gnomancer 0.3.1__py3-none-any.whl

gnomancer/__init__.py

@@ -0,0 +1,2 @@
+"""gnomancer: vendor-agnostic analysis of consumer raw DNA files."""
+__version__ = "0.3.1"

gnomancer/__main__.py

@@ -0,0 +1,2 @@
+from .cli import main
+main()

gnomancer/cli.py

@@ -0,0 +1,155 @@
+"""Command-line entry point: python -m gnomancer <raw file> [--out DIR]"""
+from __future__ import annotations
+
+import argparse
+import json
+import sys
+from pathlib import Path
+
+from . import __version__
+from .load import load
+from .qc import chromosome_table, summary
+from .roh import find_roh, interpret
+from .panel import load_markers, evaluate, combos
+from .haplogroups import call_y, call_mt
+from .report import write
+from . import clinvar as clinvar_mod
+from . import pharmcat as pharmcat_mod
+from . import yleaf as yleaf_mod
+from . import summary as summary_mod
+from .report_html import write_html
+
+from .paths import DATA, ancestry_python
+
+
+def run_ezancestry(inp: Path, loaded) -> dict | None:
+    """Run ezancestry in a separate interpreter (it needs Python <=3.11) if .venv-ancestry exists, else in-process."""
+    import json as _json
+    import subprocess
+    code = ("import json,sys\nfrom snps import SNPs\nr=SNPs(sys.argv[1]).predict_ancestry()\n"
+            "print(json.dumps({k:v for k,v in r.items() if k!='ezancestry_df'}, default=str))")
+    venv_py = ancestry_python()
+    py = str(venv_py) if venv_py else sys.executable
+    try:
+        p = subprocess.run([py, "-c", code, str(inp)], capture_output=True, text=True, timeout=1800)
+        if p.returncode != 0:
+            raise RuntimeError(p.stderr.strip().splitlines()[-1] if p.stderr.strip() else "unknown error")
+        return _json.loads(p.stdout.strip().splitlines()[-1])
+    except Exception as e:  # noqa: BLE001
+        loaded.notes.append(f"ezancestry failed: {e}")
+        return None
+
+
+def main(argv=None):
+    argv = list(sys.argv[1:] if argv is None else argv)
+    if argv[:1] == ["serve"]:
+        from .mcp_server import main as serve
+        return serve()
+    if argv[:1] == ["setup"]:
+        from .setup_cmd import run as setup
+        return setup(argv[1:])
+    ap = argparse.ArgumentParser(prog="gnomancer", description="Analyze a consumer raw DNA file locally. Subcommands: `gnomancer setup` (download tools), `gnomancer serve` (MCP server).")
+    ap.add_argument("input", help="raw data file (.txt/.csv/.tsv/.vcf, optionally .zip/.gz)")
+    ap.add_argument("--out", default=None, help="output directory (default: <input>_analysis)")
+    ap.add_argument("--markers", default=str(DATA / "markers.tsv"), help="marker panel TSV")
+    ap.add_argument("--no-y", action="store_true", help="skip Y haplogroup")
+    ap.add_argument("--no-mt", action="store_true", help="skip mtDNA haplogroup")
+    ap.add_argument("--ancestry", action="store_true", help="run ezancestry population prediction (uses .venv-ancestry if present)")
+    ap.add_argument("--no-clinvar", action="store_true", help="skip the ClinVar scan")
+    ap.add_argument("--no-yleaf", action="store_true", help="skip the Yleaf second-opinion Y call")
+    ap.add_argument("--no-pharmcat", action="store_true", help="skip PharmCAT")
+    ap.add_argument("--summary", default=None, help="Markdown/text file with a narrative summary (e.g. written by your AI agent) to show at the top of the report")
+    ap.add_argument("--ai", action="store_true", help="have Claude write the narrative summary (sends the structured results, not the raw file, to the Anthropic API; needs credentials)")
+    ap.add_argument("--version", action="version", version=__version__)
+    a = ap.parse_args(argv)
+    run_args = argv
+
+    inp = Path(a.input)
+    out = Path(a.out) if a.out else inp.with_name(inp.stem.replace(".", "_") + "_analysis")
+    out.mkdir(parents=True, exist_ok=True)
+
+    print(f"[1/6] loading {inp} ...", file=sys.stderr)
+    loaded = load(str(inp))
+    df = loaded.df
+    print(f"      {loaded.source}, build {loaded.build}, {loaded.count:,} SNPs, sex {loaded.sex or '?'}", file=sys.stderr)
+
+    print("[2/6] quality metrics ...", file=sys.stderr)
+    ct = chromosome_table(df)
+    qc = summary(df, ct, loaded.sex)
+    ct.to_csv(out / "chromosome-qc.tsv", sep="\t", index=False)
+
+    print("[3/6] runs of homozygosity ...", file=sys.stderr)
+    roh = find_roh(df)
+    roh.to_csv(out / "roh.tsv", sep="\t", index=False)
+    roh_text = interpret(roh)
+
+    print("[4/6] marker panel ...", file=sys.stderr)
+    markers = load_markers(Path(a.markers))
+    panel = evaluate(df, markers)
+    panel.to_csv(out / "marker-panel.tsv", sep="\t", index=False)
+    combo = combos(df)
+
+    is_male = qc["sex_inferred"] == "Male"
+    y = dict(error="skipped (sample is not male)" if not is_male else "skipped (--no-y)")
+    if is_male and not a.no_y:
+        print("[5/6] Y haplogroup (yhaplo) ...", file=sys.stderr)
+        y = call_y(df, out)
+        if not a.no_yleaf:
+            print("      second opinion on current trees (Yleaf) ...", file=sys.stderr)
+            y["yleaf"] = yleaf_mod.run(df, out)
