pgx.py

"""Pharmacogenomics & drug metabolism (CPIC / PharmGKB) — curated chip-callable markers."""
from .common import G, C, CHIPS_ALL, PGKB, GWAS, CLIN, CPIC


def register(add):
    add(
        "rs4244285", "CYP2C19", "CYP2C19*2 Loss-of-Function Variant (681G>A)", "10", 96541604,
        ["pharmacogenomics", "cardio"], "strong", "A", "G",
        "CYP2C19*2 is a splicing defect causing complete loss of enzyme activity. CYP2C19 metabolizes clopidogrel (Plavix, bioactivation required), SSRIs (citalopram, escitalopram, sertraline), and proton pump inhibitors (omeprazole, pantoprazole).",
        G("CYP2C19*2 Homozygote / Poor Metabolizer (AA)", "caution",
          "Two loss-of-function alleles (*2/*2). CYP2C19 Poor Metabolizer phenotype.",
          "CPIC Guideline Alert: Ineffective bioactivation of clopidogrel (Plavix) after coronary stenting—alternative antiplatelet therapy (ticagrelor/prasugrel) is recommended by CPIC. Slower clearance of escitalopram and omeprazole.",
          recover=["Critical: Discuss with your physician or pharmacist before taking clopidogrel (Plavix), PPIs, or SSRIs. Never alter medications independently."]),
        G("CYP2C19*2 Heterozygote / Intermediate Metabolizer (AG)", "caution",
          "One loss-of-function allele (*1/*2). CYP2C19 Intermediate Metabolizer phenotype.",
          "Reduced activation of clopidogrel and moderately reduced clearance of certain antidepressants and PPIs.",
          recover=["CPIC guidelines recommend evaluating antiplatelet choices with a physician if coronary interventions are planned."]),
        G("Typical CYP2C19 Function (GG)", "typical",
          "No *2 variant detected (*1/*1 unless other rare star alleles are present).",
          "Normal expected drug bioactivation and clearance for CYP2C19 substrates."),
        notes="A premier CPIC Level-1A pharmacogenomic marker.",
    )

    add(
        "rs4986893", "CYP2C19", "CYP2C19*3 Loss-of-Function Variant (636G>A)", "10", 96535281,
        ["pharmacogenomics"], "strong", "A", "G",
        "CYP2C19*3 creates a premature stop codon (Trp212Ter), abolishing enzyme activity. Prevalent in East Asian populations (~5–10%).",
        G("CYP2C19*3 Carrier / Reduced Metabolism (AG)", "caution",
          "One *3 loss-of-function allele. Intermediate or Poor metabolizer phenotype.",
          "CPIC Guidelines: Reduced clopidogrel activation and reduced clearance of diazepam, omeprazole, and citalopram.",
          recover=["Review with a clinical pharmacist prior to antiplatelet or psychiatric prescriptions."]),
        G("CYP2C19*3 Homozygote / Poor Metabolizer (AA)", "caution",
          "Two *3 alleles. Poor metabolizer phenotype.",
          "Substantially reduced clopidogrel bioactivation.",
          recover=["Consult your physician for CPIC-guided alternative antiplatelet selection."]),
        G("Typical (GG)", "typical", "No *3 variant detected.", "Normal baseline."),
    )

    add(
        "rs12248560", "CYP2C19", "CYP2C19*17 Gain-of-Function Variant (-806C>T)", "10", 96522463,
        ["pharmacogenomics"], "strong", "C", "T",
        "CYP2C19*17 is a promoter variant that increases gene transcription, resulting in Rapid (*1/*17) or Ultrarapid (*17/*17) metabolism.",
        G("CYP2C19*17 Homozygote / Ultrarapid Metabolizer (TT)", "notable",
          "Two *17 alleles. CYP2C19 Ultrarapid Metabolizer phenotype.",
          "Clears PPIs (omeprazole) and certain antidepressants (escitalopram, sertraline) much faster than normal, potentially leading to treatment failure at standard dosages.",
          recover=["Inform prescribing physicians if stomach acid or depression treatments fail to achieve expected clinical response."]),
        G("CYP2C19*17 Heterozygote / Rapid Metabolizer (CT)", "notable",
          "One *17 allele. Rapid Metabolizer phenotype.",
          "Moderately faster clearance of CYP2C19 substrates.",
          recover=["Standard clinical monitoring."]),
        G("Typical (CC)", "typical", "No *17 allele detected.", "Normal transcription."),
    )

    add(
        "rs1799853", "CYP2C9", "CYP2C9*2 Reduced-Function Variant (Arg144Cys)", "10", 96702047,
        ["pharmacogenomics", "cardio"], "strong", "C", "T",
        "CYP2C9 metabolizes warfarin (Coumadin), phenytoin (Dilantin), celecoxib (Celebrex), and NSAIDs (ibuprofen, naproxen). The *2 allele (T) reduces enzyme activity by ~30–50%.",
        G("CYP2C9*2 Carrier (CT)", "caution",
          "One *2 allele. CYP2C9 Intermediate Metabolizer.",
          "CPIC Guideline: Slower clearance of warfarin and NSAIDs. Warfarin dose requirements are typically lower; standard doses increase bleeding risk.",
          recover=["Alert your clinician if prescribed warfarin; genetic dose-prediction algorithms can prevent bleeding complications."]),
        G("CYP2C9*2 Homozygote (TT)", "caution",
          "Two *2 alleles. Significantly reduced CYP2C9 clearance.",
          "Markedly lower warfarin dose requirement and prolonged half-life of NSAIDs and sulfonylureas.",
          recover=["Clinical consultation essential for anticoagulant and anti-inflammatory dosing."]),
        G("Typical (CC)", "typical", "No *2 variant detected.", "Normal baseline clearance."),
    )

    add(
        "rs1057910", "CYP2C9", "CYP2C9*3 Loss-of-Function Variant (Ile359Leu)", "10", 96741053,
        ["pharmacogenomics", "cardio"], "strong", "A", "C",
        "The *3 allele (C) reduces CYP2C9 enzyme activity by ~80–90%, dramatically decreasing clearance of warfarin and celecoxib.",
        G("CYP2C9*3 Carrier (AC)", "caution",
          "One *3 allele. CYP2C9 Intermediate to Poor Metabolizer.",
          "CPIC Guideline: Substantially reduced warfarin metabolism, higher sensitivity, and increased bleeding hazard at standard starting doses.",
          recover=["Inform clinicians prior to warfarin, celecoxib, or phenytoin therapy."]),
        G("CYP2C9*3 Homozygote / Poor Metabolizer (CC)", "caution",
          "Two *3 alleles. Severe CYP2C9 deficiency.",
          "Extreme sensitivity to warfarin and NSAIDs.",
          recover=["Requires specialized clinical dosing algorithms."]),
        G("Typical (AA)", "typical", "No *3 variant detected.", "Normal clearance."),
    )

