Upload tasks/cnv-detection-wes/run_script.sh with huggingface_hub

tasks/cnv-detection-wes/run_script.sh

@@ -0,0 +1,429 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+# =============================================================================
+# CNV Detection from WES Pipeline
+# =============================================================================
+# DAG Structure (depth=10, convergence=4, tools=10):
+#
+#  tumor.fastq.gz         normal.fastq.gz
+#      |                       |
+#  [fastp QC] ----------- [fastp QC]             Level 1
+#      |                       |
+#  [bwa-mem2 align] ------ [bwa-mem2 align]      Level 2
+#      |                       |
+#  [picard MarkDup] ------ [picard MarkDup]      Level 3
+#      |                       |
+#  [mosdepth coverage] --- [mosdepth coverage]   Level 4
+#      |                       |
+#      +-----------+-----------+
+#                  |
+#          CONVERGENCE 1                          Level 5
+#          (T+N BAMs + coverage)
+#                  |
+#      +-----------+-----------+
+#      |           |           |
+#  [cnvkit      [freec]     [gatk4               Level 6
+#   batch]                   CollectRC->
+#                            DenoiseRC->
+#                            ModelSeg->
+#                            CallSeg]
+#      |           |           |
+#      +-----------+-----+-----+
+#                        |
+#                CONVERGENCE 2                    Level 7
+#                (3-caller merge)
+#                [bedtools intersect]
+#                        |
+#                +-------+-------+
+#                |       |       |
+#           [bcftools [bedtools [python            Level 8
+#            stats]   intersect gene-level
+#                     w/ genes] stats]
+#                |       |       |
+#                +-------+-------+
+#                        |
+#                CONVERGENCE 3                    Level 9
+#                (stats + gene-impact)
+#                        |
+#                [python report]
+#                CONVERGENCE 4 <-- QC + coverage  Level 10
+# =============================================================================
+
+THREADS=$(( $(nproc) > 8 ? 8 : $(nproc) ))
+WORKDIR="$(cd "$(dirname "$0")" && pwd)"
+DATA="${WORKDIR}/data"
+REF="${WORKDIR}/reference"
+OUT="${WORKDIR}/outputs"
+RES="${WORKDIR}/results"
+GENOME="${REF}/genome.fa"
+TARGETS="${REF}/targets.bed"
+
+mkdir -p "${OUT}"/{tumor,normal,cnvkit,freec,gatk_cnv,merged,qc} "${RES}"
+
+# =============================================================================
+# Index reference
+# =============================================================================
+if [ ! -f "${GENOME}.bwt.2bit.64" ]; then
+  echo "Indexing reference..."
+  bwa-mem2 index "${GENOME}"
+  samtools faidx "${GENOME}"
+  samtools dict "${GENOME}" > "${GENOME%.fa}.dict"
+fi
+
+# =============================================================================
+# Level 1-3: Process Tumor
+# =============================================================================
+if [ ! -f "${OUT}/tumor/tumor.dedup.bam" ]; then
+  echo "[L1-3] Processing tumor..."
+  fastp -i "${DATA}/tumor_R1.fastq.gz" -I "${DATA}/tumor_R2.fastq.gz" \
+    -o "${OUT}/tumor/R1_trimmed.fq.gz" -O "${OUT}/tumor/R2_trimmed.fq.gz" \
+    --json "${OUT}/qc/tumor_fastp.json" --thread "${THREADS}" --length_required 30
+
+  bwa-mem2 mem -t "${THREADS}" \
+    -R "@RG\tID:tumor\tSM:tumor\tPL:ILLUMINA\tLB:lib1" \
+    "${GENOME}" "${OUT}/tumor/R1_trimmed.fq.gz" "${OUT}/tumor/R2_trimmed.fq.gz" \
+    | samtools sort -@ "${THREADS}" -o "${OUT}/tumor/tumor.sorted.bam"
+
+  picard MarkDuplicates \
+    I="${OUT}/tumor/tumor.sorted.bam" \
+    O="${OUT}/tumor/tumor.dedup.bam" \
+    M="${OUT}/tumor/tumor.dedup_metrics.txt" \
+    REMOVE_DUPLICATES=false CREATE_INDEX=true 2>/dev/null
+fi
+
+# =============================================================================
+# Level 1-3: Process Normal
+# =============================================================================
+if [ ! -f "${OUT}/normal/normal.dedup.bam" ]; then
+  echo "[L1-3] Processing normal..."