+    mt = dict(error="skipped (--no-mt)")
+    if not a.no_mt:
+        print("[5/6] mtDNA haplogroup (haplogrep3) ...", file=sys.stderr)
+        mt = call_mt(df, out, loaded.build_original)
+
+    cv = dict(available=False, hits=None, note="skipped (--no-clinvar)")
+    if not a.no_clinvar:
+        print("[6/8] ClinVar scan of every typed site ...", file=sys.stderr)
+        cv = clinvar_mod.scan(df)
+        if cv["available"] and cv["hits"] is not None and not cv["hits"].empty:
+            cv["hits"].to_csv(out / "clinvar-hits.tsv", sep="\t", index=False)
+    cv_text = clinvar_mod.interpret(cv) if not a.no_clinvar else cv["note"]
+    pc = dict(available=False, calls=None, error="skipped (--no-pharmcat)", stats={})
+    if not a.no_pharmcat:
+        print("[7/8] PharmCAT pharmacogenomics ...", file=sys.stderr)
+        pc = pharmcat_mod.run(df, out)
+        if pc.get("calls") is not None and not pc["calls"].empty:
+            pc["calls"].to_csv(out / "pharmcat-calls.tsv", sep="\t", index=False)
+    pc_text = pharmcat_mod.interpret(pc) if not a.no_pharmcat else pc["error"]
+    ancestry = None
+    if a.ancestry:
+        print("[8/8] ancestry prediction (ezancestry) ...", file=sys.stderr)
+        ancestry = run_ezancestry(inp, loaded)
+
+    glance = summary_mod.at_a_glance(loaded, qc, roh, roh_text, combo, y, mt, ancestry, panel, cv, pc)
+    prompt = summary_mod.build_prompt(glance, combo, y, mt, ancestry, panel, cv, pc, qc)
+    summary_mod.write_prompt_file(out, prompt)
+    narrative = None
+    if a.summary:
+        narrative = Path(a.summary).read_text().strip()
+    elif a.ai:
+        print("      narrative summary (Claude) ...", file=sys.stderr)
+        try:
+            narrative = summary_mod.ai_summary(prompt)
+            (out / "summary.md").write_text(narrative + "\n")
+        except Exception as e:  # noqa: BLE001
+            loaded.notes.append(f"AI summary failed: {e}")
+    extra = dict(clinvar=cv, clinvar_text=cv_text, pharmcat=pc, pharmcat_text=pc_text, glance=glance, narrative=narrative)
+    rep = write(out / "REPORT.md", loaded, qc, ct, roh, roh_text, panel, combo, y, mt, ancestry, extra)
+    html = write_html(out / "REPORT.html", loaded, qc, ct, roh, roh_text, panel, combo, y, mt, ancestry, input_name=inp.name, extra=extra)
+    json.dump(dict(input=str(inp.resolve()), args=[str(Path(x).resolve()) if Path(x).exists() else x for x in run_args],
+                   report_html=str((out / "REPORT.html").resolve()), report_md=str((out / "REPORT.md").resolve()), glance=glance,
+                   source=loaded.source, build=loaded.build, count=loaded.count, qc=qc,
+                   roh_total_mb=float(roh.length_mb.sum()) if not roh.empty else 0.0, y=y, mt=mt,
+                   combos=combo, ancestry=ancestry, notes=loaded.notes,
+                   clinvar=dict(sites_in_clinvar=cv.get("sites_in_clinvar"), pathogenic_sites_typed=cv.get("pathogenic_sites_typed"),
+                                hits=(cv["hits"].to_dict("records") if cv.get("hits") is not None and not cv["hits"].empty else [])),
+                   pharmcat=dict(stats={k: v for k, v in pc.get("stats", {}).items() if k != "genes"}, error=pc.get("error"),
+                                 calls=(pc["calls"].to_dict("records") if pc.get("calls") is not None and not pc["calls"].empty else []))),
+              open(out / "summary.json", "w"), indent=1, default=str)
+    print(f"\nReport: {html}\n        {rep}", file=sys.stderr)
+    if y.get("haplogroup"):
+        print(f"Y: {y['haplogroup']}   ", end="", file=sys.stderr)
+    if mt.get("haplogroup"):
+        print(f"MT: {mt['haplogroup']}", end="", file=sys.stderr)
+    print(file=sys.stderr)

gnomancer/clinvar.py

@@ -0,0 +1,133 @@
+"""Scan every typed site against ClinVar (GRCh37 VCF) and report pathogenic / likely pathogenic alleles carried.
+
+Why this matters: the curated panel covers ~110 sites; consumer chips type several thousand sites that ClinVar
+classifies. This finds the rest. Why to be careful: arrays have a high false-positive rate at rare sites
+(Tandy-Connor et al. 2018, Genetics in Medicine, doi:10.1038/gim.2018.38: 40% of variants from consumer raw data
+were false positives on clinical confirmation), so every hit here is "worth confirming with a clinical test",
+never a diagnosis.
+"""
+from __future__ import annotations
+
+import gzip
+from pathlib import Path
+
+import pandas as pd
+
+from .paths import tool
+
+DEFAULT_VCF = tool("clinvar", "clinvar_GRCh37.vcf.gz")
+PATHOGENIC = ("Pathogenic", "Likely_pathogenic")
+STAR = {"practice_guideline": 4, "reviewed_by_expert_panel": 3, "criteria_provided,_multiple_submitters,_no_conflicts": 2,
+        "criteria_provided,_single_submitter": 1, "criteria_provided,_conflicting_classifications": 1,
+        "criteria_provided,_conflicting_interpretations": 1, "no_assertion_criteria_provided": 0, "no_classification_provided": 0}
+
+
+def _conditions(raw: str, limit: int = 4) -> str:
+    """'Disease_A|not_provided|Disease_B' -> 'Disease A; Disease B' (deduplicated, generic entries dropped, at most `limit`)."""
