TruthNexus
Pharmacogenomics · Medical

Genotype into prescribing action

7% of prescriptions involve a gene-drug pair where CPIC guidelines require dose adjustment or an alternative. 90% of patients with pharmacogenomic test results never have them applied at the point of prescribing. The data exists. The decision support doesn't — until now.

POST /api/check · 200 OK
Clopidogrel
CYP2C19 · *2/*2
Use Alternative
Phenotype
Poor Metabolizer
Confidence
0.95
CPIC Level A · ACTIONABLE

Select alternative antiplatelet. CYP2C19 *2/*2 poor metabolizers cannot convert clopidogrel to its active metabolite — inadequate platelet inhibition increases MACE risk.

Alternative: Prasugrel or ticagrelor
CPIC ×3PharmGKB ×58 findings · audit record generated
21,000+
Curated PGx findings from CPIC and PharmGKB
7%
Of prescriptions involve actionable gene-drug pairs
30%
Of serious ADEs in genotyped patients are preventable
11
Pharmacogenes with CPIC Level A clinical action guidance
See It In Action

What a response looks like

Select a gene-drug scenario to see a representative API response — recommendation tier, phenotype, dosing guidance, and sourced evidence counts.

POST /api/check · application/jsonsimulation
Scenarios
Clopidogrel
CYP2C19 · *2/*2
Use Alternative
Phenotype
Poor Metabolizer
Confidence
0.95
CPIC Level A · ACTIONABLE

Select an alternative antiplatelet agent. CYP2C19 *2/*2 poor metabolizers cannot convert clopidogrel to its active metabolite — platelet inhibition is inadequate, increasing MACE risk at standard dosing.

Recommended alternative: Prasugrel or ticagrelor (not CYP2C19-dependent)
CPIC ×3PharmGKB ×58 findings

Simulated output representative of real API responses. Production results sourced live from CPIC guidelines and PharmGKB annotations.

What It Does

Six capabilities. One API.

POST /api/check

Drug + Gene Check

Returns recommendation tier (ACTIONABLE / INFORMATIVE / PRELIMINARY), calibrated confidence score, and sourced CPIC/PharmGKB evidence for any drug-gene pair. Accepts an optional genotype for diplotype-aware results in the same call.

POST /api/dosing

Diplotype Dosing

Translates a patient's diplotype (e.g. CYP2C19 *2/*2) into a specific CPIC dosing action — contraindicated, use alternative, dose reduction, dose increase, or normal. Deterministic CPIC table lookup, not a model inference.

POST /api/panel

Multi-Gene Panel Checker

Cross-references a complete gene panel against a full medication list in one call. Returns per-drug, per-gene dosing guidance across the entire panel — surfacing every actionable interaction, not just the highest-priority one.

GET /api/gene/{name}

Gene Profile

Returns all drugs affected by a specific gene, ranked by evidence strength and recommendation tier. CYP2D6 alone influences 25% of all marketed drugs — Gene Profile maps the full prescribing landscape for any gene in the supported panel.

POST /api/gene-drug

Gene–Drug Crosswalk

Bidirectional lookup: given a gene, return all affected drugs; given a drug, return all relevant genes with evidence strength and phenotype implications. Useful for population-level screening and formulary analysis.

GET /api/pgx-drugs

PGx Drug Index

Full index of drugs with CPIC or PharmGKB evidence in the evidence base. Returns gene associations, highest-evidence tier, and CPIC level for each drug — useful for formulary gap analysis and prior-authorization workflows.

CPIC Level A

Clinically actionable gene–drug pairs

The pairs where wrong prescribing causes serious harm — and where a genotype result changes the decision.

ClopidogrelCPIC-A
CYP2C19

Antiplatelet — *2/*2 poor metabolizers cannot activate the drug; MACE risk without genotype-guided alternatives

Population · ~2–3% European · ~15% East Asian
CodeineCPIC-A
CYP2D6

Opioid — ultrarapid metabolizers convert to toxic morphine doses; contraindicated under CPIC guidelines

Population · ~7–10% European ultra-rapid metabolizers
WarfarinCPIC-A
CYP2C9 + VKORC1

Anticoagulant — genotype-guided dosing required; combined variants cause 5× dose variance

Population · ~35–40% carry a variant affecting warfarin dose
TamoxifenCPIC-A
CYP2D6

Oncology — poor metabolizers produce minimal active endoxifen, significantly reducing efficacy

Population · ~7% European poor metabolizers
FluorouracilCPIC-A
DPYD

Oncology — DPYD*2A carriers face life-threatening toxicity at standard doses; dose reduction mandatory

Population · ~1–2% carry a DPYD risk variant
AzathioprineCPIC-A
TPMT

Immunosuppressant — poor metabolizer accumulates toxic thiopurine metabolites; severe myelotoxicity

Population · ~10% intermediate · ~0.3% poor metabolizers
AbacavirCPIC-A
HLA-B*57:01

HIV antiretroviral — HLA-B*57:01 carriers face potentially fatal hypersensitivity reaction; CONTRAINDICATED. FDA black box warning. Pre-screening mandatory.

Population · ~5–8% European · ~1% South Asian · ~1% African
CarbamazepineCPIC-A
HLA-B*15:02

Antiepileptic — HLA-B*15:02 carriers face severe cutaneous reactions (SJS/TEN); CONTRAINDICATED. FDA black box warning for Asian ancestry patients.

Population · ~6–10% East/Southeast Asian · rare in European
SimvastatinCPIC-A
SLCO1B1

Statin — decreased-function variants impair hepatic uptake, raising plasma exposure and risk of myopathy; limit to ≤20 mg/day or switch to rosuvastatin.

