNP_COUNT=$(bcftools view -H "${OUT}/filtered/snps_filtered.vcf.gz" 2>/dev/null | wc -l || true)
+INDEL_COUNT=$(bcftools view -H "${OUT}/filtered/indels_filtered.vcf.gz" 2>/dev/null | wc -l || true)
+SNV_TOTAL=$((${SNP_COUNT:-0} + ${INDEL_COUNT:-0}))
+
+SV_COUNT=0
+if [ -s "${OUT}/sv/delly.vcf.gz" ]; then
+  SV_COUNT=$(bcftools view -H "${OUT}/sv/delly.vcf.gz" 2>/dev/null | wc -l || true)
+fi
+
+echo "SNVs: ${SNV_TOTAL} (SNPs: ${SNP_COUNT}, Indels: ${INDEL_COUNT}), SVs: ${SV_COUNT}"
+
+# ============================
+# Level 11: Triple annotation branch
+# ============================
+
+# ClinVar annotation
+if [ ! -f "${OUT}/annotation/clinvar_annotated.vcf" ]; then
+  echo ">>> Level 11a: ClinVar annotation"
+  SnpSift annotate "${REF}/clinvar.vcf.gz" "${OUT}/filtered/snv_merged.vcf.gz" \
+    > "${OUT}/annotation/clinvar_annotated.vcf" 2>/dev/null || true
+fi
+
+# bcftools stats
+if [ ! -f "${OUT}/stats/vcf_stats.txt" ]; then
+  echo ">>> Level 11b: bcftools stats"
+  bcftools stats "${OUT}/filtered/snv_merged.vcf.gz" > "${OUT}/stats/vcf_stats.txt" 2>/dev/null
+fi
+
+# Variant scoring (Ti/Tv, het/hom)
+echo ">>> Level 11c: Variant scoring"
+TITV=$(grep "^TSTV" "${OUT}/stats/vcf_stats.txt" | head -1 | cut -f5 || true)
+TITV=${TITV:-0}
+
+HET_COUNT=$(bcftools view -H "${OUT}/filtered/snv_merged.vcf.gz" 2>/dev/null | \
+  awk -F'\t' '{split($10,a,":"); if(a[1]=="0/1" || a[1]=="0|1") c++} END{print c+0}' || true)
+HOM_COUNT=$(bcftools view -H "${OUT}/filtered/snv_merged.vcf.gz" 2>/dev/null | \
+  awk -F'\t' '{split($10,a,":"); if(a[1]=="1/1" || a[1]=="1|1") c++} END{print c+0}' || true)
+
+CLINVAR_HITS=$(grep -c "CLNSIG=" "${OUT}/annotation/clinvar_annotated.vcf" 2>/dev/null || true)
+CLINVAR_HITS=${CLINVAR_HITS:-0}
+
+# ============================
+# Level 12: Gene panel + clinical report
+# ============================
+echo ">>> Level 12: Gene panel filtering + report"
+
+# Gene panel intersection
+PANEL_VARIANTS=0
+if [ -s "${REF}/gene_panel.bed" ]; then
+  bcftools view -H "${OUT}/filtered/snv_merged.vcf.gz" 2>/dev/null | \
+    awk -F'\t' '{print $1"\t"$2-1"\t"$2"\t"$4">"$5}' \
+    > "${OUT}/annotation/variant_positions.bed" || true
+  if [ -s "${OUT}/annotation/variant_positions.bed" ]; then
+    PANEL_VARIANTS=$(bedtools intersect -a "${OUT}/annotation/variant_positions.bed" \
+      -b "${REF}/gene_panel.bed" -u | wc -l || true)
+  fi
+fi
+
+# Coverage stats
+MEAN_COV=$(awk 'NR==2 {print $4}' "${OUT}/coverage/coverage.mosdepth.summary.txt" 2>/dev/null || echo "0")
+
+# Low coverage regions (< 5x)
+LOW_COV_REGIONS=0
+if [ -f "${OUT}/coverage/coverage.per-base.bed.gz" ]; then
+  LOW_COV_REGIONS=$(zcat "${OUT}/coverage/coverage.per-base.bed.gz" | \
+    awk '$4 < 5 && $4 > 0 {c++} END{print c+0}' || true)
+fi
+
+# Alignment stats
+MAPPED=$(grep "mapped (" "${OUT}/stats/flagstat.txt" | head -1 | awk '{print $1}')
+TOTAL_ALN=$(grep "in total" "${OUT}/stats/flagstat.txt" | awk '{print $1}')
+MAP_RATE=$(python3 -c "print(round(${MAPPED}/${TOTAL_ALN}*100,2)) if ${TOTAL_ALN} > 0 else print(0)")
+
+# QC stats
+RAW_READS=$(python3 -c "import json; d=json.load(open('${OUT}/qc/fastp.json')); print(d['summary']['before_filtering']['total_reads'])")
+CLEAN_READS=$(python3 -c "import json; d=json.load(open('${OUT}/qc/fastp.json')); print(d['summary']['after_filtering']['total_reads'])")
+Q30=$(python3 -c "import json; d=json.load(open('${OUT}/qc/fastp.json')); print(round(d['summary']['after_filtering']['q30_rate']*100,2))")
+
+DUP_RATE=$(grep -A1 "^LIBRARY" "${OUT}/processed/markdup_metrics.txt" | tail -1 | awk -F'\t' '{printf "%.2f", $9*100}') || DUP_RATE=0
+
+HET_HOM_RATIO=0
+if [ "${HOM_COUNT}" -gt 0 ] 2>/dev/null; then
+  HET_HOM_RATIO=$(python3 -c "print(round(${HET_COUNT}/${HOM_COUNT},2))")
+fi
+
+# MultiQC
+multiqc "${OUT}" -o "${OUT}/multiqc" --force --quiet 2>/dev/null || true
+
+# Write report
+cat > "${RESULTS}/report.csv" << CSVEOF
+metric,value
+raw_reads,${RAW_READS}
+clean_reads,${CLEAN_READS}
+q30_rate,${Q30}
+mapped_reads,${MAPPED}
+mapping_rate,${MAP_RATE}
+duplication_rate,${DUP_RATE}
+mean_coverage,${MEAN_COV}
+snp_count,${SNP_COUNT}
+indel_count,${INDEL_COUNT}
+total_snv,${SNV_TOTAL}
+sv_count,${SV_COUNT}
+ti_tv_ratio,${TITV}
+het_count,${HET_COUNT}
+hom_count,${HOM_COUNT}
+het_hom_ratio,${HET_HOM_RATIO}
+clinvar_annotated,${CLINVAR_HITS}
+gene_panel_variants,${PANEL_VARIANTS}
+low_coverage_regions,${LOW_COV_REGIONS}
+CSVEOF
+
+echo "=== Final Report ==="
+cat "${RESULTS}/report.csv"
+echo ""
+echo "Pipeline complete!"