What is Pharmacogenomics?

The science of how your genes affect your response to medications — choosing the right drug, at the right dose, guided by DNA.

Try the search My PGx Report See dashboard

💡 The big idea

"

Two people take the exact same dose of codeine for the exact same headache. One feels relief in 30 minutes. The other feels nothing. A third has dangerously slow breathing within an hour.

Why? Their genes are different. Specifically, the gene CYP2D6 that converts codeine into morphine in the body. Some people have variants that make this gene hyperactive (rapid metabolizers). Others have variants that make it inactive (poor metabolizers).

Pharmacogenomics is the science that turns this insight into actionable medicine.

🔗 From DNA to drug response: the chain

In five steps, how a single "typo" in your cells decides whether your prescription works:

1. DNA — your body's instruction book

Every cell carries a ~3-billion-letter instruction book written in just four letters (A T G C). Everyone's book is 99.9% identical — the remaining 0.1% makes the difference.

2. Gene — one recipe in the book

A gene is a recipe for one specific job. The CYP2D6 gene, for example, is the recipe for a liver enzyme that breaks down drugs. Humans have ~20,000 genes; a few hundred of them influence drug response.

3. Enzyme — the worker the recipe builds

Your body reads the recipe and builds a "worker" (a protein/enzyme). Most drug enzymes work in the liver: they either activate drugs or break them down so the body can clear them.

4. Variant — a typo in the recipe

A variant is a one-letter change in the recipe. Some typos don't matter; others make the worker slower, faster, or broken entirely. Every known typo has an address: an "rs" number like rs4244285.

5. Drug response — the outcome

If the worker is too slow, the drug piles up (side-effect risk) or never gets activated (no effect). Too fast, and the drug is cleared before it works — or too much active form is produced. The right dose depends on your worker's speed.

📖 The dictionary: every term, in plain words

Every concept you'll meet on these pages — each with an everyday analogy and a real example.

Variant & rsID

A single-letter difference in DNA (also called a SNP). The rsID is the universal address science gives that difference — the same number in every study worldwide.

💭 Think of it as the page-and-line number of a typo in a book.

rs4244285 — the most common "broken enzyme" variant in CYP2C19

Allele

You carry two copies of every gene: one from your mother, one from your father. Each version of a gene is called an "allele".

💭 Two copies of the same recipe — one from each parent.

Genotype AG = one copy carries A, the other carries G

Star allele (2, 17…)

Shorthand names for known variant combinations in a gene. *1 usually means the "normal" version; names like *2 or *3 describe specific packages of changes.

💭 A version number for the recipe — v1 is original, v2 is modified.

CYP2C19*2 = the version that makes the enzyme non-functional

Diplotype (1/2)

Your two alleles written together. It's your personal "version pair" for a gene, and it determines your phenotype (your enzyme speed).

💭 The version numbers of both your recipe copies, side by side.

*1/*2 = one normal + one broken copy → intermediate speed

Phenotype

The observable result of your genotype: how fast your enzyme actually runs in the real world. This is what drives dosing decisions (see the speed scale below).

💭 Not the engine's spec sheet — its actual speed on the road.

Poor / Intermediate / Normal / Rapid / Ultra-rapid Metabolizer

Activity score

Each allele gets a score (normal = 1, reduced = 0.5, broken = 0) and the two are added up. The total maps to the phenotype: 2 = normal, 0 = poor metabolizer.

💭 The combined performance rating of your two workers.

*1/*2 → 1 + 0 = 1.0 → Intermediate Metabolizer

CYP450 enzymes

Your liver's drug-processing family. About 75% of prescription drugs pass through a handful of them: CYP2D6, CYP2C19, CYP2C9, CYP3A4. See a "CYP" name? It's about drug metabolism.

💭 Specialist crews in the liver's recycling plant.

CYP2D6: codeine, tamoxifen, many antidepressants

Transporters (SLCO1B1)

Some genes build "doorman" proteins that carry drugs into cells rather than breaking them down. SLCO1B1 carries statins into the liver — if it underperforms, statin builds up in blood and muscle-damage risk rises.

💭 A forklift moving goods into the warehouse — too slow, and goods pile up on the dock.

SLCO1B1 phenotypes use "function" terms: Poor / Decreased / Normal Function

HLA

Your immune system's ID-checking system. Certain HLA versions mistakenly flag specific drugs as "threats", triggering life-threatening allergic reactions. There is no dose adjustment: the drug is avoided entirely.

💭 An overzealous security guard who mistakes an innocent visitor for an intruder.

HLA-B*57:01 positive → abacavir is CONTRAINDICATED

CPIC

The Clinical Pharmacogenetics Implementation Consortium: the world's leading experts writing the actual "with this genotype, use this drug like this" rules. Level A = the strongest, actionable evidence.

💭 The committee that writes the official rulebook for genetic prescribing.

Drugly syncs these rules directly from CPIC monthly

PharmGKB / ClinPGx

The Stanford-based gold-standard knowledge base curating 25+ years of pharmacogenomic research. It's the source of Drugly's gene, variant, and annotation data.

💭 The peer-reviewed encyclopedia of the whole field.

Every annotation links to real studies on PubMed

Level of evidence (1A–4)

How solidly a gene-drug relationship is proven. 1A = replicated, guideline-backed, strong evidence. 3–4 = early findings, not yet enough for clinical decisions.