    add(
        "rs9923231", "VKORC1", "Vitamin K Epoxide Reductase (-1639G>A)", "16", 31107660,
        ["pharmacogenomics", "cardio", "vitamins"], "strong", "A", "G",
        "VKORC1 is the target of warfarin. The A allele (-1639A) reduces VKORC1 promoter activity and protein expression, conferring high warfarin sensitivity and requiring significantly lower maintenance doses.",
        G("High Warfarin Sensitivity (AA)", "caution",
          "Two A alleles. Low VKORC1 expression.",
          "CPIC Guideline: Highly sensitive to warfarin. Standard initial doses can cause severe over-anticoagulation and hemorrhage; typical required doses are 50–70% lower.",
          recover=["Crucial clinical info for anticoagulation management. Share with your prescribing doctor."]),
        G("Intermediate Warfarin Sensitivity (AG)", "notable",
          "One A allele. Moderate VKORC1 expression.",
          "Requires lower-than-average warfarin maintenance doses.",
          recover=["Standard clinical INR monitoring and dose titration."]),
        G("Standard Warfarin Sensitivity (GG)", "typical",
          "Two G alleles. Normal VKORC1 expression.",
          "Typical baseline warfarin dose requirements."),
        notes="Pairs directly with CYP2C9 to determine personalized warfarin dosing.",
    )

    add(
        "rs3892097", "CYP2D6", "CYP2D6*4 Splicing Defect (1846G>A)", "22", 42524947,
        ["pharmacogenomics"], "strong", "A", "G",
        "CYP2D6*4 (A) is the most common loss-of-function allele in Europeans. CYP2D6 metabolizes ~25% of all clinical drugs: codeine/tramadol (prodrug activation), tamoxifen, beta-blockers (metoprolol), and antidepressants (venlafaxine, TCAs).",
        G("CYP2D6*4 Homozygote / Poor Metabolizer (AA)", "caution",
          "Two *4 null alleles. CYP2D6 Poor Metabolizer phenotype.",
          "CPIC Guideline: Inability to convert codeine and tramadol into active morphine/M1 (lack of pain relief). Ineffective activation of tamoxifen for breast cancer. Slow clearance of metoprolol and TCAs (higher drug levels/side effects).",
          recover=["Essential medical note: Report CYP2D6 Poor Metabolizer status to your doctor when discussing pain medications, antidepressants, or oncology therapies."]),
        G("CYP2D6*4 Carrier / Intermediate Metabolizer (AG)", "notable",
          "One *4 null allele. CYP2D6 Intermediate Metabolizer.",
          "Reduced bioactivation of prodrugs and moderately slowed clearance of substrates.",
          recover=["Discuss appropriate pain management and beta-blocker dosing with a physician."]),
        G("Typical (GG)", "typical", "No *4 allele detected.", "Normal baseline enzyme function."),
    )

    add(
        "rs4149056", "SLCO1B1", "Organic Anion Transporter OATP1B1 (521T>C / *5)", "12", 21331549,
        ["pharmacogenomics", "cardio", "lipids"], "strong", "C", "T",
        "SLCO1B1 encodes the OATP1B1 hepatic uptake transporter. The C allele (521C / *5) reduces hepatic uptake of statins (especially simvastatin and atorvastatin), raising systemic plasma statin exposure and significantly increasing risk of statin-induced myopathy/rhabdomyolysis.",
        G("High Statin Myopathy Risk (CC)", "caution",
          "Two C alleles (*5/*5). Severely reduced hepatic statin uptake.",
          "CPIC Guideline: ~15–20x higher risk of statin-induced muscle pain and myopathy with simvastatin. CPIC recommends avoiding high-dose simvastatin and considering alternative statins (pravastatin, rosuvastatin) or lower doses.",
          recover=["Never stop a prescribed statin abruptly. Discuss SLCO1B1 status with your physician to select the safest statin type and dose."]),
        G("Moderate Statin Myopathy Risk (CT)", "caution",
          "One C allele (*1/*5). Moderately reduced hepatic statin uptake.",
          "CPIC Guideline: ~4–5x increased risk of simvastatin-induced muscle toxicity.",
          recover=["Inform your doctor if starting statin therapy; lower doses or alternative statins may be preferred."]),
        G("Typical Statin Uptake (TT)", "typical",
          "Two T alleles. Normal hepatic OATP1B1 transporter function.",
          "Standard baseline tolerance for statin medications."),
        notes="Premier CPIC Level-1A guideline for statin safety.",
    )

    add(
        "rs3918290", "DPYD", "Dihydropyrimidine Dehydrogenase (DPYD*2A / IVS14+1G>A)", "1", 97915614,
        ["pharmacogenomics"], "strong", "A", "G",
        "DPYD metabolizes fluoropyrimidine chemotherapies (5-Fluorouracil, Capecitabine). The *2A allele (A) causes complete loss of enzyme activity, leading to lethal drug toxicity if given standard chemotherapy doses.",
        G("Severe Fluoropyrimidine Toxicity Risk (AA)", "caution",
          "Two *2A null alleles. Complete DPYD deficiency.",
          "CPIC Guideline: Absolute contraindication for standard 5-FU / Capecitabine due to life-threatening myelosuppression and GI toxicity.",
          recover=["Critical oncology warning: Must be communicated to oncologist before any fluoropyrimidine chemotherapy."]),
        G("High Toxicity Risk Carrier (AG)", "caution",
          "One *2A null allele. DPYD Intermediate Metabolizer (~50% activity).",
          "CPIC Guideline: Requires 50% initial dose reduction of 5-FU/capecitabine with therapeutic drug monitoring to prevent severe toxicity.",
          recover=["Mandatory oncology consultation if considering chemotherapy."]),
        G("Typical DPYD Activity (GG)", "typical", "No *2A mutation detected.", "Normal fluoropyrimidine clearance."),
    )

    add(
        "rs1800460", "TPMT", "Thiopurine S-Methyltransferase (TPMT*3B / 460G>A)", "6", 18139228,
        ["pharmacogenomics", "immunity"], "strong", "A", "G",
        "TPMT inactivates thiopurine immunosuppressants (Azathioprine, 6-Mercaptopurine). Deficient alleles cause bone marrow suppression (severe leukopenia).",
        G("Typical TPMT (GG)", "typical", "Normal thiopurine inactivation.", "Standard dosing baseline."),
        G("Reduced TPMT Activity (AG)", "caution",
          "One non-functional allele. TPMT Intermediate Metabolizer.",
          "CPIC Guideline: Requires 30–70% reduction in azathioprine/6-MP starting dose to avoid life-threatening bone marrow toxicity.",
          recover=["Inform gastroenterologist, rheumatologist, or oncologist before thiopurine therapy."]),
        G("Severe TPMT Deficiency (AA)", "caution",
          "Two non-functional alleles. TPMT Poor Metabolizer.",
          "CPIC Guideline: Severe myelosuppression risk; requires 90% dose reduction or alternative agents.",
          recover=["Clinical consultation required prior to immunosuppressive thiopurine therapy."]),
    )

