SystemOnline
WorkflowsClopidogrel + Warfarin + Simvastatin
Reference PopulationsAFR EUR EAS SAS AMR
1000G SamplesPhase-3 (n=2504)
Decision EngineDeterministic PGx
Evidence LayerV1–V10 / R1–R12 / U1–U9
Export StackCSV PDF FHIR
Data ModeResearch Simulation
GovernanceAudit Enabled
Engineanukriti-pgx-core==0.2.1
SystemOnline
WorkflowsClopidogrel + Warfarin + Simvastatin
Reference PopulationsAFR EUR EAS SAS AMR
1000G SamplesPhase-3 (n=2504)
Decision EngineDeterministic PGx
Evidence LayerV1–V10 / R1–R12 / U1–U9
Export StackCSV PDF FHIR
Data ModeResearch Simulation
GovernanceAudit Enabled
Engineanukriti-pgx-core==0.2.1
SystemOnline
WorkflowsClopidogrel + Warfarin + Simvastatin
Reference PopulationsAFR EUR EAS SAS AMR
1000G SamplesPhase-3 (n=2504)
Decision EngineDeterministic PGx
Evidence LayerV1–V10 / R1–R12 / U1–U9
Export StackCSV PDF FHIR
Data ModeResearch Simulation
GovernanceAudit Enabled
Engineanukriti-pgx-core==0.2.1
SystemOnline
WorkflowsClopidogrel + Warfarin + Simvastatin
Reference PopulationsAFR EUR EAS SAS AMR
1000G SamplesPhase-3 (n=2504)
Decision EngineDeterministic PGx
Evidence LayerV1–V10 / R1–R12 / U1–U9
Export StackCSV PDF FHIR
Data ModeResearch Simulation
GovernanceAudit Enabled
Engineanukriti-pgx-core==0.2.1
Anukriti logo

Anukriti

Case Studies & Evidence

The Genome Doesn't
Speak One Language

Pharmacogenomic variant frequencies shift dramatically across populations. The same drug can be life-saving in one ancestry group and life-threatening in another. Anukriti is built around this reality. Below are documented case studies, regional summaries, and the structural evidence that supports population-aware simulation.

Research and education only. Not a clinical diagnostic system. All evidence linked back to peer-reviewed sources.

Reference Framework

  • CPIC guidelines
  • PharmVar star-alleles
  • 1000 Genomes superpops
  • gnomAD frequencies
  • PharmGKB annotations
Helix · Reference

~90%

Drug candidates fail in clinical trials

Toxicity drives ~30% of failures

99.5%

Of individuals carry ≥1 PGx-relevant variant

UK Yellow Card pharmacovigilance analysis

9%

Of ADR reports involve known PGx risk drugs

75% tied to just CYP2C19, CYP2D6, SLCO1B1

83.8%

GWAS participants of European ancestry (2021)

Representation gap that drives bias

30%

ADR reduction with pre-emptive PGx testing

PREPARE trial, U-PGx consortium

5 superpops

AFR · EUR · EAS · SAS · AMR

1000 Genomes reference framework

↻ helix
biotechCase 01 · HLA-B*15:02

Carbamazepine

Carbamazepine is an antiepileptic drug. HLA-B*15:02 carriers are at sharply elevated risk of severe cutaneous adverse reactions (SJS/TEN). Allele frequency varies dramatically by ancestry, which is why Western trials initially failed to detect the risk.

Source: FDA recommendation context (PMC)
Region / PopulationEvidence & ObservationsImplicationRisk
Han Chinese (Taiwan)In a case-control study, ~all (44/44) Han Chinese SJS/TEN patients carried HLA-B*15:02. Carriage was ~3% in tolerant patients and ~8.6% in healthy controls. Odds ratio for SJS/TEN >2,500.Taiwan mandated genetic testing before carbamazepine. Anukriti can stratify Asian vs non-Asian cohorts and surface population-specific risk early.high
Thailand · Singapore · Malaysia · PhilippinesHLA-B*15:02 frequency: ~11.6% Singapore, 8.4% Malays, 6.1% Thais, 5.3% Filipinos. Carbamazepine is among the most common SJS/TEN causes regionally.Population-wide screening introduced. Anukriti can simulate high-risk vs low-risk genotype distributions when modelling alternative therapies.high
IndiaReports of carbamazepine-induced SJS/TEN, especially in Tamil and Telugu ancestry. HLA-B*15:02 frequency varies across Indian groups; pre-therapy testing is now recommended.Anukriti supports multi-ethnic Indian cohort exploration to evaluate the impact of genotype screening.medium
Europe & North AmericaHLA-B*15:02 is virtually absent (~0%) in European and African-descent populations. Routine screening is not required.Low frequency explains why early Western clinical trials missed the risk — exactly the bias Anukriti is designed to expose.low

