PGx CN: pgx copy-number train|call (CLI-031/032, PCN-001..017, IN-050), region BED v1, demo identities (K-028)
- simurg/pgx/cn.py: per-region CN mixture, mu_c = delta + gamma*log2((c+1)/3), free lower CN-0 component, Hardy-Weinberg-constrained weights over haplotype copy numbers, EM from a grid of label starts. - pipeline/pgx_cn.py: model file (.simurg-pgxcn), batch rules (PCN-004), per-sample CN VCF + tbi, batch and sample summaries. - The GSA v4 PGx demo arrays are 30 CEPH/1000G individuals (scripts/pgx/demo36_identity.tsv); against NYGC 30x SV RD_CN 224/240 region calls agree (GSTM1, UGT2B17 30/30), in-sample (docs/validation/V2_pgx_cn.md). Co-Authored-By: Claude Opus 5.5 <[email protected]> Claude-Session: https://claude.ai/code/session_01814GKVzZVL3FuF3DeiZXhv
This commit is contained in:
kemalandClaude Opus 5.5 committed
1 parent
8c425182c3
commit
58e70bc894
14 files changed
+1306
-1
No files matched your search
@@ -0,0 +1,85 @@
|
||||
#!/usr/bin/env python3
|
||||
"""Build the PGx CN target-region BED (spec PCN-001) from UCSC ncbiRefSeq (GRCh38).
|
||||
|
||||
python scripts/pgx/build_cn_regions.py ncbiRefSeq.hg38.txt.gz OUT.bed [--version N]
|
||||
|
||||
Regions (0-based half-open BED: chrom start0 end region gene), from the MANE/first-listed transcript:
|
||||
CYP2D6.p5 2 kb 5' flank of CYP2D6 (upstream of exon 1; the gene is on the minus strand)
|
||||
CYP2D6.intron.2 CYP2D6 intron 2
|
||||
CYP2D6.exon.9 CYP2D6 exon 9
|
||||
CYP2A6.exon.1 CYP2A6 exon 1 (± 500 bp)
|
||||
CYP2A6.intron.7 CYP2A6 intron 7
|
||||
GSTM1, GSTT1 (on chr22_KI270879v1_alt, IN-062), UGT2B17, SULT1A1, CYP2E1: whole gene.
|
||||
Exons and introns are numbered in transcript order (strand-aware). [DB: NCBI RefSeq via UCSC ncbiRefSeq]
|
||||
"""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import argparse
|
||||
import gzip
|
||||
from pathlib import Path
|
||||
|
||||
TRANSCRIPT = {"CYP2D6": "NM_000106", "CYP2A6": "NM_000762", "GSTM1": "NM_000561", "GSTT1": "NM_000853",
|
||||
"UGT2B17": "NM_001077", "SULT1A1": "NM_001055", "CYP2E1": "NM_000773"} # fmt: skip
|
||||
FLANK = 2000
|
||||
PAD = 500
|
||||
|
||||
|
||||
def main() -> int:
|
||||
ap = argparse.ArgumentParser(description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter)
|
||||
ap.add_argument("refseq", type=Path)
|
||||
ap.add_argument("out", type=Path)
|
||||
ap.add_argument("--version", type=int, default=1)
|
||||
a = ap.parse_args()
|
||||
tx: dict[str, tuple[str, str, list[int], list[int]]] = {}
|
||||
with gzip.open(a.refseq, "rt") as fh:
|
||||
for line in fh:
|
||||
f = line.rstrip("\n").split("\t")
|
||||
name, chrom, strand, gene = f[1], f[2], f[3], f[12]
|
||||
if gene not in TRANSCRIPT or not name.startswith(TRANSCRIPT[gene] + "."):
|
||||
continue
|
||||
if "_" in chrom and not (gene == "GSTT1" and chrom == "chr22_KI270879v1_alt"):
|
||||
continue # primary assembly, except GSTT1 which is only on the alt contig
|
||||
starts = [int(x) for x in f[9].rstrip(",").split(",")]
|
||||
ends = [int(x) for x in f[10].rstrip(",").split(",")]
|
||||
tx.setdefault(gene, (chrom, strand, starts, ends))
|
||||
missing = set(TRANSCRIPT) - set(tx)
|
||||
if missing:
|
||||
raise SystemExit(f"transcripts not found: {sorted(missing)}")
|
||||
rows: list[tuple[str, int, int, str, str]] = []
|
||||
|
||||
def exon(gene: str, k: int) -> tuple[int, int]:
|
||||
chrom, strand, s, e = tx[gene]
|
||||
i = k - 1 if strand == "+" else len(s) - k
|
||||
return s[i], e[i]
|
||||
|
||||
def intron(gene: str, k: int) -> tuple[int, int]:
|
||||
chrom, strand, s, e = tx[gene]
|
||||
a_, b_ = exon(gene, k), exon(gene, k + 1)
|
||||
return (a_[1], b_[0]) if strand == "+" else (b_[1], a_[0])
|
||||
|