    add(
        "rs887829", "UGT1A1", "UDP-Glucuronosyltransferase 1A1 (UGT1A1*28 / *80 Proxy)", "2", 234674751,
        ["pharmacogenomics", "detox"], "strong", "C", "T",
        "UGT1A1 conjugates bilirubin and active drug metabolites (e.g. SN-38 from irinotecan chemotherapy). The T allele tags the *28 TA-repeat expansion in the promoter, causing Gilbert's syndrome (benign unconjugated hyperbilirubinemia) and irinotecan toxicity.",
        G("Typical UGT1A1 (CC)", "typical", "Standard glucuronidation rate.", "Normal bilirubin clearance."),
        G("Gilbert's Syndrome Carrier (CT)", "typical", "One *28 tag allele. Mildly reduced glucuronidation.",
          "Generally asymptomatic; occasional mild bilirubin elevations during fasting or illness."),
        G("Gilbert's Syndrome / Irinotecan Sensitivity (TT)", "notable",
          "Two *28 tag alleles. ~70% reduction in UGT1A1 promoter activity.",
          "Classic Gilbert's syndrome: benign mild jaundice during fasting, dehydration, or strenuous exercise. CPIC Alert: Increased toxicity risk with the cancer drug irinotecan.",
          eat=["Stay well-hydrated and avoid prolonged severe fasting if you experience yellowing of the eyes."]),
    )

    add(
        "rs2395029", "HCP5", "HLA-B*5701 Tag SNP (Abacavir Hypersensitivity)", "6", 31432296,
        ["pharmacogenomics", "immunity"], "strong", "G", "T",
        "rs2395029 is in near-perfect linkage disequilibrium with the HLA-B*5701 allele. HLA-B*5701 carriers suffer severe, potentially fatal multi-organ hypersensitivity reactions to the HIV antiretroviral drug Abacavir.",
        G("Typical (GG)", "typical", "Negative for HLA-B*5701 tag.", "Standard low risk for abacavir hypersensitivity."),
        G("HLA-B*5701 Positive Tag (GT)", "caution",
          "One T allele. Strong tag for HLA-B*5701.",
          "FDA / CPIC Black Box Warning: Abacavir is strictly contraindicated due to severe hypersensitivity reaction risk.",
          recover=["Inform medical providers: HLA-B*5701 positive tag. Never take Abacavir."]),
        G("HLA-B*5701 Homozygote Tag (TT)", "caution",
          "Two T alleles. Positive for HLA-B*5701.",
          "Abacavir is strictly contraindicated.",
          recover=["Mandatory medical documentation: Avoid Abacavir."]),
    )

    add(
        "rs762551", "CYP1A2", "Cytochrome P450 1A2 (Caffeine & Drug Clearance / *1F)", "15", 75041050,
        ["pharmacogenomics", "caffeine", "cardio", "sleep"], "strong", "A", "C",
        "CYP1A2 clears caffeine, theophylline, melatonin, and antipsychotics (olanzapine, clozapine). The A allele (*1F) confers high inducibility and fast clearance, whereas the C allele causes slow clearance.",
        G("Fast Caffeine & Drug Metabolizer (AA)", "advantageous",
          "Two A alleles (*1F/*1F). Highly inducible CYP1A2.",
          "Rapid caffeine clearance (half-life ~2.5–4 hours). Moderate coffee consumption is cardioprotective in fast metabolizers and unlikely to impair sleep if consumed early.",
          substances=["Can tolerate moderate morning coffee with rapid hepatic clearance."]),
        G("Intermediate Metabolizer (AC)", "typical",
          "One C allele. Intermediate clearance rate.",
          "Standard caffeine half-life (~5–6 hours).",
          substances=["Maintain an 8-hour caffeine cutoff before bedtime."]),
        G("Slow Caffeine Metabolizer (CC)", "caution",
          "Two C alleles. Slower CYP1A2 clearance (half-life ~7–10 hours).",
          "Caffeine lingers in circulation. High coffee intake (>2–3 cups/day) in slow metabolizers is associated with higher resting blood pressure and increased myocardial infarction risk.",
          substances=["Limit to 1 cup in the early morning or switch to green tea/decaf.", "Strict 10-hour caffeine cutoff before sleep."],
          sleep=["Caffeine significantly disrupts slow-wave deep sleep in slow metabolizers."]),
    )


    add(
        "rs1065852", "CYP2D6", "CYP2D6*10 Reduced Function (100C>T / Pro34Ser)", "22", 42526694,
        ["pharmacogenomics"], "strong", "C", "T",
        "CYP2D6*10 is the predominant reduced-function allele in East Asian populations (~40–50% allele frequency). It reduces enzymatic clearance of beta-blockers, antidepressants, and antiarrhythmics.",
        G("Typical (CC)", "typical", "No *10 allele.", "Standard enzyme function."),
        G("CYP2D6*10 Carrier / Intermediate Metabolizer (CT)", "notable", "One *10 allele. Moderately reduced CYP2D6 clearance.",
          "CPIC Guideline: Slower clearance of metoprolol, flecainide, and codeine.",
          recover=["Review with doctor when dosing CYP2D6-metabolized medications."]),
        G("CYP2D6*10 Homozygote / Intermediate Metabolizer (TT)", "caution", "Two *10 alleles. Significantly reduced CYP2D6 activity.",
          "Common in East Asia; standard doses of psychotropics or beta-blockers produce higher serum drug levels.",
          recover=["Discuss dose titration with a prescribing clinician."]),
    )

    add(
        "rs776746", "CYP3A5", "CYP3A5*3 Splicing Defect (6986A>G / Tacrolimus Metabolism)", "7", 99270539,
        ["pharmacogenomics", "immunity"], "strong", "A", "G",
        "CYP3A5 metabolizes calcineurin inhibitors (Tacrolimus) used in organ transplantation and autoimmune treatment. The G allele (*3) is a splicing defect causing complete lack of CYP3A5 expression (CYP3A5 Non-Expressor).",
        G("CYP3A5 Expressor / Rapid Clearance (AA)", "caution",
          "Two *1 functional alleles. High hepatic and intestinal CYP3A5 expression.",
          "CPIC Guideline: Requires 1.5–2x higher starting doses of tacrolimus to reach therapeutic blood levels.",
          recover=["Critical for immunosuppressive and transplant dosing protocols."]),
        G("CYP3A5 Intermediate Expressor (AG)", "notable", "One *1 allele. Moderate CYP3A5 expression.",
          "CPIC Guideline: Intermediate tacrolimus dose requirements.",
          recover=["Standard therapeutic drug monitoring."]),
        G("CYP3A5 Non-Expressor (GG / *3/*3)", "typical",
          "Two *3 null alleles. Standard Non-Expressor phenotype (~85% of Europeans).",
          "Standard baseline tacrolimus dosing requirements.",
          recover=["Standard clinical monitoring."]),
    )