Why this case matters

A drug deemed safe in Western trials caused catastrophic reactions in specific Asian populations. Population-aware pre-clinical assessment changes this picture.

biotechCase 02 · HLA-B*57:01

Abacavir

Abacavir, an antiretroviral, can trigger severe multisystem hypersensitivity in HLA-B*57:01 carriers. Routine screening collapsed adverse-reaction incidence and is now standard.

Source: HLA-B*57:01 fact sheet (ARUP Consult)
Region / PopulationEvidence & ObservationsImplicationRisk
Europeans / White populationsHypersensitivity occurred in 5–8% of White patients pre-screening. HLA-B*57:01 frequency ~6.8% in Europeans. Routine genotyping reduced incidence to <0.5%.Demonstrates how a single allele test transforms safety. Anukriti can simulate the "screen vs no screen" delta on synthetic cohorts.high
Southwest Asians & IndiansAllele frequency reaches ~11% in Southwest Asian groups; ~3–5% in some Indian studies. Pre-prescription screening recommended.Anukriti can model Indian and Southwest Asian cohorts to plan region-specific protocols.high
African Americans / AfricansAllele frequency ~1% in African populations; observed hypersensitivity ~2–3%. Routine screening still advised.Illustrates how allele-frequency differences shift risk. Useful for global protocol planning.medium
Latin AmericaFrequencies range 2–3% (Mexicans 2.2%, South Americans 2.6%). Many countries implement screening.Even moderate-frequency populations benefit from genotyping; Anukriti can quantify cohort-level outcomes.medium

Why this case matters

Abacavir is the cleanest example of how one HLA test transforms safety outcomes. Population-aware modelling makes this benefit visible in design phase.

biotechCase 03 · CYP2C9 + VKORC1

Warfarin

Warfarin dosing depends on combined CYP2C9 and VKORC1 genotypes; ancestry is a strong modifier of dose requirements. Empiric dosing can lead to bleeding or thrombosis.

Source: Hawai‘i warfarin dosing study (PMC)
Region / PopulationEvidence & ObservationsImplicationRisk
Hawai‘i (multi-ethnic cohort, n=113)Caucasians required highest mean dose (~4.4 mg/day); Asians lowest (~2.5 mg/day); Native Hawaiians and Portuguese intermediate (~3.8 mg/day). Differences traced to CYP2C9 and VKORC1 variants.Anukriti can model dose distribution by ancestry and quantify risk of mis-dosing.high
East Asians (China · Japan · Korea)High frequency of VKORC1 −1639 A allele drives lower maintenance doses (~2–3 mg/day) than in Caucasians.Stratified dosing protocols; Anukriti can simulate regional cohorts to plan trial dosing.high
Europeans / AmericansHigher CYP2C9*2/*3 prevalence; typical doses 4–6 mg/day. Without genotyping, dose adjusted empirically.Anukriti synthesizes how genotype-aware vs empiric dosing changes adverse outcome rates.medium

Why this case matters

Warfarin shows that empiric, one-size dosing breaks down across ancestries. Genotype-aware simulation is the cleanest pathway to safer dosing strategies.

biotechCase 04 · CYP2C19

Clopidogrel

Clopidogrel is an antiplatelet pro-drug requiring CYP2C19 activation. Loss-of-function alleles produce poor or intermediate metabolizers, leading to higher cardiovascular event risk.