+    seen, out = set(), []
+    for c in raw.replace("_", " ").split("|"):
+        c = c.strip()
+        if not c or c.lower() in ("not provided", "not specified", "see cases") or c in seen:
+            continue
+        seen.add(c)
+        out.append(c)
+    return "; ".join(out[:limit]) + (f" (+{len(out) - limit} more)" if len(out) > limit else "")
+
+
+def _info(s: str) -> dict:
+    d = {}
+    for kv in s.split(";"):
+        if "=" in kv:
+            k, v = kv.split("=", 1)
+            d[k] = v
+    return d
+
+
+def scan(df: pd.DataFrame, vcf: Path = DEFAULT_VCF) -> dict:
+    """df: normalized SNP table (chrom, pos, genotype). Returns dict with hits table and coverage stats."""
+    res = dict(available=vcf.exists(), vcf=str(vcf), sites_in_clinvar=0, pathogenic_sites_typed=0, hits=pd.DataFrame(), note="")
+    if not vcf.exists():
+        res["note"] = "ClinVar VCF not found; run `gnomancer setup` (downloads ~200 MB from NCBI)."
+        return res
+    called = df[df.genotype.notna()]
+    key = {}
+    for chrom, pos, gt in zip(called.chrom, called.pos, called.genotype):
+        key[(chrom, int(pos))] = gt
+    hits = []
+    n_in = n_path = 0
+    with gzip.open(vcf, "rt") as f:
+        for line in f:
+            if line[0] == "#":
+                continue
+            p = line.rstrip("\n").split("\t", 8)
+            chrom = p[0]
+            if chrom == "MT":
+                chrom = "MT"
+            k = (chrom, int(p[1]))
+            gt = key.get(k)
+            if gt is None:
+                continue
+            n_in += 1
+            info = _info(p[7])
+            sig = info.get("CLNSIG", "")
+            if not any(s in sig for s in PATHOGENIC) or sig.startswith("Benign") or sig.startswith("Likely_benign"):
+                continue
+            n_path += 1
+            ref, alts = p[3], p[4].split(",")
+            if len(ref) != 1 or any(len(a) != 1 for a in alts):
+                continue  # indels are not reliably comparable to array I/D calls
+            carried = [a for a in gt if a in alts]
+            if not carried:
+                continue
+            if any(a not in alts and a != ref for a in gt):
+                continue  # allele mismatch (likely strand issue); do not report
+            zyg = "homozygous" if len(gt) == 2 and gt[0] == gt[1] else ("hemizygous" if len(gt) == 1 else "heterozygous")
+            rev = info.get("CLNREVSTAT", "")
+            # Allele frequency of the ALT allele, from whichever population resource ClinVar carries.
+            af = None
+            for k_ in ("AF_EXAC", "AF_TGP", "AF_ESP"):
+                if k_ in info:
+                    try:
+                        af = max(float(x) for x in info[k_].split(","))
+                    except ValueError:
+                        pass
+                    break
+            flags = []
+            ambiguous = {ref, alts[0]} in ({"A", "T"}, {"C", "G"})
+            if ambiguous:
+                flags.append("strand-ambiguous SNP (A/T or C/G): a homozygous-reference call reported on the other strand looks exactly like this")
+            if af is not None and af >= 0.05:
+                flags.append(f"ALT allele is common (AF {af:.1%}): classification is likely outdated or low-penetrance")
+            if zyg != "heterozygous" and (af is None or af < 0.01):
+                flags.append("homozygous for a rare allele: statistically far more likely a genotyping/strand artifact than a true homozygote")
+            if "conflicting" in rev:
+                flags.append("submitters disagree on classification")
+            confidence = "low" if (ambiguous or (zyg != "heterozygous" and (af is None or af < 0.01)) or STAR.get(rev, 0) == 0) else ("medium" if (af or 0) >= 0.05 else "high")
+            hits.append(dict(rsid=("rs" + info["RS"]) if info.get("RS") else p[2], chrom=chrom, pos=k[1], genotype=gt,
+                             ref=ref, alt=",".join(alts), zygosity=zyg, gene=info.get("GENEINFO", "").split(":")[0],
+                             significance=sig.replace("_", " "), review=rev.replace("_", " "), stars=STAR.get(rev, 0),
+                             condition=_conditions(info.get("CLNDN", "")),
+                             variant=info.get("CLNHGVS", ""), alt_af=af, confidence=confidence, flags="; ".join(flags)))
+    res["sites_in_clinvar"], res["pathogenic_sites_typed"] = n_in, n_path
+    out = pd.DataFrame(hits)
+    if not out.empty:
+        out["conf_rank"] = out.confidence.map({"high": 0, "medium": 1, "low": 2})
+        out = out.sort_values(["conf_rank", "stars"], ascending=[True, False]).drop(columns="conf_rank")
+    res["hits"] = out
+    return res
+
+
+def interpret(res: dict) -> str:
+    if not res["available"]:
+        return res["note"]
+    h = res["hits"]
+    base = (f"{res['sites_in_clinvar']:,} typed sites have a ClinVar record; {res['pathogenic_sites_typed']:,} of those are "
+            f"classified pathogenic or likely pathogenic for at least one submitter. ")
+    if h.empty:
+        return base + "No pathogenic or likely-pathogenic alternate allele was called at any of them."
+    hi = h[h.confidence == "high"]
+    return base + (f"{len(h)} site(s) carry a pathogenic/likely-pathogenic allele; {len(hi)} pass the plausibility checks "
+                   "(heterozygous or common, unambiguous strand, reviewed). The rest are flagged as probable artifacts. "
+                   "Array calls at rare sites are often wrong: when a clinical lab re-tested variants reported in consumer raw data, "
+                   "40% were false positives (Tandy-Connor et al., Genetics in Medicine 2018, doi:10.1038/gim.2018.38). Treat every row as "
+                   "'confirm with a clinical test', not as a diagnosis. Heterozygous hits for recessive conditions are carrier states.")