+  fastp -i "${DATA}/normal_R1.fastq.gz" -I "${DATA}/normal_R2.fastq.gz" \
+    -o "${OUT}/normal/R1_trimmed.fq.gz" -O "${OUT}/normal/R2_trimmed.fq.gz" \
+    --json "${OUT}/qc/normal_fastp.json" --thread "${THREADS}" --length_required 30
+
+  bwa-mem2 mem -t "${THREADS}" \
+    -R "@RG\tID:normal\tSM:normal\tPL:ILLUMINA\tLB:lib1" \
+    "${GENOME}" "${OUT}/normal/R1_trimmed.fq.gz" "${OUT}/normal/R2_trimmed.fq.gz" \
+    | samtools sort -@ "${THREADS}" -o "${OUT}/normal/normal.sorted.bam"
+
+  picard MarkDuplicates \
+    I="${OUT}/normal/normal.sorted.bam" \
+    O="${OUT}/normal/normal.dedup.bam" \
+    M="${OUT}/normal/normal.dedup_metrics.txt" \
+    REMOVE_DUPLICATES=false CREATE_INDEX=true 2>/dev/null
+fi
+
+# =============================================================================
+# Level 4: mosdepth coverage
+# =============================================================================
+if [ ! -f "${OUT}/tumor/tumor.mosdepth.global.dist.txt" ]; then
+  echo "[L4] Running mosdepth coverage..."
+  mosdepth --by "${TARGETS}" -t "${THREADS}" \
+    "${OUT}/tumor/tumor" "${OUT}/tumor/tumor.dedup.bam" || true
+  mosdepth --by "${TARGETS}" -t "${THREADS}" \
+    "${OUT}/normal/normal" "${OUT}/normal/normal.dedup.bam" || true
+fi
+
+# =============================================================================
+# Level 5: CONVERGENCE 1 — T+N BAMs ready
+# =============================================================================
+echo "[L5] Convergence 1: T+N BAMs ready"
+
+# =============================================================================
+# Level 6a: CNVkit batch
+# =============================================================================
+if [ ! -f "${OUT}/cnvkit/tumor.cns" ]; then
+  echo "[L6a] Running CNVkit..."
+  cnvkit.py batch \
+    "${OUT}/tumor/tumor.dedup.bam" \
+    --normal "${OUT}/normal/normal.dedup.bam" \
+    --targets "${TARGETS}" \
+    --fasta "${GENOME}" \
+    --output-dir "${OUT}/cnvkit" \
+    --diagram --scatter \
+    -p "${THREADS}" || true
+
+  # Call CNVs
+  if [ -f "${OUT}/cnvkit/tumor.cns" ]; then
+    cnvkit.py call "${OUT}/cnvkit/tumor.cns" -o "${OUT}/cnvkit/tumor.call.cns" || true
+  fi
+fi
+
+# =============================================================================
+# Level 6b: Control-FREEC
+# =============================================================================
+if [ ! -f "${OUT}/freec/tumor_CNVs" ]; then
+  echo "[L6b] Running Control-FREEC..."
+  # Create FREEC config
+  cat > "${OUT}/freec/config.txt" << FREEC_CFG
+[general]
+chrLenFile = ${GENOME}.fai
+ploidy = 2
+window = 50000
+outputDir = ${OUT}/freec
+maxThreads = ${THREADS}
+
+[sample]
+mateFile = ${OUT}/tumor/tumor.dedup.bam
+inputFormat = BAM
+mateOrientation = FR
+
+[control]
+mateFile = ${OUT}/normal/normal.dedup.bam
+inputFormat = BAM
+mateOrientation = FR
+
+[target]
+captureRegions = ${TARGETS}
+FREEC_CFG
+
+  freec -conf "${OUT}/freec/config.txt" || true
+fi
+
+# =============================================================================
+# Level 6c: GATK4 CNV
+# =============================================================================
+if [ ! -f "${OUT}/gatk_cnv/tumor.called.seg" ]; then
+  echo "[L6c] Running GATK4 CNV pipeline..."