    add(
        "rs1142345", "TPMT", "Thiopurine S-Methyltransferase (TPMT*3C / Tyr240Cys)", "6", 18130918,
        ["pharmacogenomics"], "strong", "A", "G",
        "TPMT*3C is the predominant thiopurine loss-of-function allele in East Asian and African populations.",
        G("Typical (AA)", "typical", "Normal TPMT activity.", "Standard thiopurine dosing."),
        G("TPMT*3C Carrier / Intermediate Metabolizer (AG)", "caution", "One *3C null allele.",
          "CPIC Guideline: 30–70% dose reduction of azathioprine/6-MP required.",
          recover=["Alert clinical team prior to autoimmune or leukemia thiopurine therapy."]),
        G("TPMT*3C Homozygote / Severe Deficiency (GG)", "caution", "Two *3C null alleles.",
          "CPIC Guideline: Severe thiopurine-induced bone marrow suppression risk.",
          recover=["Mandatory oncology/gastroenterology consultation."]),
    )

    add(
        "rs12979860", "IFNL3", "Interferon Lambda 3 (IL28B / Hepatitis C & Viral Clearance)", "19", 39738787,
        ["immunity", "pharmacogenomics"], "strong", "C", "T",
        "IFNL3 governs innate mucosal antiviral defense. The C allele is strongly associated with spontaneous clearance of Hepatitis C and superior response to pegylated interferon-alpha therapies.",
        G("Favorable Viral Clearance Allele (CC)", "advantageous",
          "Two C alleles. Robust interferon-lambda mucosal antiviral induction.",
          "High natural spontaneous viral clearance rates and optimal interferon response.",
          immunity=["Robust baseline mucosal innate immune defense."]),
        G("Intermediate (CT)", "typical", "One C allele. Balanced antiviral signaling.", "Standard viral response."),
        G("Unfavorable Antiviral Allele (TT)", "notable", "Two T alleles. Slower spontaneous viral clearance.",
          "Lower baseline innate mucosal interferon-lambda expression.",
          immunity=["Support mucosal immunity with vitamin D, zinc, and gut health."]),
    )