gnomancer/context.py

@@ -0,0 +1,32 @@
+"""Short, population-level context for Y and mtDNA haplogroups (data/haplogroup_context.tsv)."""
+from __future__ import annotations
+
+import csv
+from pathlib import Path
+
+from .paths import DATA as _DATA
+
+DATA = _DATA / "haplogroup_context.tsv"
+
+
+def _rows():
+    with open(DATA, newline="") as f:
+        lines = [l for l in f if l.strip() and not l.startswith("#")]
+    return list(csv.DictReader(lines, delimiter="\t"))
+
+
+def lookup(kind: str, *labels: str) -> list[dict]:
+    """Return context rows for a haplogroup, most specific first.
+
+    kind: 'Y' or 'MT'. labels: any names for the call (e.g. short 'R-M269' and long 'R1b1a2a1a2').
+    Matching is by prefix on each label; results are the chain of matching prefixes, longest first,
+    so a report can show the lineage from broad to specific.
+    """
+    hits = {}
+    for r in _rows():
+        if r["type"] != kind:
+            continue
+        for lab in labels:
+            if lab and (lab == r["prefix"] or lab.startswith(r["prefix"])):
+                hits.setdefault(r["name"], r)  # one entry per lineage name, whichever prefix matched first
+    return sorted(hits.values(), key=lambda r: -int(r.get("depth") or len(r["prefix"])))

gnomancer/data/haplogroup_context.tsv

@@ -0,0 +1,72 @@
+# Population-level context for haplogroups. Matched by prefix against the caller's labels
+# (Y: ISOGG 2016 long form as emitted by yhaplo, and the short SNP form; MT: PhyloTree 17 name).
+# Keep entries short, hedged, and about populations, never individuals.
+type	prefix	depth	name	text
+Y	A	1	Haplogroup A	The deepest branches of the human Y tree, found mainly in Southern and Eastern Africa (Khoisan, Ethiopian and Sudanese groups).
+Y	B	1	Haplogroup B	An old African lineage, most common among Central African hunter-gatherer groups and in East Africa.
+Y	C	1	Haplogroup C	Spread across Asia, Oceania and the Americas: C-M217 is common in Mongolia, Siberia and East Asia; C-M38 in New Guinea and Melanesia; C-M130 branches reach Australia.
+Y	D	1	Haplogroup D	Found mainly in Japan (D-M55, including Ainu), Tibet and the Andaman Islands.
+Y	E	1	Haplogroup E	The dominant African paternal lineage, with branches that also spread into the Near East and around the Mediterranean.
+Y	E1b1a	2	E-M2	The lineage carried by the Bantu expansion: the majority Y lineage across West, Central and Southern Africa and in the African diaspora.
+Y	E1b1b	2	E-M215 / E-M35	Originated in East Africa and spread to North Africa, the Near East and Southern Europe; a major lineage in Ethiopia, Somalia, the Maghreb, Greece and the Balkans.
+Y	E1b1b1a	3	E-M78	Common from Egypt and Sudan through the Balkans and Southern Europe; E-V13 is a signature Balkan branch.
+Y	E1b1b1b1	3	E-M81	The characteristic Berber lineage of the Maghreb, also present in Iberia.
+Y	E1b1b1b2	3	E-M123 / E-M34	A Near-Eastern and East-Mediterranean branch: frequent in Ethiopia, the Levant, Anatolia and Arabia, a well-known component (around a tenth) of Ashkenazi, Sephardic and Mizrahi Jewish paternal lineages, and present at low frequency in Italy, Iberia and the Balkans.
+Y	E1b1b1b2a1a1	4	E-M84	The main sub-branch of E-M34 (YFull: E-M84). Frequent in Ethiopia and Eritrea, the Levant, Arabia and Anatolia; one of the characteristic lineages of Jewish communities from all three main diasporas; sporadic in Southern Europe.
+Y	E-M84	4	E-M84	The main sub-branch of E-M34. Frequent in Ethiopia and Eritrea, the Levant, Arabia and Anatolia; one of the characteristic lineages of Jewish communities from all three main diasporas; sporadic in Southern Europe.
+Y	E-L795	5	E-L795	A branch within E-M84 (FamilyTreeDNA naming), found in the Levant, Arabia and among Jewish lineages.
+Y	G	1	Haplogroup G	Peaks in the Caucasus and Anatolia; carried by early Neolithic farmers into Europe (the Tyrolean Iceman was G2a), today scattered at low frequency across Europe and the Near East.
+Y	H	1	Haplogroup H	Mostly South Asian; H1a1 is the main lineage of the Roma, reflecting their Indian origin.
+Y	I	1	Haplogroup I	A European lineage present since the Palaeolithic, with little presence outside Europe.
+Y	I1	2	I-M253	Concentrated in Scandinavia and Germanic-speaking regions; strongly associated with Norse and Anglo-Saxon migrations.
+Y	I2	2	I-M438	I2a1 peaks in the Balkans and Sardinia; I2a2 is found across Northwest Europe including Germany, Britain and Ireland.
+Y	J	1	Haplogroup J	Originated in the Near East and spread with Neolithic farming and later movements across the Mediterranean, Caucasus and South Asia.
+Y	J1	2	J-M267	Most frequent in Arabia, the Levant and the Caucasus (Dagestan); the J-P58 branch is associated with Semitic-speaking populations.
+Y	J2	2	J-M172	Peaks in the Fertile Crescent, Anatolia, the Caucasus and Iran; common around the Mediterranean (Italy, Greece, Balkans) and in South Asia.
+Y	L	1	Haplogroup L	Mainly South Asian, especially Pakistan and western India, with pockets in the Near East.
+Y	N	1	Haplogroup N	Spread from Siberia westward with Uralic speakers; the main lineage in Finland, the Baltic states and northern Russia.