💭 Forecast confidence: 1A is "rain, guaranteed"; 4 is "might get cloudy".

In your reports, 1A/1B variants are flagged as "high-impact"

🏎️ The metabolizer speed scale

For enzyme genes, your phenotype lands on one of five speeds. The same dose behaves very differently at each:

PM

Poor Metabolizer

Enzyme doesn't work. Drug piles up — or a prodrug never activates.

IM

Intermediate Metabolizer

Slower than normal. Often needs a dose adjustment.

NM

Normal Metabolizer

Standard doses are designed for this speed.

RM

Rapid Metabolizer

Faster than normal. Some drugs clear before they can work.

UM

Ultra-rapid Metabolizer

Very fast. With prodrugs, dangerous levels of active drug can form.

The critical detail — direction depends on the drug: Codeine is a "prodrug": CYP2D6 must convert it into morphine before it works. A poor metabolizer gets no pain relief; an ultra-rapid metabolizer makes too much morphine (danger!). For a drug given in its active form, the picture flips: poor metabolizers accumulate it, rapid ones clear it too soon. The same "slow engine" produces two opposite outcomes depending on the drug — which is exactly why CPIC rules are drug-specific.

📊 Three things to know

~99% of people

have at least one variant that affects how they respond to a common medication.

30-50% of adverse reactions

are estimated to be preventable with pharmacogenomic testing.

200+ FDA-labeled drugs

have pharmacogenomic information on their official labels (and that number keeps growing).

🧭 The five things you can search

Drugly.AI organizes pharmacogenomic data into five interlinked entity types. From any one, you can navigate to the others.

Medications with known pharmacogenomic considerations. You'll see CPIC dosing recommendations, FDA/EMA label info, clinical evidence, and which genes affect each drug.

Try: warfarin, clopidogrel, codeine

Human genes whose variations affect drug response. VIP (Very Important Pharmacogene) genes have the strongest evidence. Each gene page shows all drugs it affects.

Try: CYP2C9, CYP2D6, VKORC1

Specific genetic differences (SNPs, haplotypes, star alleles). Each variant page lists studies that found drug-response effects, with PubMed links.

Try: rs1799853, rs4244285, *1/*3

Diseases, side effects, and outcomes that pharmacogenomics affects: drug efficacy, toxicity, hypersensitivity reactions. Linked to drugs and genes.

Try: myopathy, Stevens-Johnson, bleeding

Visual diagrams from PharmGKB showing how drugs move through and act on the body, gene by gene, step by step.

Try: Browse all pathways →

🎯 Four ways to use it

Whatever you start with — a drug, a gene, a variant, or a DNA file — you'll find what you need.

A

"I want to know about a specific drug"

Search for the drug. You'll see all PGx-relevant genes for it, dose recommendations from CPIC, regulatory label info from FDA / EMA / others, and the strongest clinical evidence.

Try with warfarin

B

"I want to know about a specific gene"

Search for the gene symbol. You'll see all drugs it affects, the variants in that gene with the most evidence, and clinically actionable summary annotations.

Try with CYP2D6

C

"I want to research a specific variant or rsID"

Search by rsID. You'll see the PMID-linked studies that found drug-response effects, the drugs affected, and any clinical haplotype pairings.

Try with rs1799853

D

"I want to interpret a DNA file"

Upload a 23andMe, AncestryDNA or VCF file. Drugly calls the star alleles, assigns metabolizer phenotypes, flags HLA risks — then "Check my drugs" matches that genotype against real CPIC dosing recommendations.

Open My PGx Report

👥 Who uses this and how

Pharmacists

Patient counseling: Explain why a patient might respond differently to their new prescription.
Drug selection: When two drugs are options, check which has clearer PGx guidance for the patient's genotype.
Dose adjustment: Verify CPIC-recommended dose modifications for known metabolizer phenotypes.
Adverse event review: Investigate unexpected side effects — is there a genetic explanation?
Continuing education: Stay current with FDA biomarker label additions across major drug classes.

Researchers & students

Literature review: Find PMID-anchored variant annotation studies for a hypothesis-generating gene.
Cross-population studies: Use specialty-population tags on summary annotations to find relevant cohorts.
Pathway exploration: See visual diagrams of how a drug moves through the body, gene by gene.
Teaching: Show pharmacogenomic concepts with real examples and beautiful interactive diagrams.
Grant preparation: Quickly survey the evidence landscape for a drug-gene pair you're proposing to study.

⚠️ Important to know

Educational reference, not clinical advice

Drugly.AI's pharmacogenomic module is a knowledge reference tool designed for healthcare professionals and students. It is not a clinical decision support system and does not replace professional judgment.
Always consult licensed healthcare providers — pharmacists, physicians, clinical pharmacologists — for any medication decisions.
Data is sourced from PharmGKB / ClinPGx 2025, the gold standard for pharmacogenomic research; CPIC dosing recommendations sync monthly, directly from the CPIC API.
Evidence levels (LOE 1A through 4) reflect the strength of supporting research, not the certainty of an individual response. Even Level 1A guidelines describe population trends.
Pharmacogenomic testing in clinical practice requires certified laboratory testing and proper interpretation by qualified clinicians. Consumer DNA tests (23andMe etc.) are screening-level; confirm before any clinical decision.

🚀 Ready to explore?

Start with a single search — a drug name, a gene symbol, a variant rsID — and let the connections guide you.

Open PGx Search