    # --- 2.0 expansion: warfarin triad completion, transporters, NAT2, DPYD/UGT alleles ---
    add(
        "rs2108622", "CYP4F2", "Vitamin K Oxidase (V433M) — Warfarin Triad", "19", 15990431,
        ["pharmacogenomics", "vitamins"], "strong", "C", "T",
        "CYP4F2 removes vitamin K from the cycle. The T (433M) allele reduces enzyme level, leaving more vitamin K available — carriers need ~1 mg/day MORE warfarin on average. Completes the VKORC1 + CYP2C9 warfarin triad.",
        G("Typical vitamin K oxidation (CC)", "typical", "Standard CYP4F2 activity.", "No warfarin-dose shift from this gene."),
        G("One 433M allele (CT)", "notable", "Modestly higher vitamin K availability.", "Associated with slightly higher warfarin dose requirement — strictly a note for the prescriber, never self-adjust.",
          recover=["If warfarin is ever prescribed, mention CYP4F2 V433M carrier status alongside VKORC1/CYP2C9."]),
        G("Two 433M alleles (TT)", "notable", "Higher vitamin K availability.", "~1 mg/day higher average warfarin requirement in European-ancestry studies. Also modestly shifts vitamin K/E handling."),
        tier="well-replicated", transferability="multi-ancestry",
        effect="~+0.5-1 mg/day warfarin per allele — actionable as part of dosing algorithms (CPIC)",
        citations=[C(18250228, "Caldwell 2008, Blood — CYP4F2 V433M and warfarin dose"),
                   C(CPIC, "CPIC warfarin guideline (triad algorithm)")],
    )
    add(
        "rs2231142", "ABCG2", "Efflux Transporter Q141K — Statins & Urate", "4", 89052323,
        ["pharmacogenomics", "metabolism"], "strong", "G", "T",
        "ABCG2 (BCRP) pumps drugs and urate out of cells. Q141K halves transport: rosuvastatin exposure rises ~40-70% per allele (CPIC-annotated), and serum urate rises — the strongest common gout locus.",
        G("Typical efflux (GG)", "typical", "Standard BCRP function.", "Standard rosuvastatin exposure and urate handling."),
        G("One 141K allele (GT)", "notable", "Reduced efflux.", "Higher rosuvastatin exposure (prescriber note) and ~0.2-0.3 mg/dL higher average urate — hydration and purine awareness if gout runs in the family.",
          eat=["If urate/gout is a theme: hydration, limit fructose-sweetened drinks and beer."]),
        G("Two 141K alleles (TT)", "caution", "Substantially reduced efflux.", "Rosuvastatin exposure roughly doubles (CPIC suggests dose awareness — prescriber's call) and gout OR ~2. Urate labs are cheap.",
          recover=["Mention ABCG2 Q141K to any prescriber of rosuvastatin/sulfasalazine; ask about a urate level at the next physical."]),
        tier="well-replicated", transferability="multi-ancestry",
        effect="Rosuvastatin AUC +40-70%/allele; urate +0.2 mg/dL/allele; gout OR ~1.7/allele",
        citations=[C(CPIC, "CPIC rosuvastatin (SLCO1B1/ABCG2) guideline 2022"),
                   C(18834626, "Dehghan 2008, Lancet — ABCG2 and gout")],
    )
    add(
        "rs116855232", "NUDT15", "Thiopurine Toxicity (R139C) — East-Asian Axis", "13", 48611934,
        ["pharmacogenomics", "immunity"], "strong", "C", "T",
        "NUDT15 R139C cripples degradation of active thiopurine metabolites. Intermediate metabolizers need reduced azathioprine/6-MP dosing; poor metabolizers face severe myelosuppression at standard doses. Common in East-Asian ancestry where TPMT variants are rare — the mirror image of the European TPMT story.",
        G("Normal NUDT15 (CC)", "typical", "Typical thiopurine inactivation at this gene.", "TPMT still applies — see its card."),
        G("Intermediate metabolizer (CT)", "caution", "One inactive copy (CPIC: reduced starting dose).",
          "If a thiopurine (azathioprine, 6-MP) is ever prescribed, this genotype belongs in the conversation BEFORE the first dose.",
          recover=["Carry this in your clinician notes: NUDT15 intermediate — CPIC recommends dose reduction for thiopurines."]),
        G("Poor metabolizer (TT)", "caution", "Both copies inactive — severe myelosuppression risk at standard doses.",
          "CPIC recommends drastically reduced dosing or alternatives. Rare but critical to know in advance.",
          recover=["Flag prominently to any prescriber of azathioprine/6-mercaptopurine/thioguanine."]),
        tier="well-replicated", transferability="east-asian-derived",
        effect="Leukopenia OR ~7 for het at standard dosing — large, actionable (CPIC A)",
        citations=[C(26878724, "Yang 2015/2016 — NUDT15 and thiopurine-induced leukopenia"),
                   C(CPIC, "CPIC thiopurine guideline (TPMT + NUDT15)")],
    )
    add(
        "rs35599367", "CYP3A4", "CYP3A4*22 (Reduced Expression)", "7", 99270539,
        ["pharmacogenomics"], "strong", "C", "T",
        "CYP3A4 clears more drugs than any other enzyme. The *22 intronic allele lowers expression ~40%: higher exposure to sensitive substrates (some statins, tacrolimus, quetiapine).",
        G("Typical CYP3A4 (CC)", "typical", "Standard clearance of CYP3A4 substrates.", "Baseline."),
        G("One *22 allele (CT)", "notable", "Reduced CYP3A4 expression.", "Simvastatin/atorvastatin exposure runs higher; a prescriber note, especially combined with SLCO1B1 variants.",
          recover=["Mention CYP3A4*22 if statin muscle symptoms ever appear — alternatives exist."]),
        G("Two *22 alleles (TT)", "caution", "Substantially reduced expression (rare).", "Meaningfully higher exposure to sensitive substrates — prescriber-level information."),
        tier="well-replicated", transferability="euro-biased",
        effect="~40% expression reduction per allele; substrate-dependent exposure increase",
        citations=[C(21691474, "Wang 2011 — CYP3A4*22 discovery"), C(PGKB, "CYP3A4*22 clinical annotations")],
    )
    add(
        "rs1045642", "ABCB1", "P-glycoprotein C3435T", "7", 87138645,
        ["pharmacogenomics"], "moderate", "A", "G",
        "ABCB1 (P-gp) pumps xenobiotics across gut, liver, kidney, and blood-brain barriers. C3435T subtly changes expression/folding; meta-analyses show drug-specific, modest exposure effects (digoxin, some DOACs).",
        G("Variant pair (AA)", "typical", "Lower-expression-associated pair (3435TT by legacy naming).", "Modest, drug-specific exposure shifts — informational."),
        G("Heterozygote (AG)", "typical", "Intermediate.", "Informational."),
        G("Reference pair (GG)", "typical", "Typical P-gp expression.", "Baseline."),
        tier="replicated", transferability="multi-ancestry",
        effect="Modest, drug-specific; meta-analyses heterogeneous",
        citations=[C(PGKB, "ABCB1 C3435T annotations — extensive but heterogeneous")],
    )
    add(
        "rs2032582", "ABCB1", "P-glycoprotein G2677T (A893S/T)", "7", 87160618,
        ["pharmacogenomics"], "moderate", "G", "T",
        "Coding ABCB1 variant (tri-allelic in reality; the rare A allele is not modeled here). Forms the functional haplotype with C3435T.",
        G("Reference (GG)", "typical", "Typical P-gp.", "Baseline."),
        G("Heterozygote (GT)", "typical", "Haplotype-intermediate.", "Informational."),
        G("Variant pair (TT)", "typical", "Altered-transport haplotype.", "Modest drug-specific effects; interpret with C3435T."),