+Y	O	1	Haplogroup O	The dominant lineage of East and Southeast Asia: O-M122 (Han Chinese and neighbours), O-M119 (coastal China, Taiwan, Austronesians), O-M95 (Southeast Asia).
+Y	Q	1	Haplogroup Q	Siberian in origin; Q-M3 is the principal Native American paternal lineage, and other branches are found in Siberia and Central Asia.
+Y	R	1	Haplogroup R	A widespread Eurasian lineage whose main branches dominate Europe, Central Asia and South Asia.
+Y	R1a	2	R-M420	Most common in Eastern Europe (Poland, Russia, Ukraine), Central Asia and northern India; associated with Bronze Age steppe migrations.
+Y	R1b	2	R-M343	The most common paternal lineage in Western Europe (peaking in Ireland, Wales, the Basque Country); the R-V88 branch is found in Central Africa.
+Y	R1b1a2	3	R-M269	The main Western European branch, carried by Bronze Age migrations; R-U106 is typical of Germanic regions, R-P312 and R-L21 of Atlantic Europe, R-U152 of the Alpine region and Italy.
+Y	R2	2	R-M479	Mainly South Asian.
+Y	T	1	Haplogroup T	A scattered lineage with pockets in the Near East, East Africa (Somalia), Sardinia and parts of Iberia.
+MT	L0	2	Haplogroup L0	Among the oldest maternal lineages, concentrated in Southern Africa (Khoisan) and East Africa.
+MT	L1	2	Haplogroup L1	A deep African lineage found mainly in Central and West Africa.
+MT	L2	2	Haplogroup L2	Widespread across Sub-Saharan Africa and the most common maternal lineage among African Americans.
+MT	L3	2	Haplogroup L3	East African in origin and the ancestor of every non-African maternal lineage; also common across Africa.
+MT	M	1	Haplogroup M	One of the two founding non-African lineages, now concentrated in South Asia, East Asia and among Native Americans (via branches C and D).
+MT	D	1	Haplogroup D	An East Asian and Siberian lineage; D1 and D4h3 are Native American founder lineages.
+MT	C	1	Haplogroup C	Found across Siberia, Central Asia and East Asia, and a Native American founder lineage (C1).
+MT	N	1	Haplogroup N	The other founding non-African lineage; its branches cover Europe, the Near East, Asia and the Americas.
+MT	A	1	Haplogroup A	East Asian and Siberian, and one of the main Native American founder lineages (A2).
+MT	B	1	Haplogroup B	East and Southeast Asian; B4 spread with Austronesians to Polynesia and B2 is a Native American founder lineage.
+MT	I	1	Haplogroup I	A minor European and Near Eastern lineage at a few percent frequency.
+MT	W	1	Haplogroup W	A minor lineage of Europe, the Near East and South Asia.
+MT	X	1	Haplogroup X	A rare lineage in Europe and the Near East; the X2a branch is a Native American founder lineage.
+MT	R	1	Haplogroup R	A large branch of N whose descendants include most European lineages.
+MT	F	1	Haplogroup F	Common in Southeast Asia and southern China.
+MT	U	1	Haplogroup U	An old West Eurasian lineage with branches in Europe, the Near East, North Africa and South Asia.
+MT	U5	2	Haplogroup U5	The characteristic lineage of European hunter-gatherers; today most frequent in Finland and the Saami.
+MT	U6	2	Haplogroup U6	Mainly North African (Berber populations) and Iberian.
+MT	K	1	Haplogroup K	A branch of U8 common across Europe and the Near East; three K sub-lineages account for about a third of Ashkenazi maternal lines.
+MT	HV	1	Haplogroup HV	The parent of H and V, found across Europe, the Near East and the Caucasus.
+MT	H	1	Haplogroup H	The most common maternal lineage in Europe (around 40%), also frequent in the Near East and North Africa; it spread widely after the last Ice Age.
+MT	H1	2	Haplogroup H1	The largest H branch and the most common maternal lineage in Western Europe; it peaks in Iberia and the Basque Country and is also found in North Africa and the Near East.
+MT	H2	2	Haplogroup H2	Found across Europe and the Caucasus; H2a is more frequent in Eastern Europe.
+MT	H3	2	Haplogroup H3	A Western European branch with a peak in Iberia and Western France.
+MT	H5	2	Haplogroup H5	Found across Europe with higher frequency in the Balkans and Central Europe.
+MT	H13	2	Haplogroup H13	Most frequent in the Caucasus and Near East, with a presence in Sardinia.
+MT	V	1	Haplogroup V	A sister of H found across Europe, with high frequency among the Saami, in Iberia and the Basque Country.
+MT	J	1	Haplogroup J	Spread from the Near East into Europe with Neolithic farmers; around a tenth of Europeans, higher in the Near East.
+MT	J1	2	Haplogroup J1	The main European J branch (J1c is the most common form in Europe).
+MT	J2	2	Haplogroup J2	Found around the Mediterranean and the Near East.
+MT	T	1	Haplogroup T	A Near Eastern and European lineage at about a tenth of the population in parts of Europe; associated with early farming and pastoralism.
+MT	T2	2	Haplogroup T2	The more common T branch in Europe, especially Northern and Central Europe.
+MT	N1	2	Haplogroup N1	Includes N1b, common among Ashkenazi Jews, and N1a, carried by early Neolithic farmers into Europe.
+MT	G	1	Haplogroup G	Mainly East Asian and Siberian, including Japan and Kamchatka.
+MT	Z	1	Haplogroup Z	Found in Siberia, Central Asia and among the Saami.
+MT	Q	1	Haplogroup Q	Found in New Guinea and Melanesia.
+MT	P	1	Haplogroup P	Found in Australia and New Guinea.

gnomancer/data/markers.tsv

@@ -0,0 +1,115 @@
+# Curated marker panel. Alleles are on the GRCh37 forward strand (what all major vendors report).