Source: NCBI Medical Genetics Summary
Region / PopulationEvidence & ObservationsImplicationRisk
Europeans & African Americans~2% of Caucasians and ~4% of African Americans are CYP2C19 poor metabolizers. ~20% are intermediate metabolizers. FDA labels carry boxed warnings for poor metabolizers.Even Western populations show meaningful variability. Anukriti can model the benefit of switching to prasugrel or ticagrelor.medium
East Asians & OceaniansLoss-of-function alleles much more prevalent: ~14% PM in Chinese, up to ~57% in Pacific Islanders. Intermediate metabolizers >45% in East Asians.Region-wide alternative therapy may be justified. Anukriti can simulate East Asian and Oceanian cohorts to quantify failure risk.high
South & Central AsiansIntermediate metabolizer frequency >40%. Many patients potentially benefit from genotype-guided therapy.South Asian cohort modelling helps plan pre-therapy genotyping policies.high
Pacific Islanders (Oceania)Highest poor-metabolizer frequency (~57%). Clopidogrel may be ineffective in over half the population.Strongest case for population-wide alternative agents; Anukriti makes the magnitude visible.high

Why this case matters

Pro-drug activation can vary dramatically by genotype. Synthetic cohort modelling reveals where clopidogrel will likely underperform and where alternatives are warranted.

biotechCase 05 · CYP2D6

Codeine

Codeine is metabolised to morphine by CYP2D6. Both extremes of metabolism are unsafe: ultra-rapid metabolizers risk overdose; poor metabolizers receive negligible analgesia.

Source: NCBI Medical Genetics Summary
Region / PopulationEvidence & ObservationsImplicationRisk
North Africa · Ethiopia · Arab nationsUp to ~28% are CYP2D6 ultra-rapid metabolizers. Rapid morphine production raises risk of life-threatening respiratory depression at standard doses.FDA contraindications followed pediatric tonsillectomy reports. Anukriti can quantify risk in synthetic high-UM cohorts.high
Caucasians & African populationsDuplicated CYP2D6 alleles in ~10% of Caucasians, ~3% in African Americans. PM alleles (*4, *5, *6) more common in Europeans; *17 more common in Africans.Moderate but meaningful risk; especially relevant in pediatric and breastfeeding contexts.medium
East AsiansOnly ~1% UMs in Chinese/Japanese populations, but >50% of CYP2D6 alleles show reduced function or duplications.Highly variable analgesic efficacy; Anukriti models analgesic response distribution by region.medium
Latin America & EuropeUM frequencies 1–10%; PM phenotypes 5–10%. Caution advised in pediatric and postpartum populations.Highlights global heterogeneity; cohort-level modelling helps protocol planning.medium

Why this case matters

Codeine safety varies strongly by CYP2D6 genotype. Anukriti can model alternative analgesic strategies for high-risk populations.

biotechCase 06 · G6PD

Primaquine

Primaquine treats Plasmodium vivax and ovale malaria but causes oxidative hemolysis in G6PD-deficient patients. Screening is essential, and prevalence varies sharply by region.

Source: NCBI Medical Genetics Summary
Region / PopulationEvidence & ObservationsImplicationRisk
GlobalG6PD deficiency affects ~400 million people globally. Prevalence ~5% globally; >25% in malaria-endemic regions. ~12% of African Americans in the US are G6PD deficient.Anukriti models hemolytic-event frequency under different dosing strategies and the value of screening.high
AfricaPrevalence often >25%. G6PD A− is common and produces mild-to-moderate hemolysis. WHO recommends weekly dosing for deficient individuals.Anukriti can simulate African cohorts and evaluate alternative regimens like tafenoquine or weekly primaquine.high
Mediterranean & Middle EastG6PD Mediterranean variant causes severe hemolysis. Frequencies: 0.5% Portuguese males, 6.4% Cypriots, 8.3% newborn Greek males.Demonstrates that "European" is not homogeneous; Anukriti supports safer dosing protocols.high
Southeast Asia (Thailand · Myanmar · Cambodia · Laos · Vietnam)Wide variation: 9–31% Thailand, ~30% Kachin (Myanmar/China border), 8% Laos, 9% Vietnam, 15.8% Myanmar. Variants include G6PD Canton and Viangchan.Anukriti enables Southeast Asian cohort modelling to evaluate weekly dosing and alternative regimens.high
AmericasLower prevalence (2–3%) but present in specific ethnic groups. Screening still recommended.Mixed-ancestry simulation is critical in Latin American populations.medium

Why this case matters

A widely used antimalarial is hazardous in a substantial fraction of global populations. Cohort simulation directly informs safe dosing strategy.

biotechCase 07 · NAT2

Isoniazid

Isoniazid is a first-line tuberculosis drug metabolised by NAT2. Slow acetylators accumulate hepatotoxic metabolites; rapid acetylators may underdose.