        notes="Tri-allelic site: the rare 2677A allele is ignored by this biallelic model.",
        tier="replicated", effect="Modest, haplotype-dependent",
        citations=[C(PGKB, "ABCB1 haplotype annotations")],
    )
    add(
        "rs1128503", "ABCB1", "P-glycoprotein C1236T", "7", 87179601,
        ["pharmacogenomics"], "moderate", "A", "G",
        "Third member of the common ABCB1 haplotype (1236-2677-3435).",
        G("Variant pair (AA)", "typical", "1236TT-equivalent.", "Interpret as haplotype."),
        G("Heterozygote (AG)", "typical", "Intermediate.", "Informational."),
        G("Reference (GG)", "typical", "Typical.", "Baseline."),
        tier="replicated", effect="Haplotype member — modest",
        citations=[C(PGKB, "ABCB1 haplotype annotations")],
    )
    add(
        "rs2306283", "SLCO1B1", "OATP1B1 *1B (N130D)", "12", 21329738,
        ["pharmacogenomics"], "moderate", "A", "G",
        "SLCO1B1 N130D (*1B) modestly INCREASES hepatic uptake for some substrates — the counterpoint to the *5 decreased-function allele; *1B+*5 together define *15.",
        G("Reference (AA)", "typical", "*1A-type uptake.", "Baseline."),
        G("Heterozygote (AG)", "typical", "Mixed uptake haplotype.", "Informational; the *5 allele (rs4149056) carries the actionable signal."),
        G("*1B pair (GG)", "typical", "Higher-uptake haplotype.", "Slightly lower systemic exposure for some statins — informational."),
        tier="replicated", effect="Modest, substrate-specific; interpret with rs4149056",
        citations=[C(PGKB, "SLCO1B1 haplotype definitions"), C(CPIC, "CPIC statin guideline")],
    )
    add(
        "rs717620", "ABCC2", "MRP2 Promoter (-24C>T)", "10", 101542578,
        ["pharmacogenomics"], "moderate", "C", "T",
        "ABCC2 (MRP2) exports conjugated drugs into bile. The -24T allele modestly alters expression with drug-specific exposure effects (methotrexate, some antiepileptics) in cohort studies.",
        G("Reference (CC)", "typical", "Typical MRP2.", "Baseline."),
        G("One copy (CT)", "typical", "Small expression shift.", "Informational."),
        G("Two copies (TT)", "typical", "Reported exposure shifts for some substrates.", "Informational; evidence heterogeneous."),
        tier="tendency", effect="Drug-specific, modest, heterogeneous",
        citations=[C(PGKB, "ABCC2 -24C>T annotations")],
    )
    add(
        "rs2231137", "ABCG2", "BCRP V12M (Secondary Variant)", "4", 89061114,
        ["pharmacogenomics"], "moderate", "C", "T",
        "Secondary ABCG2 coding variant (V12M) with smaller functional impact than Q141K; more common in East-Asian ancestry.",
        G("Reference (CC)", "typical", "Typical BCRP at this site.", "Q141K (rs2231142) carries the main signal."),
        G("One copy (CT)", "typical", "Small transport shift.", "Informational."),
        G("Two copies (TT)", "typical", "Modest transport reduction.", "Interpret with Q141K."),
        tier="replicated", transferability="east-asian-derived",
        effect="Smaller than Q141K",
        citations=[C(PGKB, "ABCG2 V12M annotations")],
    )
    add(
        "rs3745274", "CYP2B6", "CYP2B6*6 (Q172H) — Efavirenz/Bupropion Axis", "19", 41512841,
        ["pharmacogenomics"], "strong", "G", "T",
        "CYP2B6*6 reduces metabolism of efavirenz (CPIC guideline exists), bupropion, and methadone. Poor-metabolizer status concentrates in *6/*6 homozygotes.",
        G("Typical CYP2B6 (GG)", "typical", "Standard clearance.", "Baseline."),
        G("Intermediate (GT)", "notable", "One reduced-function allele.", "Higher efavirenz/bupropion exposure on average — prescriber information.",
          recover=["If efavirenz or bupropion are ever prescribed, CYP2B6*6 carrier status is worth mentioning."]),
        G("Poor-metabolizer pattern (TT)", "caution", "*6/*6-like: substantially higher exposure.", "CPIC recommends consideration of reduced efavirenz dosing — strictly prescriber territory."),
        tier="well-replicated", transferability="multi-ancestry",
        effect="Efavirenz exposure roughly doubles in *6/*6 — actionable (CPIC B/A by drug)",
        citations=[C(CPIC, "CPIC efavirenz-CYP2B6 guideline"), C(PGKB, "CYP2B6*6 Level 1A")],
    )
    add(
        "rs28371686", "CYP2C9", "CYP2C9*5 (African-Ancestry Allele)", "10", 96741058,
        ["pharmacogenomics"], "strong", "C", "G",
        "CYP2C9*5 is a decreased-function allele found almost exclusively in African-ancestry populations — omitted by Euro-centric panels, which systematically under-calls warfarin sensitivity in African-ancestry patients.",
        G("No *5 (CC)", "typical", "Reference at this site.", "See CYP2C9 *2/*3 for the European-common alleles."),
        G("One *5 allele (CG)", "notable", "Reduced CYP2C9 function.", "Warfarin/NSAID exposure runs higher — prescriber note."),
        G("Two *5 alleles (GG)", "caution", "Rare; substantially reduced function.", "Warfarin sensitivity comparable to *3 carriers."),
        notes="Palindromic C/G pair — orientation reviewed.",
        tier="well-replicated", transferability="african-derived", reviewed=True,
        effect="Decreased function; part of CPIC warfarin algorithm for African-ancestry patients",
        citations=[C(CPIC, "CPIC warfarin guideline — African-ancestry alleles"), C(PGKB, "CYP2C9*5")],
    )
    add(
        "rs9332131", "CYP2C9", "CYP2C9*6 (Null, African-Ancestry Allele)", "10", 96731043,
        ["pharmacogenomics"], "strong", "A", "G",
        "CYP2C9*6 is a frameshift null allele in African-ancestry populations. Note: this is an indel-type variant — some arrays encode it poorly; treat a call here with extra caution.",
        G("No *6 detected (AA)", "typical", "Reference.", "Baseline."),
        G("One *6 allele (AG)", "caution", "One null copy.", "Meaningful warfarin/phenytoin sensitivity — prescriber note."),
        G("Two *6 alleles (GG)", "caution", "Very rare; no functional CYP2C9 from this allele pair.", "Major dose implications for CYP2C9 substrates — clinical confirmation essential."),
        tier="well-replicated", transferability="african-derived", chips=["23andme_v5"],
        effect="Null allele; large effect when present",
        citations=[C(CPIC, "CPIC warfarin guideline"), C(PGKB, "CYP2C9*6")],
    )
    add(
        "rs7900194", "CYP2C9", "CYP2C9*8 (African-Ancestry Allele)", "10", 96702066,
        ["pharmacogenomics"], "strong", "G", "A",
        "CYP2C9*8 reduces function and is the most common decreased-function CYP2C9 allele in African-ancestry populations — a flagship example of why Euro-only panels mis-serve everyone else.",
        G("No *8 (GG)", "typical", "Reference.", "Baseline."),
        G("One *8 allele (AG)", "notable", "Reduced function.", "Lower average warfarin requirement — prescriber note."),
        G("Two *8 alleles (AA)", "caution", "Substantially reduced function.", "Comparable to *3 heterozygosity in dose terms."),
        tier="well-replicated", transferability="african-derived",
        effect="Decreased function; in CPIC warfarin algorithm",
        citations=[C(CPIC, "CPIC warfarin guideline — *8"), C(PGKB, "CYP2C9*8")],
    )
    add(
        "rs12777823", "CYP2C cluster", "Warfarin Dose Modifier (African Ancestry)", "10", 96405502,