||||
for gene in ("CYP2D6", "CYP2A6"):
|
||||
chrom, strand, s, e = tx[gene]
|
||||
if gene == "CYP2D6":
|
||||
p5 = (e[-1], e[-1] + FLANK) if strand == "-" else (s[0] - FLANK, s[0])
|
||||
rows += [(chrom, *p5, "CYP2D6.p5", gene), (chrom, *intron(gene, 2), "CYP2D6.intron.2", gene),
|
||||
(chrom, *exon(gene, 9), "CYP2D6.exon.9", gene)] # fmt: skip
|
||||
else:
|
||||
x1 = exon(gene, 1)
|
||||
rows += [(chrom, x1[0] - PAD, x1[1] + PAD, "CYP2A6.exon.1", gene),
|
||||
(chrom, *intron(gene, 7), "CYP2A6.intron.7", gene)] # fmt: skip
|
||||
for gene in ("GSTM1", "GSTT1", "UGT2B17", "SULT1A1", "CYP2E1"):
|
||||
chrom, strand, s, e = tx[gene]
|
||||
rows.append((chrom, s[0], e[-1], gene, gene))
|
||||
order = {f"chr{i}": i for i in range(1, 23)}
|
||||
rows.sort(key=lambda r: (order.get(r[0].split("_")[0], 99), r[0], r[1]))
|
||||
with a.out.open("w") as fh:
|
||||
fh.write(f"# simurg PGx CN regions v{a.version}, GRCh38, from UCSC ncbiRefSeq (PCN-001)\n")
|
||||
for r in rows:
|
||||
fh.write("\t".join(map(str, r)) + "\n")
|
||||
print(a.out.read_text())
|
||||
return 0
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
raise SystemExit(main())
|
||||
@@ -0,0 +1,39 @@
|
||||
# GSA v4 PGx demo (idats_Demo_36) sample identities: best match among the 3202 1000 Genomes 30x individuals
|
||||
# by genotype concordance on 4647 chr22 SNPs with 1000G AF 0.05-0.95 (next best <= 0.91: relatives)
|
||||
sample_id individual concordance second_best second_concordance
|
||||
207895980019_R03C01 NA12875 0.9920 NA12865 0.7038
|
||||
207895980019_R03C02 NA12156 0.9933 NA10831 0.7021
|
||||
207895980019_R04C01 NA12156 0.9931 NA10831 0.7027
|
||||
207895980019_R05C01 NA12044 0.9931 NA10857 0.6969
|
||||
207895980019_R06C01 NA11992 0.9931 NA10860 0.6848
|
||||
207895980019_R07C01 NA06993 0.9940 NA06991 0.7209
|
||||
207895980019_R07C03 NA07019 0.9953 NA07056 0.7051
|
||||
207895980019_R08C01 NA12239 0.9938 NA10847 0.7022
|
||||
207895980019_R09C01 NA12716 0.9944 NA12707 0.6896
|
||||
207895980019_R09C03 NA06991 0.9931 NA06993 0.7215
|
||||
207895980019_R10C01 NA12878 0.9933 NA12892 0.7022
|
||||
207895980019_R11C01 NA12760 0.9948 NA12752 0.7171
|
||||
207895980019_R11C02 NA06985 0.9948 NA06991 0.6996
|
||||
207895980019_R12C01 NA11993 0.9961 NA10860 0.7168
|
||||
207895980019_R12C02 NA12003 0.9935 NA10838 0.6998
|
||||
207895980019_R13C03 NA06994 0.9948 NA07029 0.7114
|
||||
207895980019_R14C03 NA11993 0.9959 NA10860 0.7163
|
||||
207895980019_R15C02 NA12003 0.9931 NA10838 0.6993
|
||||
207895980035_R03C03 NA07029 0.9952 NA07000 0.7237
|
||||
207895980035_R04C03 NA12892 0.9957 NA12878 0.7005
|
||||
207895980035_R05C03 NA12248 0.9950 NA10835 0.7143
|
||||
207895980035_R06C03 NA10846 0.9920 NA12145 0.7056
|
||||
207895980035_R07C01 NA07000 0.9957 NA07029 0.7231
|
||||
207895980035_R07C03 NA12155 0.9935 NA10831 0.6903
|
||||
207895980035_R08C01 NA11832 0.9942 NA10855 0.7252
|
||||
207895980035_R08C03 NA06993 0.9940 NA06991 0.7205
|
||||
207895980035_R09C03 NA10838 0.9948 NA12004 0.7007
|
||||
207895980035_R10C01 NA12813 0.9937 NA07045 0.9103
|
||||
207895980035_R10C03 NA12249 0.9938 NA10835 0.6858
|
||||
207895980035_R11C03 NA12812 0.9942 NA12801 0.7061
|
||||
207895980035_R12C03 NA11881 0.9935 NA10859 0.7138
|
||||
207895980035_R13C01 NA11995 0.9953 NA10861 0.6857
|
||||
207895980035_R13C03 NA12248 0.9950 NA10835 0.7145
|
||||
207895980035_R14C01 NA12891 0.9942 NA12878 0.6998
|
||||
207895980035_R15C03 NA12043 0.9946 NA10857 0.7162
|
||||
207895980042_R15C02 NA11832 0.9942 NA10855 0.7253
|
||||
|
@@ -0,0 +1,72 @@
|
||||
#!/usr/bin/env python3
|
||||
"""Validate `simurg pgx copy-number call` against the NYGC 30x SV callset (1KGP_3202 freeze V3) — PCN-017.