Source: NAT2 hepatotoxicity multi-ethnic cohort (PMC)
Region / PopulationEvidence & ObservationsImplicationRisk
Multi-ethnic cohort (Milan, n=102)~21.6% developed antituberculosis drug-induced hepatotoxicity (ATDH). Slow acetylators: 16.7% ATDH vs 4.9% in rapid/intermediate. Onset earlier in slow acetylators (median 0.5 vs 2 months). Adjusted hazard ratio ~3.05.Anukriti can model liver injury risk and explore dose adjustments by NAT2 status.high
Global distribution (53-study systematic review)Slow acetylator (SA) prevalence: ~60.16% Africa, ~58.93% Europe, ~41.94% Americas, ~34.42% Asia. DILI in SA carriers: 41.86% Africa, 42.83% Asia, 27.36% Europe, 29.06% Americas. Rapid acetylators (RA) more common in Asia (28.54%) than Europe (6.24%).Anukriti models global NAT2 phenotype variation to guide genotype-aware dosing strategy.high
Japan & East AsiaRapid acetylators more prevalent (some studies >40%). Standard dosing may underexpose patients, leading to treatment failure.Anukriti can explore RA-dominant cohorts to test higher-dose protocols.medium

Why this case matters

NAT2 polymorphism is a textbook example where one-size-fits-all dosing is suboptimal. Modelling phenotype distributions enables safer, more effective regimens.

biotechCase 08 · DPYD

5-Fluorouracil (5-FU)

5-FU and fluoropyrimidines are widely used in chemotherapy. DPYD variants reduce DPD activity, causing severe toxicity. Variant spectrums differ across populations.

Source: DPYD review (PMC)
Region / PopulationEvidence & ObservationsImplicationRisk
Europe (Caucasians)Key variants: c.1905+1G>A, c.2846A>T, c.1236G>A, c.1679T>G. DPYD*2A frequency: ~0.024 Finnish, 0.006 non-Finnish Europeans. Heterozygotes can develop life-threatening toxicity at standard doses.Pre-treatment DPYD genotyping adopted in several European countries; Anukriti models prevention impact.high
South Asian & Middle EasternDPYD*2A: ~0.004 South Asian, 0.003 Middle Eastern. c.85T>C frequency: 0.072 East Asians, 0.255 South Asians. c.577A>G in 2.2% of Africans. Novel Indian/Japanese variants cause severe toxicity.European panels miss variants relevant to non-European populations. Anukriti can evaluate universal vs region-specific genotyping.high
Africa & East AsiaSeveral Caucasian DPYD variants (c.1905+1G>A, c.1679T>G, c.1601G>A) absent in East Africans and Japanese. East Asians carry novel variants of unknown significance.Highlights diversity of DPYD variation and need for region-specific risk allele research.medium
Meta-analysis (rs1801160, Val732Ile)Carriers of rs1801160 A allele had 1.73× higher overall toxicity and 2.37× higher hematological toxicity vs GG. Allele frequency: 0.046 Europeans, 0.098 South Asians.Even moderate-frequency variants meaningfully shift toxicity. Anukriti can quantify cohort-level adverse-event reduction.medium

Why this case matters

DPYD variant spectrum and frequency differ across populations. European-only screening panels can miss meaningful risk in other regions.

↻ helix

Cross-Region Summary

Pharmacogenomics is regional

Allele frequencies and risk patterns differ markedly by region. The table below condenses the case study evidence into a regional view.

East & Southeast Asia

Key concerns

High HLA-B*15:02 (carbamazepine SJS/TEN); high CYP2C19 LOF (clopidogrel failure); variable G6PD deficiency including Canton/Viangchan; moderate NAT2 SA (~34%); novel DPYD variants absent from Caucasian panels.

Documented risks

Carbamazepine-induced SJS/TEN in Han Chinese; clopidogrel-related stent thrombosis in East Asians; primaquine-induced hemolysis in Thai/Cambodian patients; NAT2-driven INH outcomes; unexpected 5-FU toxicity from novel DPYD variants.

South Asia (India · Pakistan · Bangladesh)

Key concerns

Mixed HLA-B*57:01 and HLA-B*15:02; high TB burden with variable NAT2 acetylator status; significant G6PD deficiency; DPYD variants like c.85T>C, c.2194G>A, c.1627G>A causing 5-FU toxicity.

Documented risks

SJS/TEN from carbamazepine; abacavir hypersensitivity; INH-induced hepatotoxicity; primaquine-related hemolysis; severe 5-FU toxicity from novel DPYD variants.