        ["pharmacogenomics"], "strong", "G", "A",
        "Regulatory-region variant near CYP2C18 associated with lower warfarin dose requirement specifically in African-American cohorts; CPIC recommends factoring it in for African-ancestry dosing.",
        G("Reference (GG)", "typical", "No dose-modifier allele.", "Baseline."),
        G("One A allele (AG)", "notable", "Lower average warfarin requirement (African-ancestry data).", "Prescriber note."),
        G("Two A alleles (AA)", "notable", "Larger dose-requirement reduction.", "Prescriber note; ancestry-specific evidence base."),
        tier="well-replicated", transferability="african-derived",
        effect="~7-9 mg/week dose reduction per allele in African-American cohorts",
        citations=[C(23755828, "Perera 2013, Lancet — warfarin dosing in African Americans"),
                   C(CPIC, "CPIC warfarin guideline addendum")],
    )
    add(
        "rs28371725", "CYP2D6", "CYP2D6*41 (Splice-Reducing Allele)", "22", 42523805,
        ["pharmacogenomics"], "strong", "C", "T",
        "CYP2D6*41 reduces (does not abolish) splicing efficiency — a decreased-function allele refining intermediate-metabolizer calls beyond *4/*10.",
        G("No *41 (CC)", "typical", "Reference at this site.", "See the CYP2D6 gene card."),
        G("One *41 allele (CT)", "notable", "Decreased-function contribution.", "Counts toward intermediate-metabolizer status."),
        G("Two *41 alleles (TT)", "notable", "Reduced-function pair.", "Intermediate metabolism likely — full status needs the whole allele set plus CNV (which arrays cannot see)."),
        tier="well-replicated", transferability="multi-ancestry",
        effect="Decreased function (activity value 0.5 per CPIC)",
        citations=[C(CPIC, "CPIC CYP2D6 allele definitions"), C(PGKB, "CYP2D6*41")],
    )
    add(
        "rs35742686", "CYP2D6", "CYP2D6*3 (Frameshift Null)", "22", 42524244,
        ["pharmacogenomics"], "strong", "A", "T",
        "CYP2D6*3 is a frameshift null allele (~1-2% of European alleles). Indel-encoded on arrays — calls can be unreliable; treat with care.",
        G("Deletion-carrier pattern (AA)", "caution", "Null-allele signal on both reads.", "Poor metabolism contribution — clinical confirmation essential before any decision."),
        G("One null signal (AT)", "notable", "One *3 read.", "Counts toward reduced metabolism."),
        G("No *3 (TT)", "typical", "Reference.", "Baseline."),
        notes="Palindromic A/T site and indel-adjacent — array reliability limited; reviewed.",
        tier="well-replicated", reviewed=True, chips=["23andme_v5"],
        effect="Null allele when truly present",
        citations=[C(CPIC, "CPIC CYP2D6 allele definitions"), C(PGKB, "CYP2D6*3")],
    )
    add(
        "rs5030655", "CYP2D6", "CYP2D6*6 (Deletion Null)", "22", 42525086,
        ["pharmacogenomics"], "moderate", "T", "A",
        "CYP2D6*6 single-base deletion null allele (~1% European). Array encoding of indels varies by vendor.",
        G("No *6 (TT)", "typical", "Reference.", "Baseline."),
        G("One null signal (AT)", "notable", "One *6 read.", "Counts toward reduced metabolism; confirm clinically."),
        G("Null pair signal (AA)", "caution", "Rare.", "Poor-metabolizer contribution; confirm clinically."),
        notes="Palindromic A/T and indel-adjacent — reviewed; low array reliability.",
        tier="replicated", reviewed=True, chips=["23andme_v5"],
        effect="Null allele when truly present",
        citations=[C(PGKB, "CYP2D6*6")],
    )
    add(
        "rs1801280", "NAT2", "NAT2*5 (Slow Acetylator Allele, I114T)", "8", 18257854,
        ["pharmacogenomics", "detox"], "strong", "T", "C",
        "NAT2 acetylates isoniazid, sulfonamides, hydralazine, and dietary heterocyclic amines. *5 (I114T) is the major European slow allele: slow acetylators clear these drugs slower and handle charred-meat amines differently.",
        G("Fast at this site (TT)", "typical", "No *5 allele.", "Overall status needs the three-allele NAT2 panel."),
        G("Intermediate (CT)", "typical", "One slow allele.", "Counts toward slow-acetylator status."),
        G("Slow pair (CC)", "notable", "*5/*5-type slow acetylation.", "Higher exposure to NAT2-cleared drugs (isoniazid flag for any TB treatment) and slower clearance of well-done-meat amines.",
          eat=["Prefer gentler cooking for meats (braise/sous-vide over charring) — slow acetylators process heterocyclic amines less efficiently."]),
        tier="well-replicated", transferability="multi-ancestry",
        effect="Slow acetylation: isoniazid hepatotoxicity OR ~2 in slow acetylators",
        citations=[C(PGKB, "NAT2 acetylator phenotype annotations"), C(GWAS, "NAT2 meta-analyses")],
    )
    add(
        "rs1799930", "NAT2", "NAT2*6 (Slow Acetylator Allele, R197Q)", "8", 18258103,
        ["pharmacogenomics", "detox"], "strong", "G", "A",
        "Second major slow-acetylator allele (*6), common across ancestries.",
        G("Fast at this site (GG)", "typical", "No *6.", "Interpret with the NAT2 set."),
        G("Intermediate (AG)", "typical", "One slow allele.", "Counts toward slow status."),
        G("Slow pair (AA)", "notable", "*6/*6-type slow acetylation.", "Same implications as *5 homozygosity."),
        tier="well-replicated", transferability="multi-ancestry",
        effect="Slow allele — see NAT2*5",
        citations=[C(PGKB, "NAT2*6 annotations")],
    )
    add(
        "rs1799931", "NAT2", "NAT2*7 (Slow Allele, G286E — East-Asian Common)", "8", 18258370,
        ["pharmacogenomics", "detox"], "strong", "G", "A",
        "Third major slow allele (*7), relatively more common in East-Asian ancestry.",
        G("Fast at this site (GG)", "typical", "No *7.", "Interpret with the NAT2 set."),
        G("Intermediate (AG)", "typical", "One slow allele.", "Counts toward slow status."),
        G("Slow pair (AA)", "notable", "*7/*7-type slow acetylation.", "Same implications as other slow pairs."),
        tier="well-replicated", transferability="east-asian-derived",
        effect="Slow allele — see NAT2*5",
        citations=[C(PGKB, "NAT2*7 annotations")],
    )
    add(
        "rs1048943", "CYP1A1", "CYP1A1*2C (I462V) — PAH Activation", "15", 75012985,
        ["pharmacogenomics", "detox"], "moderate", "T", "C",
        "CYP1A1 bio-activates polycyclic aromatic hydrocarbons (smoke, char) into reactive intermediates. The 462V allele shows higher inducible activity in vitro; cancer associations are smoking-dependent and modest.",
        G("Reference (TT)", "typical", "Typical PAH activation.", "Smoke exposure is the variable that matters."),
        G("One 462V allele (CT)", "typical", "Higher inducible activity in vitro.", "Only meaningful with heavy smoke/char exposure."),
        G("Two 462V alleles (CC)", "notable", "Highest inducible activation.", "Another reason (as if needed) to avoid smoke exposure and heavy charring.",
          eat=["Marinate and cook gentler; avoid charred surfaces if you grill often."]),
        tier="replicated", transferability="multi-ancestry",
        effect="In-vitro activity difference; exposure-dependent clinical relevance",