|
||||
|
||||
python scripts/pgx/validate_cn_truth.py --calls CALL_DIR --identity scripts/pgx/demo36_identity.tsv \
|
||||
--sv 1KGP_3202.sv_freeze_V3.vcf.gz --bcftools BCFTOOLS --out summary.json
|
||||
|
||||
Truth per region = RD_CN of one SV record per region (below, chosen as the record spanning the whole region; for
|
||||
GSTM1 the read-depth DUP/CNV record, whose RD_CN gives 0/1/2 in Hardy–Weinberg proportions — the DEL record there
|
||||
reports CN 1 for 27/30 individuals). CYP2D6 sub-regions share one gene-level record, so exon-9 conversions (*36) and
|
||||
2D7-2D6 hybrids (*13, *68) legitimately differ from it. GSTT1 lies on an alt contig with no SV calls: only its CN
|
||||
distribution is reported. Replicate arrays of one individual are counted once (the first array).
|
||||
"""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import argparse
|
||||
import collections
|
||||
import csv
|
||||
import json
|
||||
import subprocess
|
||||
from pathlib import Path
|
||||
|
||||
TRUTH_RECORD = {"GSTM1": "HGSV_9691", "UGT2B17": "HGSV_63272", "SULT1A1": "HGSV_207805",
|
||||
"CYP2A6.intron.7": "HGSV_232737", "CYP2A6.exon.1": "HGSV_232737", "CYP2D6.exon.9": "HGSV_247317",
|
||||
"CYP2D6.intron.2": "HGSV_247317", "CYP2D6.p5": "HGSV_247317"} # fmt: skip
|
||||
|
||||
|
||||
def main() -> int:
|
||||
ap = argparse.ArgumentParser(description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter)
|
||||
ap.add_argument("--calls", required=True, type=Path)
|
||||
ap.add_argument("--identity", required=True, type=Path)
|
||||
ap.add_argument("--sv", required=True)
|
||||
ap.add_argument("--bcftools", default="bcftools")
|
||||
ap.add_argument("--out", required=True, type=Path)
|
||||
a = ap.parse_args()
|
||||
ident = {}
|
||||
with a.identity.open() as fh:
|
||||
for r in csv.DictReader((ln for ln in fh if not ln.startswith("#")), delimiter="\t"):
|
||||
ident[r["sample_id"]] = r["individual"]
|
||||
calls: dict[str, dict[str, str]] = {}
|
||||
with (a.calls / "pgx_cn_sample_summary.tsv").open() as fh:
|
||||
for r in csv.DictReader(fh, delimiter="\t"):
|
||||
ind = ident.get(r["sample_id"])
|
||||
if ind and ind not in calls:
|
||||
calls[ind] = {k[3:]: v for k, v in r.items() if k.startswith("cn_")}
|
||||
ids = sorted(set(TRUTH_RECORD.values()))
|
||||
q = subprocess.run([a.bcftools, "query", "-H", "-s", ",".join(sorted(calls)), "--force-samples", "-i",
|
||||
"ID=@/dev/stdin", "-f", "%ID[\t%RD_CN]\n", a.sv], input="\n".join(ids), capture_output=True,
|
||||
text=True, check=True).stdout.splitlines() # fmt: skip
|
||||
samples = [h.split("]")[1].split(":")[0] for h in q[0].split("\t")[1:]]
|
||||
truth = {ln.split("\t")[0]: dict(zip(samples, ln.split("\t")[1:], strict=True)) for ln in q[1:]}
|
||||
res: dict[str, dict[str, object]] = {}
|
||||
tot = [0, 0]
|
||||
for region, rid in TRUTH_RECORD.items():
|
||||
pairs = [(truth[rid][ind], c[region]) for ind, c in calls.items() if ind in truth.get(rid, {})]
|
||||
pairs = [(t, c) for t, c in pairs if t not in (".", "")]
|
||||
ok = sum(t == c for t, c in pairs)
|
||||
tot[0] += ok
|
||||
tot[1] += len(pairs)
|
||||
res[region] = {"record": rid, "concordant": ok, "n": len(pairs),
|
||||
"rate": round(ok / len(pairs), 4) if pairs else None,
|
||||
"truth_to_call_mismatches": dict(collections.Counter(f"{t}->{c}" for t, c in pairs if t != c))} # fmt: skip
|
||||
summary = {"individuals": len(calls), "regions": res, "overall": {"concordant": tot[0], "n": tot[1],
|
||||
"rate": round(tot[0] / tot[1], 4) if tot[1] else None},
|
||||
"gstt1_cn_distribution": dict(sorted(collections.Counter(c.get("GSTT1", ".") for c in calls.values()).items()))} # fmt: skip
|
||||
a.out.write_text(json.dumps(summary, indent=1))
|
||||
print(json.dumps(summary, indent=1))
|
||||
return 0
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
raise SystemExit(main())
|
||||
Reference in new issue
Block a user