+# category: ancestry | trait | health | carrier | pgx.  effect = the allele being counted.
+# interp0/1/2 = text for 0/1/2 copies of the effect allele (optional).
+rsid	category	gene	label	ref	effect	interp0	interp1	interp2	note
+rs1426654	ancestry	SLC24A5	Skin pigmentation A111T	G	A	ancestral allele on both copies (typical of Sub-Saharan African and East Asian ancestry)	one light-skin allele	light-skin allele on both copies (near-fixed in Europe, West and South Asia)	
+rs16891982	ancestry	SLC45A2	Skin/hair pigmentation L374F	C	G	ancestral on both copies	one European light-pigmentation allele	European light-pigmentation allele on both copies	
+rs3827760	ancestry	EDAR	V370A hair thickness / shovel incisors	A	G	ancestral; no East Asian/Native American hair-thickness allele	one copy of the East Asian/Native American allele	homozygous for the East Asian/Native American allele	
+rs2814778	ancestry	ACKR1 (DARC)	Duffy null (FY*O)	T	C	Duffy-positive	one Duffy-null allele (common in West/Central African ancestry)	Duffy-null: resistant to P. vivax malaria; benign lower neutrophil counts	
+rs4988235	ancestry	LCT	Lactase persistence -13910	G	A	non-persistent at this site	persistent (dominant)	persistent	the European persistence allele; other populations use different alleles
+rs182549	ancestry	LCT	Lactase persistence -22018	C	T	non-persistent at this site	persistent (linked to -13910)	persistent	
+rs671	ancestry	ALDH2	ALDH2*2 alcohol flush	G	A	no flush allele	flush reaction, reduced acetaldehyde clearance (common in East Asia)	strong flush; alcohol poorly tolerated	
+rs1229984	ancestry	ADH1B	ADH1B*2 (His48)	C	T	*1/*1 typical of Europe and Africa	one fast-metabolizing *2 allele	*2/*2 fast alcohol metabolism (common in East Asia)	
+rs12913832	trait	HERC2	Eye color (main determinant)	A	G	brown eyes likely	brown or hazel likely, green possible	blue or grey eyes likely	
+rs1800407	trait	OCA2	Eye color modifier R419Q	C	T	no modifier	green/hazel modifier present	homozygous modifier	
+rs12896399	trait	SLC24A4	Eye/hair color	G	T	darker-associated	intermediate	lighter-associated	
+rs1042602	trait	TYR	S192Y pigmentation / freckling	C	A	no lighter allele	one lighter/freckling allele	lighter skin, freckling more likely	
+rs1393350	trait	TYR	Eye/hair color	G	A	darker-associated	intermediate	lighter-associated	
+rs12203592	trait	IRF4	Freckles, lighter hair, sun sensitivity	C	T	no freckle allele	one freckle / lighter-hair allele	freckles and sun sensitivity likely	
+rs1805007	trait	MC1R	R151C red hair	C	T	no variant	red-hair carrier (R allele)	red hair likely	
+rs1805008	trait	MC1R	R160W red hair	C	T	no variant	red-hair carrier (R allele)	red hair likely	
+rs1805009	trait	MC1R	D294H red hair	G	C	no variant	red-hair carrier (R allele)	red hair likely	
+rs1805006	trait	MC1R	D84E red hair	C	A	no variant	red-hair carrier (R allele)	red hair likely	
+rs11547464	trait	MC1R	R142H red hair	G	A	no variant	red-hair carrier (R allele)	red hair likely	
+rs1110400	trait	MC1R	I155T red hair	T	C	no variant	red-hair carrier (R allele)	red hair likely	
+rs885479	trait	MC1R	R163Q (weak variant)	G	A	no variant	one weak r allele	homozygous weak variant (common in East Asia)	
+rs11803731	trait	TCHH	Hair curl (European)	T	A	curlier-associated	intermediate	straight-associated	
+rs17822931	trait	ABCC11	Earwax type / underarm odor	C	T	wet earwax, typical body odor	wet earwax (C is dominant)	dry earwax, little underarm odor	
+rs4481887	trait	OR2M7 region	Asparagus metabolite smell	G	A	less likely to smell it	intermediate	likely smells asparagus metabolites	
+rs10427255	trait	ZEB2 region	Photic sneeze reflex	T	C	less likely	intermediate	more likely to sneeze in bright light	
+rs1815739	trait	ACTN3	R577X sprint/power muscle	C	T	RR: two working copies, power-leaning	RX: mixed	XX: no alpha-actinin-3, endurance-leaning (~18% of Europeans)	
+rs762551	trait	CYP1A2	Caffeine metabolism *1F	A	C	AA: fast caffeine metabolizer	intermediate	slow caffeine metabolizer	
+rs2472297	trait	CYP1A1/2 region	Caffeine intake tendency	C	T	lower-intake alleles	intermediate	higher-intake alleles	
+rs4410790	trait	AHR	Caffeine intake tendency	T	C	lower-intake alleles	intermediate	higher-intake alleles	
+rs713598	trait	TAS2R38	Bitter taste PTC (A49P)	C	G	non-taster allele on both copies	taster	taster	see combined TAS2R38 reading
+rs1726866	trait	TAS2R38	Bitter taste (V262A)	A	G	non-taster allele on both copies	taster	taster	see combined TAS2R38 reading