        citations=[C(PGKB, "CYP1A1*2C annotations")],
    )
    add(
        "rs4646903", "CYP1A1", "CYP1A1 MspI (3'UTR) Variant", "15", 75011882,
        ["pharmacogenomics", "detox"], "moderate", "T", "C",
        "Classic CYP1A1 MspI restriction-site variant in LD with *2C; historical smoking-related cancer literature, effects modest and exposure-dependent.",
        G("Reference (TT)", "typical", "Typical.", "Baseline."),
        G("One copy (CT)", "typical", "LD-partner of *2C.", "Exposure-dependent."),
        G("Two copies (CC)", "typical", "Higher-inducibility haplotype.", "Avoid smoke exposure — true for every genotype."),
        tier="replicated", effect="Modest, exposure-dependent",
        citations=[C(PGKB, "CYP1A1 MspI annotations")],
    )
    add(
        "rs1695", "GSTP1", "Glutathione Transferase Pi (I105V)", "11", 67352689,
        ["pharmacogenomics", "detox", "antioxidant"], "strong", "A", "G",
        "GSTP1 conjugates electrophiles (including chemo drugs and air-pollution metabolites) to glutathione. I105V changes substrate kinetics — faster for some substrates, slower for others.",
        G("Ile/Ile (AA)", "typical", "Reference kinetics.", "Baseline conjugation."),
        G("Ile/Val (AG)", "typical", "Mixed kinetics.", "Substrate-dependent shifts."),
        G("Val/Val (GG)", "notable", "Altered substrate kinetics.", "Associated with different oxaliplatin/anthracycline handling in oncology cohorts (specialist territory) and with pollution-response differences.",
          recover=["Cruciferous vegetables and sulfur-rich foods support glutathione conjugation for every genotype."]),
        tier="replicated", transferability="multi-ancestry",
        effect="Substrate-dependent kinetic shift",
        citations=[C(PGKB, "GSTP1 I105V annotations")],
    )
    add(
        "rs55886062", "DPYD", "DPYD*13 (I560S) — Fluoropyrimidine Toxicity", "1", 97981343,
        ["pharmacogenomics"], "strong", "A", "C",
        "DPYD*13 is a rare decreased-function allele; CPIC treats carriers like *2A carriers for 5-FU/capecitabine dose reduction.",
        G("Reference (AA)", "typical", "No *13.", "See the DPYD gene card."),
        G("One *13 allele (AC)", "caution", "Decreased DPD activity.", "CPIC: substantial dose reduction for fluoropyrimidines — oncologist territory, flagged in advance."),
        G("Two *13 alleles (CC)", "caution", "Extremely rare; DPD deficiency.", "Fluoropyrimidines potentially life-threatening — clinical confirmation mandatory."),
        tier="well-replicated", effect="Large per-allele toxicity effect (CPIC A)",
        citations=[C(CPIC, "CPIC fluoropyrimidine guideline — *13"), C(CLIN, "DPYD I560S")],
    )
    add(
        "rs67376798", "DPYD", "DPYD D949V — Fluoropyrimidine Toxicity", "1", 97547947,
        ["pharmacogenomics"], "strong", "T", "A",
        "D949V halves DPD activity per allele; in the CPIC activity-score model alongside *2A and *13.",
        G("Reference (TT)", "typical", "No D949V.", "See DPYD gene card."),
        G("One D949V allele (AT)", "caution", "Half-activity allele.", "CPIC: reduced fluoropyrimidine starting dose — flag to oncology in advance."),
        G("Two D949V alleles (AA)", "caution", "Very rare; severe deficiency.", "Clinical confirmation mandatory."),
        notes="Palindromic A/T site — reviewed.",
        tier="well-replicated", reviewed=True,
        effect="Half-activity allele (CPIC activity score 0.5)",
        citations=[C(CPIC, "CPIC fluoropyrimidine guideline — D949V")],
    )
    add(
        "rs75017182", "DPYD", "DPYD HapB3 (Intronic Splice Variant)", "1", 97740414,
        ["pharmacogenomics"], "strong", "G", "C",
        "HapB3 (c.1129-5923C>G) causes partial exon skipping — the most common European decreased-function DPYD haplotype (~4.7% carriers).",
        G("Reference (GG)", "typical", "No HapB3.", "See DPYD gene card."),
        G("One HapB3 allele (CG)", "caution", "Partial splice defect (activity 0.5).", "CPIC: consider reduced fluoropyrimidine starting dose."),
        G("Two HapB3 alleles (CC)", "caution", "Rare; substantially reduced DPD.", "Clinical confirmation mandatory."),
        notes="Palindromic C/G site — reviewed.",
        tier="well-replicated", reviewed=True, transferability="euro-biased",
        effect="Activity 0.5 per allele; most common European DPYD risk haplotype",
        citations=[C(CPIC, "CPIC fluoropyrimidine guideline — HapB3")],
    )
    add(
        "rs4148323", "UGT1A1", "UGT1A1*6 (G71R — East-Asian Gilbert/Irinotecan)", "2", 234669144,
        ["pharmacogenomics"], "strong", "G", "A",
        "UGT1A1*6 causes Gilbert-type bilirubin elevation and irinotecan sensitivity in East-Asian populations where *28 typing alone under-calls risk.",
        G("No *6 (GG)", "typical", "Reference at this site.", "See UGT1A1 rs887829 for the *28 tag."),
        G("One *6 allele (AG)", "notable", "Reduced glucuronidation.", "Mild bilirubin elevations possible; irinotecan note for oncology."),
        G("Two *6 alleles (AA)", "caution", "Gilbert-pattern conjugation reduction.", "Benign jaundice under fasting/stress; irinotecan dose caution (CPIC)."),
        tier="well-replicated", transferability="east-asian-derived",
        effect="Comparable to *28 homozygosity for irinotecan risk in East-Asian cohorts",
        citations=[C(CPIC, "CPIC UGT1A1 guidance"), C(PGKB, "UGT1A1*6")],
    )
    add(
        "rs10264272", "CYP3A5", "CYP3A5*6 (African-Ancestry Splice Allele)", "7", 99262835,
        ["pharmacogenomics"], "strong", "C", "T",
        "CYP3A5*6 is a splice-defect non-expressor allele common in African-ancestry populations; needed alongside *3 for accurate tacrolimus phenotype in African-ancestry patients.",
        G("No *6 (CC)", "typical", "Reference at this site.", "Interpret with *3 (rs776746)."),
        G("One *6 allele (CT)", "notable", "One non-expressor copy.", "Affects tacrolimus dose prediction — transplant-team information."),
        G("Two *6 alleles (TT)", "notable", "Non-expressor pair at this allele.", "Same clinical meaning as *3/*3."),
        tier="well-replicated", transferability="african-derived",
        effect="Non-expressor allele (CPIC tacrolimus algorithm)",
        citations=[C(CPIC, "CPIC tacrolimus guideline — *6/*7"), C(PGKB, "CYP3A5*6")],
    )
    add(
        "rs41303343", "CYP3A5", "CYP3A5*7 (African-Ancestry Frameshift)", "7", 99247567,
        ["pharmacogenomics"], "moderate", "T", "A",
        "CYP3A5*7 frameshift non-expressor allele, essentially African-ancestry-specific; indel-encoded — array calls may be unreliable.",
        G("No *7 signal (TT)", "typical", "Reference.", "Interpret with *3/*6."),
        G("One *7 signal (AT)", "notable", "One non-expressor copy.", "Transplant-team information; confirm clinically."),
        G("*7 pair signal (AA)", "notable", "Non-expressor pair.", "Confirm clinically."),
        notes="Palindromic A/T, indel-adjacent — reviewed; limited array reliability.",
        tier="replicated", reviewed=True, transferability="african-derived", chips=["23andme_v5"],
        effect="Non-expressor allele when truly present",
        citations=[C(CPIC, "CPIC tacrolimus guideline"), C(PGKB, "CYP3A5*7")],
    )