Population · ~15% carry SLCO1B1 *5 variant
The combinatorial problem

A patient with 3 genes tested and 4 active medications has 12 gene-drug combinations to evaluate. Panel Checker returns all of them — with per-drug, per-gene dosing guidance — in one call.

4
combos
1G × 4Rx
12
combos
3G × 4Rx
50
combos
5G × 10Rx
150
combos
10G × 15Rx
At population scale

In a health system with 500,000 members, approximately 35,000 patients are currently on at least one medication with a CPIC Level A gene-drug interaction. Without systematic PGx integration at the point of prescribing, most of those interactions go unchecked every time the prescription is renewed.

~7%
of all prescriptions, every day
Evidence Framework

Three tiers, calibrated by evidence level

Every result is classified by the strength of underlying CPIC and PharmGKB evidence — so clinicians know whether to act, consider, or monitor.

Actionable
CPIC Level A / B · PharmGKB Level 1

Genotype-guided prescribing is required or strongly recommended. Wrong prescribing at this tier causes measurable patient harm — the gene result must change the drug or dose before the prescription is written.

Informative
CPIC Level C · PharmGKB Level 2–3

Credible evidence of gene-drug effect. Consider genotyping for high-risk patients, high-cost therapies, or therapeutic areas where response variability is clinically significant.

Preliminary
CPIC Level D · PharmGKB Level 4

Early-stage evidence. No prescribing action required. Monitor response and flag for review as evidence matures — useful for research, registry programs, and clinical trial stratification.

Known preventable events

CYP2D6 ultra-rapid metabolizers converting codeine to toxic morphine doses. Warfarin over-anticoagulation in CYP2C9 poor metabolizers. Clopidogrel treatment failure in CYP2C19 poor metabolizers. Abacavir hypersensitivity in HLA-B*57:01 carriers — potentially fatal on re-challenge.

These are known, preventable events that continue to occur because pharmacogenomics is not systematically integrated at point of care. Skippy Pharmacogenomics closes that gap — surfacing the right action at the moment of prescribing.

The integration gap
“Over 90% of patients who have had pharmacogenomic testing do not have their results systematically applied at the point of prescribing.”

CPIC / PharmGKB Implementation Consortium · Clin Pharmacol Ther, 2023

The data exists — in genotyping reports, EHR problem lists, lab systems. It is almost never surfaced when a prescriber is writing the order. Skippy Pharmacogenomics closes that gap at themedication-prescribehook — before the prescription is written, not after.

Who It's For

Health systems, PBMs, and labs

Health Systems
Integrate PGx alerts into EHR prescribing workflows — surfaced at the point of order, not buried in a genomics report.
Pharmacy Benefit Managers
Run population-level pharmacogenomics screening to identify members at risk from known gene-drug interactions at formulary level.
Specialty Pharmacies
Add genotype-guided dosing recommendations to oncology, cardiology, and psychiatry medication reports.
Clinical Laboratories
Append prescribing action guidance to genotype reports, turning a test result into a clinical recommendation.
Regulatory

CPIC guidelines, citable in prescribing records

All recommendations sourced from CPIC peer-reviewed guidelines (Level A = strongest evidence) and PharmGKB clinical annotations (Level 1 = highest clinical impact). Confidence scores are weighted by evidence level with documented methodology.

Output is citable in prescribing documentation as a pharmacogenomics-aware safety check. Every check generates an audit record documenting that a PGx-guided review was performed — relevant for malpractice documentation and prescribing liability. HIPAA-ready with BAA available.

HIPAA ReadyCPIC Level A/BPharmGKB Level 1Audit Record Per Check
Why Skippy PGx

Not a drug database. Not an LLM.

Deterministic CPIC table lookup

Recommendations come from the CPIC dosing table — not a language model inference. Diplotype *2/*2 in → "Use Alternative" out. Same answer every time, citable in prescribing records.

Panel × medication matrix in one call

A patient on 4 medications with 3 genes tested has 12 gene-drug combinations to evaluate. One call to /api/panel returns all of them — with per-drug, per-gene action guidance and a priority-ranked summary.

Audit record per check

Every check generates an immutable record: drug queried, gene tested, diplotype, recommendation returned, and timestamp. Citable in prescribing documentation. Relevant to malpractice liability.

Calibrated confidence, not a severity flag

CPIC Level A returns confidence 1.00. Level D returns 0.55. The tier (ACTIONABLE / INFORMATIVE / PRELIMINARY) tells prescribers whether to act now, consider, or monitor — not just whether an interaction exists.

EHR Delivery

Embedded at the moment of prescribing

PGx guidance delivered via CDS Hooks — the HL7 standard for real-time clinical decision support inside Epic, Cerner, and other EHR workflows. Clinicians see the alert without leaving the prescribing screen.

medication-prescribe
At prescription time

Genotype-guided dosing alert fires when a prescriber writes the order — before it's signed. The highest-impact integration point.

patient-view
On chart open

Surface unresolved PGx flags when a clinician opens the patient record — useful for medication reconciliation and rounds.

order-review
At order review

Pharmacist-facing alerts at order verification — catch gene-drug conflicts before dispense, not after.

FHIR R4
Native output format
< 90 days
Agreement to live in Epic
REST API
Direct integration available
Clinical Use

Skippy Pharmacogenomics is a decision-support tool for clinical professionals. It does not replace laboratory genotyping, clinical pharmacist review, or physician judgment. Recommendations should be validated against current institutional protocols and patient-specific factors before prescribing. HIPAA-ready; BAA available.

See Pharmacogenomics in your prescribing workflow

We work with health systems, PBMs, specialty pharmacies, and clinical labs. Let's talk about integrating genotype-guided prescribing.