+rs10246939	trait	TAS2R38	Bitter taste (I296V)	T	C	non-taster allele on both copies	taster	taster	see combined TAS2R38 reading
+rs72921001	trait	OR6A2	Cilantro soapy taste	A	C	no soapy-taste allele	one soapy-taste allele	cilantro likely tastes soapy	
+rs9939609	trait	FTO	BMI/obesity risk	T	A	no risk allele	one risk allele (~+1.2 kg on average)	two risk alleles (~+3 kg on average)	
+rs17782313	trait	MC4R	BMI	T	C	no risk allele	one risk allele	two risk alleles	
+rs16969968	trait	CHRNA5	Nicotine dependence D398N	G	A	no risk allele	one risk allele: heavier smoking if a smoker	two risk alleles	
+rs4680	trait	COMT	Val158Met	G	A	Val/Val: higher COMT activity	Val/Met	Met/Met: lower COMT activity	popular but weak-effect marker
+rs6265	trait	BDNF	Val66Met	C	T	Val/Val (most common)	Val/Met	Met/Met	weak-effect marker
+rs1799971	trait	OPRM1	A118G	A	G	reference	one G allele	two G alleles	weak-effect marker
+rs601338	trait	FUT2	Secretor status W154X	G	A	secretor	secretor (one non-secretor allele)	non-secretor: resistant to common norovirus strains; altered gut flora	
+rs2180439	trait	PAX1 region	Male pattern hair loss	T	C	no risk allele	one risk allele	two risk alleles	
+rs53576	trait	OXTR	Oxytocin receptor	G	A	no variant allele	one variant allele	two variant alleles	weak-effect marker, popular online but poorly replicated
+rs1800497	trait	ANKK1/DRD2	TaqIA	G	A	A2/A2	A1/A2	A1/A1	weak-effect marker
+rs8176719	trait	ABO	O allele (261delG)	I	D	no O allele	one O allele	blood group O	see combined ABO reading
+rs8176746	trait	ABO	B allele	G	T	no B allele	one B allele	B/B	see combined ABO reading
+rs1050828	health	G6PD	G6PD A- (V68M)	C	T	no deficiency allele	carrier / affected if male	deficient	X-linked; hemizygous in males
+rs429358	health	APOE	e4-defining site (C112R)	T	C	no e4 allele	one e4 allele	two e4 alleles	see combined APOE reading
+rs7412	health	APOE	e2-defining site (R158C)	C	T	no e2 allele	one e2 allele	two e2 alleles	see combined APOE reading
+rs6025	health	F5	Factor V Leiden R506Q	C	T	negative	heterozygous: ~5x venous thrombosis risk; relevant for surgery, estrogen, long flights	homozygous: substantially elevated clotting risk	
+rs1799963	health	F2	Prothrombin G20210A	G	A	negative	heterozygous: ~2-3x venous thrombosis risk	homozygous	
+rs1800562	health	HFE	Hemochromatosis C282Y	G	A	negative	carrier; iron overload uncommon unless compound heterozygous with H63D	C282Y homozygous: high risk of iron overload, check ferritin/transferrin saturation	
+rs1799945	health	HFE	Hemochromatosis H63D	C	G	negative	carrier; little risk alone	H63D homozygous: mild iron overload possible	
+rs1800730	health	HFE	S65C	A	T	negative	carrier	homozygous	
+rs1801133	health	MTHFR	C677T (A222V)	G	A	normal activity	one T allele, ~65% activity, usually no consequence	TT: ~30% activity; mildly higher homocysteine if folate is low	widely over-interpreted online
+rs1801131	health	MTHFR	A1298C	T	G	normal	heterozygous, benign	homozygous, little effect alone	
+rs2187668	health	HLA-DQA1	DQ2.5 tag (celiac)	C	T	no DQ2.5	DQ2.5 present (most celiac patients carry it; most carriers never develop it)	DQ2.5 homozygous	
+rs7454108	health	HLA-DQB1	DQ8 tag (celiac)	T	C	no DQ8	DQ8 present	DQ8 homozygous	
+rs3135388	health	HLA-DRB1	DRB1*15:01 tag (multiple sclerosis)	G	A	non-carrier	carrier (~3x MS odds, still low absolute risk)	homozygous	
+rs2476601	health	PTPN22	R620W autoimmunity	G	A	negative	one risk allele (RA, T1D, thyroid autoimmunity)	two risk alleles	
+rs2066844	health	NOD2	R702W (Crohn's)	C	T	negative	carrier	homozygous	
+rs2066845	health	NOD2	G908R (Crohn's)	G	C	negative	carrier	homozygous	
+rs7903146	health	TCF7L2	Type 2 diabetes	C	T	no risk allele	one risk allele (~1.4x odds)	two risk alleles (~2x odds; ~10% of Europeans)	strongest common T2D variant
+rs1061170	health	CFH	Y402H macular degeneration	T	C	no risk allele	one risk allele (~2.5x AMD odds)	two risk alleles (~6x AMD odds)	smoking interacts strongly
+rs10490924	health	ARMS2	A69S macular degeneration	G	T	no risk allele	one risk allele	two risk alleles	
+rs1333049	health	CDKN2B-AS1 (9p21)	Coronary artery disease	G	C	no risk haplotype	one copy (~1.25x)	two copies (~1.5x)	
+rs10455872	health	LPA	High Lp(a) tag	A	G	no high-Lp(a) allele	elevated Lp(a) likely (~2x CAD odds); a direct Lp(a) blood test is cheap	homozygous	
+rs3798220	health	LPA	I4399M high Lp(a)	T	C	no high-Lp(a) allele	elevated Lp(a) likely	homozygous	
+rs11136000	health	CLU	Alzheimer's (protective T)	C	T	no protective allele	one protective allele (small effect)	two protective alleles	