Africa (Sub-Saharan · Ethiopia · Kenya · Nigeria)

Key concerns

Very high G6PD deficiency (>25%); high NAT2 SA (~60.16%); high CYP2D6 ultra-rapid metabolizers (up to 28%); presence of G6PD A− and novel NAT2 alleles.

Documented risks

Severe primaquine/dapsone hemolysis; early INH hepatotoxicity; codeine overdose risk in UMs; variable clopidogrel efficacy; warfarin dose modulation.

Middle East & Mediterranean

Key concerns

G6PD Mediterranean (severe hemolysis); moderate HLA-B*57:01; some NAT2 SA; DPYD*2A in Turkish/Portuguese cohorts (~0.6–0.9%).

Documented risks

Primaquine and rasburicase hemolysis in G6PD Mediterranean carriers; abacavir hypersensitivity; moderate INH hepatotoxicity; 5-FU toxicity in DPYD*2A carriers.

Europe (Western & Northern)

Key concerns

Low HLA-B*15:02; moderate HLA-B*57:01 (~6.8%); higher warfarin doses; moderate NAT2 SA (58.93%); low CYP2C19 LOF; common DPYD variants (frequency 0.006–0.024).

Documented risks

Abacavir hypersensitivity (mitigated by screening); warfarin dose differences across subgroups; codeine toxicity in UMs; 5-FU toxicity reduced by DPYD genotyping; INH dosing adjustments for SAs.

Americas (North & South)

Key concerns

Mixed ancestry alters allele distribution. NAT2 SA ~41.94%; G6PD deficiency 2–3%; CYP2D6 duplications in some groups; moderate DPYD frequencies.

Documented risks

Abacavir hypersensitivity; warfarin dose variation; codeine caution in UMs; need for DPYD genotyping in chemotherapy; variable clopidogrel response in Hispanic populations.

↻ helix

Why the Static-Rules Era Falls Short

Limits of hard-coded PGx

Hard-coded CPIC tables make outputs auditable, but they have known structural limitations. Anukriti is designed to mitigate them transparently.

Limit · 01

Limited gene & variant coverage

CPIC panels test a subset of common star alleles. Newly discovered variants may default to "normal". GeT-RM analysis flagged ~339 star alleles as outdated/inconsistent.

Mitigation

Programmatic PharmVar/PharmGKB sync with version tracking; integrate Aldy/PyPGx for novel variants.

Limit · 02

Missing polygenic & gene–environment effects

Single-gene rules miss combinations and environmental modifiers. PRS, epigenetics, age, sex, and comorbidities all influence response.

Mitigation

Couple CPIC rules with polygenic models and non-genetic factors as evidence emerges.

Limit · 03

Under-representation of diverse ancestries

CPIC guidance largely derives from European cohorts; rare variants and gene–gene interactions in other ancestries may be missed.

Mitigation

Population-specific allele frequency calibration and synthetic cohort generation across AFR/EUR/EAS/SAS/AMR.

Limit · 04

Reliance on single guideline source

CPIC vs DPWG vs FDA can disagree. Comparative testing showed ~12.8% of phenotype translations and ~42.9% of medication recommendations differ from CPIC.

Mitigation

Modular knowledge layer that ingests multiple guideline sources and surfaces discrepancies via RAG.

Limit · 05

Limited large-scale clinical utility data

Few well-powered trials demonstrate outcome improvement, slowing payer/regulator adoption.

Mitigation

Use simulation tools to prioritise highest-impact gene–drug pairs and keep prototypes positioned as research/education.

↻ helix

How Anukriti uses this evidence

Built for population-aware simulation

  • check_circleSynthetic cohort construction across AFR · EUR · EAS · SAS · AMR using 1000 Genomes-aligned frequencies.
  • check_circleDeterministic CPIC + PharmVar rules with explicit call statuses (called / cannot_call / insufficient_data).
  • check_circleSide-by-side cohort comparison (e.g. EAS vs EUR) with explicit Δ% deltas.
  • check_circleEvidence map and reproducibility pack export for every run.
  • check_circleConfidence calibration so reviewers see uncertainty, not just numbers.
  • check_circleReviewer reports + partner handoff bundles formatted for trial-design teams.

Sources & Further Reading

Primary references

Anukriti is a research simulation platform. It is not a diagnostic device, prescribing system, or substitute for medical advice.