+
+  # Create interval list from BED
+  gatk BedToIntervalList \
+    -I "${TARGETS}" \
+    -O "${OUT}/gatk_cnv/targets.interval_list" \
+    -SD "${GENOME%.fa}.dict" || true
+
+  if [ -f "${OUT}/gatk_cnv/targets.interval_list" ]; then
+    # Preprocess intervals
+    gatk PreprocessIntervals \
+      -R "${GENOME}" \
+      -L "${OUT}/gatk_cnv/targets.interval_list" \
+      --bin-length 0 \
+      --padding 250 \
+      -O "${OUT}/gatk_cnv/preprocessed.interval_list" || true
+
+    # Collect read counts
+    for sample in tumor normal; do
+      gatk CollectReadCounts \
+        -I "${OUT}/${sample}/${sample}.dedup.bam" \
+        -L "${OUT}/gatk_cnv/preprocessed.interval_list" \
+        -R "${GENOME}" \
+        -O "${OUT}/gatk_cnv/${sample}.counts.hdf5" || true
+    done
+
+    # Create PoN from normal
+    if [ -f "${OUT}/gatk_cnv/normal.counts.hdf5" ]; then
+      gatk CreateReadCountPanelOfNormals \
+        -I "${OUT}/gatk_cnv/normal.counts.hdf5" \
+        -O "${OUT}/gatk_cnv/pon.hdf5" || true
+    fi
+
+    # Denoise read counts
+    if [ -f "${OUT}/gatk_cnv/tumor.counts.hdf5" ] && [ -f "${OUT}/gatk_cnv/pon.hdf5" ]; then
+      gatk DenoiseReadCounts \
+        -I "${OUT}/gatk_cnv/tumor.counts.hdf5" \
+        --count-panel-of-normals "${OUT}/gatk_cnv/pon.hdf5" \
+        --standardized-copy-ratios "${OUT}/gatk_cnv/tumor.standardized.tsv" \
+        --denoised-copy-ratios "${OUT}/gatk_cnv/tumor.denoised.tsv" || true
+    fi
+
+    # Model segments
+    if [ -f "${OUT}/gatk_cnv/tumor.denoised.tsv" ]; then
+      gatk ModelSegments \
+        --denoised-copy-ratios "${OUT}/gatk_cnv/tumor.denoised.tsv" \
+        -O "${OUT}/gatk_cnv" \
+        --output-prefix tumor || true
+    fi
+
+    # Call segments
+    if [ -f "${OUT}/gatk_cnv/tumor.cr.seg" ]; then
+      gatk CallCopyRatioSegments \
+        -I "${OUT}/gatk_cnv/tumor.cr.seg" \
+        -O "${OUT}/gatk_cnv/tumor.called.seg" || true
+    fi
+  fi
+fi
+
+# =============================================================================
+# Level 7: CONVERGENCE 2 — merge callers
+# =============================================================================
+echo "[L7] Merging CNV calls..."
+
+export OUT
+python3 << 'MERGE'
+import os
+
+OUT = os.environ.get("OUT", "outputs")
+os.makedirs(f"{OUT}/merged", exist_ok=True)
+
+all_cnvs = []
+
+# CNVkit calls
+cns_file = f"{OUT}/cnvkit/tumor.call.cns"
+if os.path.exists(cns_file):
+    with open(cns_file) as f:
+        next(f)  # header
+        for line in f:
+            parts = line.strip().split("\t")
+            if len(parts) >= 5:
+                chrom, start, end = parts[0], int(parts[1]), int(parts[2])
+                try:
+                    cn = int(float(parts[4]))
+                    if cn != 2:
+                        all_cnvs.append((chrom, start, end, "CNVkit", cn))
+                except:
+                    pass
+    print(f"CNVkit: {sum(1 for c in all_cnvs if c[3]=='CNVkit')} CNVs")
+
+# FREEC calls
+freec_file = f"{OUT}/freec/tumor_CNVs"
+if os.path.exists(freec_file):
+    cnt = 0
+    with open(freec_file) as f:
+        for line in f:
+            parts = line.strip().split("\t")
+            if len(parts) >= 5:
+                chrom, start, end = parts[0], int(parts[1]), int(parts[2])
+                cn = int(parts[3])
+                if cn != 2:
+                    all_cnvs.append((chrom, start, end, "FREEC", cn))
+                    cnt += 1
+    print(f"FREEC: {cnt} CNVs")
+
+# GATK calls
+gatk_file = f"{OUT}/gatk_cnv/tumor.called.seg"
+if os.path.exists(gatk_file):
+    cnt = 0
+    with open(gatk_file) as f:
+        for line in f:
+            if line.startswith("@") or line.startswith("CONTIG"):
+                continue
+            parts = line.strip().split("\t")
+            if len(parts) >= 6:
+                chrom, start, end = parts[0], int(parts[1]), int(parts[2])
+                call = parts[5] if len(parts) > 5 else ""
+                if call in ("+", "-", "AMP", "DEL"):
+                    all_cnvs.append((chrom, start, end, "GATK", call))
+                    cnt += 1
+    print(f"GATK: {cnt} CNVs")
+
+# Write merged BED
+with open(f"{OUT}/merged/all_cnvs.bed", "w") as f:
+    for chrom, start, end, caller, cn in sorted(all_cnvs):
+        f.write(f"{chrom}\t{start}\t{end}\t{caller}\t{cn}\n")
+
+print(f"Total merged CNVs: {len(all_cnvs)}")
+MERGE
+
+# =============================================================================
+# Level 8-10: Stats + Report
+# =============================================================================
+echo "[L8-10] Generating report..."