+rs75932628	health	TREM2	R47H Alzheimer's	C	T	negative	carrier (~3x AD odds)	homozygous	
+rs76763715	health	GBA	N370S (Gaucher / Parkinson's)	T	C	negative	carrier: ~5x Parkinson's odds; Gaucher carrier	Gaucher type 1 likely	
+rs34637584	health	LRRK2	G2019S Parkinson's	G	A	negative	carrier: elevated Parkinson's risk (penetrance ~30% by 80)	homozygous	
+rs2802292	health	FOXO3	Longevity-associated	T	G	no longevity allele	one copy	two copies (modest association with reaching 90+)	
+rs1800629	health	TNF	-308 promoter	G	A	no variant allele	one variant allele	two variant alleles	weak-effect marker
+rs113993960	carrier	CFTR	deltaF508 cystic fibrosis	I	D	negative	CF carrier (most common CF allele)	affected	arrays type 1-3 CFTR sites of 2000+
+rs333	carrier	CCR5	delta32	I	D	negative	one delta32 copy	homozygous: resistant to R5-tropic HIV-1	
+rs80357906	carrier	BRCA1	185delAG (Ashkenazi founder)	I	D	negative at this site	pathogenic variant present: confirm clinically	homozygous	
+rs80359550	carrier	BRCA2	6174delT (Ashkenazi founder)	I	D	negative at this site	pathogenic variant present: confirm clinically	homozygous	
+rs11571833	carrier	BRCA2	K3326X	A	T	negative	modest cancer risk increase (not a classic BRCA2 pathogenic allele)	homozygous	
+rs334	carrier	HBB	Sickle cell HbS	T	A	negative	sickle cell trait	sickle cell disease	
+rs33930165	carrier	HBB	HbC	C	T	negative	HbC trait	HbC disease	
+rs33950507	carrier	HBB	HbE	C	T	negative	HbE trait	HbE	
+rs1799853	pgx	CYP2C9	*2 (R144C)	C	T	no *2	one *2: reduced warfarin/NSAID/phenytoin metabolism	*2/*2	
+rs1057910	pgx	CYP2C9	*3 (I359L)	A	C	no *3	one *3: markedly reduced metabolism	*3/*3 poor metabolizer	
+rs28371686	pgx	CYP2C9	*5	C	G	no *5	one *5 (African ancestry)	*5/*5	
+rs9332131	pgx	CYP2C9	*6	I	D	no *6	one *6	*6/*6	
+rs28371685	pgx	CYP2C9	*11	C	T	no *11	one *11	*11/*11	
+rs4244285	pgx	CYP2C19	*2 (loss of function)	G	A	no *2	one *2: intermediate metabolizer (clopidogrel less effective)	*2/*2 poor metabolizer	
+rs4986893	pgx	CYP2C19	*3 (loss of function)	G	A	no *3	one *3	*3/*3	mostly East Asian
+rs12248560	pgx	CYP2C19	*17 (increased function)	C	T	no *17	one *17: rapid metabolizer	*17/*17 ultrarapid	
+rs9923231	pgx	VKORC1	-1639G>A warfarin sensitivity	C	T	normal warfarin dose requirement	intermediate: lower warfarin dose	high sensitivity: low warfarin dose	
+rs9934438	pgx	VKORC1	1173C>T (proxy for -1639)	G	A	normal	intermediate	sensitive	in near-perfect LD with rs9923231
+rs2108622	pgx	CYP4F2	V433M	C	T	normal	slightly higher warfarin dose	higher warfarin dose	
+rs3892097	pgx	CYP2D6	*4 (null)	C	T	no *4	one *4 allele	*4/*4 poor metabolizer	CNV not detectable on arrays
+rs1065852	pgx	CYP2D6	*10 (reduced) / also on *4	G	A	no *10	one reduced-function allele	two reduced-function alleles	
+rs28371725	pgx	CYP2D6	*41 (reduced)	C	T	no *41	one *41	*41/*41	
+rs776746	pgx	CYP3A5	*3 vs *1 (expressor)	C	T	*3/*3 non-expressor (most Europeans; standard tacrolimus dosing)	*1/*3 expressor: higher tacrolimus dose	*1/*1 expressor	
+rs3745274	pgx	CYP2B6	*6 (516G>T)	G	T	*1/*1	one *6: slower efavirenz/bupropion clearance	*6/*6 poor metabolizer	
+rs4149056	pgx	SLCO1B1	*5 (V174A) statin myopathy	T	C	normal statin transport	one C: ~4x simvastatin myopathy risk	two C: ~17x; avoid high-dose simvastatin	
+rs2231142	pgx	ABCG2	Q141K	G	T	normal	reduced transport: higher rosuvastatin/allopurinol exposure	homozygous reduced	
+rs3918290	pgx	DPYD	*2A (IVS14+1G>A)	C	T	normal	DPD deficiency carrier: fluoropyrimidine (5-FU, capecitabine) toxicity risk, dose reduction needed	DPD deficient: avoid fluoropyrimidines	
+rs67376798	pgx	DPYD	D949V	T	A	normal	reduced DPD activity	homozygous	
+rs887829	pgx	UGT1A1	*28 proxy (Gilbert)	C	T	no *28	one *28	*28/*28 Gilbert syndrome: irinotecan toxicity, benign high bilirubin	
+rs4148323	pgx	UGT1A1	*6 (G71R)	G	A	no *6	one *6	*6/*6	mostly East Asian
+rs1801280	pgx	NAT2	*5 (I114T)	T	C	no *5	one slow allele	slow	combine with *6/*7
+rs1799930	pgx	NAT2	*6 (R197Q)	G	A	no *6	one slow allele	slow	
+rs1799931	pgx	NAT2	*7 (G286E)	G	A	no *7	one slow allele	slow	
+rs1495741	pgx	NAT2	acetylator tag	A	G	slow acetylator (tag)	intermediate	rapid acetylator (tag)	
+rs1800460	pgx	TPMT	*3B	C	T	no *3B	one *3B (thiopurine dose reduction)	*3B/*3B	
+rs1142345	pgx	TPMT	*3C	T	C	no *3C	one *3C (thiopurine dose reduction)	*3C/*3C	
+rs2395029	pgx	HCP5	HLA-B*57:01 tag	T	G	negative	HLA-B*57:01 carrier: abacavir hypersensitivity, avoid	homozygous	
+rs1800566	pgx	NQO1	P187S (*2)	G	A	normal	one null allele	NQO1 null