+
+python3 << 'REPORT'
+import os, json
+
+OUT = os.environ.get("OUT", "outputs")
+RES_DIR = os.environ.get("RES", "results")
+
+metrics = {}
+
+# fastp stats
+for sample in ["tumor", "normal"]:
+    fp = f"{OUT}/qc/{sample}_fastp.json"
+    if os.path.exists(fp):
+        with open(fp) as f:
+            fj = json.load(f)
+        metrics[f"{sample}_reads"] = fj["summary"]["before_filtering"]["total_reads"] // 2
+        metrics[f"{sample}_reads_after_trim"] = fj["summary"]["after_filtering"]["total_reads"] // 2
+
+# mosdepth coverage
+for sample in ["tumor", "normal"]:
+    summ = f"{OUT}/{sample}/{sample}.mosdepth.summary.txt"
+    if os.path.exists(summ):
+        with open(summ) as f:
+            for line in f:
+                if line.startswith("total_region") or line.startswith("total"):
+                    parts = line.strip().split("\t")
+                    if len(parts) >= 4:
+                        metrics[f"{sample}_mean_coverage"] = float(parts[3])
+                        break
+
+# Picard dedup metrics
+for sample in ["tumor", "normal"]:
+    dup_file = f"{OUT}/{sample}/{sample}.dedup_metrics.txt"
+    if os.path.exists(dup_file):
+        with open(dup_file) as f:
+            header = None
+            for line in f:
+                if line.startswith("LIBRARY"):
+                    header = line.strip().split("\t")
+                elif header and not line.startswith("#") and line.strip():
+                    vals = line.strip().split("\t")
+                    if len(vals) > 8:
+                        try:
+                            metrics[f"{sample}_dup_pct"] = round(float(vals[8]) * 100, 2)
+                        except:
+                            pass
+                    break
+
+# CNV caller counts
+for caller, prefix in [("cnvkit", "cnvkit"), ("freec", "freec"), ("gatk", "gatk")]:
+    count = 0
+    if caller == "cnvkit":
+        f = f"{OUT}/cnvkit/tumor.call.cns"
+        if os.path.exists(f):
+            with open(f) as fh:
+                next(fh)
+                for line in fh:
+                    parts = line.strip().split("\t")
+                    if len(parts) >= 5:
+                        try:
+                            cn = int(float(parts[4]))
+                            if cn != 2:
+                                count += 1
+                        except:
+                            pass
+    elif caller == "freec":
+        f = f"{OUT}/freec/tumor_CNVs"
+        if os.path.exists(f):
+            with open(f) as fh:
+                for line in fh:
+                    parts = line.strip().split("\t")
+                    if len(parts) >= 4 and parts[3] != "2":
+                        count += 1
+    elif caller == "gatk":
+        f = f"{OUT}/gatk_cnv/tumor.called.seg"
+        if os.path.exists(f):
+            with open(f) as fh:
+                for line in fh:
+                    if not line.startswith("@") and not line.startswith("CONTIG"):
+                        parts = line.strip().split("\t")
+                        if len(parts) > 5 and parts[5] in ("+", "-", "AMP", "DEL"):
+                            count += 1
+    metrics[f"{prefix}_cnv_count"] = count
+
+# Merged CNVs
+merged_f = f"{OUT}/merged/all_cnvs.bed"
+if os.path.exists(merged_f):
+    with open(merged_f) as f:
+        metrics["total_merged_cnvs"] = sum(1 for l in f if l.strip())
+
+# Target info
+targets_f = os.environ.get("REF", "reference") + "/targets.bed"
+if os.path.exists(targets_f):
+    with open(targets_f) as f:
+        metrics["target_intervals"] = sum(1 for l in f if l.strip())
+
+# Write CSV
+with open(f"{RES_DIR}/report.csv", "w") as f:
+    f.write("metric,value\n")
+    for k, v in metrics.items():
+        f.write(f"{k},{v}\n")
+
+print("Report written:")
+for k, v in metrics.items():
+    print(f"  {k} = {v}")
+REPORT
+
+echo "=== Pipeline Complete ==="
+cat "${RES}/report.csv"