How Pharmacogenomic Testing Works

If you've taken a consumer DNA test from 23andMe, AncestryDNA, or a similar service, your raw data file contains hundreds of thousands of genetic variants — including many that are relevant to how you metabolize medications. Here's how that data gets transformed into a pharmacogenomic report.

Step 1: Raw Data and Genotyping Arrays

Consumer DNA tests use genotyping arrays — chips that test for specific known genetic variants (SNPs, or single nucleotide polymorphisms) at predetermined positions across your genome. Your raw data file is essentially a list of these positions and the genotype (the two alleles you carry) at each one.

For example, one line in your raw data might show that at position rs4244285 on chromosome 10, your genotype is CT. This particular variant is part of the CYP2C19 gene *2 allele — a well-known pharmacogenomic variant.

Step 2: Extracting Pharmacogenomic Variants

Not all of the hundreds of thousands of variants in your raw data are pharmacogenomically relevant. The first step in analysis is to extract only the variants that are known to affect drug metabolism or response.

These relevant positions are defined by organizations like PharmVar (the Pharmacogene Variation Consortium), which maintains the official allele definitions for pharmacogenes. Each pharmacogene has a set of "defining variants" — specific genetic positions that determine which star allele you carry.

Step 3: Star Allele Assignment

Pharmacogenes use a star allele nomenclature system. The normal (reference) allele is typically designated *1. Variant alleles are numbered sequentially: *2, *3, *4, and so on. Each star allele is defined by a specific combination of genetic variants.

For example, CYP2C19*1 is the normal-function allele. CYP2C19*2 (defined by rs4244285) is a no-function allele. CYP2C19*17 (defined by rs12248560) is an increased-function allele.

The analysis maps your genotypes at the defining positions to determine which star alleles you carry for each pharmacogene.

Step 4: Diplotype and Phenotype

Since you have two copies of each gene (one from each parent), the combination of your two star alleles is called your diplotype. For example, CYP2C19 *1/*2 means you have one normal allele and one no-function allele.

Your diplotype translates to a metabolizer phenotype using CPIC-standardized categories:

• Normal Metabolizer (NM) — Expected enzyme activity. Standard drug doses typically apply.
• Intermediate Metabolizer (IM) — Reduced enzyme activity. Some drugs may need dose adjustments.
• Poor Metabolizer (PM) — Significantly reduced or absent enzyme activity. Higher risk of toxicity or therapeutic failure for certain drugs.
• Ultrarapid Metabolizer (UM) — Increased enzyme activity. Drugs may be metabolized too quickly, reducing effectiveness or increasing active metabolite levels.

Step 5: Drug-Gene Interaction Lookup

Once your phenotype is determined for each gene, it is matched against clinical guidelines published by CPIC and DPWG. These guidelines specify, for each gene-drug combination, what clinical actions (if any) should be considered based on your phenotype.

For example, CPIC guidelines state that CYP2C19 poor metabolizers should consider an alternative to clopidogrel because the drug may not be effectively converted to its active form, potentially reducing its antiplatelet effect.

Step 6: The Report

The final report presents your results organized by gene, showing your diplotype, metabolizer phenotype, and any clinically significant drug-gene interactions. Each finding references the underlying variant, the clinical guideline, and the evidence level.

This report is designed to be shared with your healthcare provider, who can interpret the results in the context of your full medical history and current medications. For a guide to reading your results, see Understanding Your Medication Safety Report.

What Consumer Testing Cannot Do

Consumer genotyping arrays test for a predefined set of variants. They do not sequence your full genome, which means some rare or novel variants may not be detected. Structural variants like gene deletions or duplications (particularly important for the CYP2D6 gene) cannot be reliably identified from array data alone. For these reasons, consumer-grade analysis is informational and not a replacement for clinical pharmacogenomic testing.

Get Your Medication Safety Report

Upload your raw DNA data from 23andMe, AncestryDNA, or other consumer tests to get your personalized Medication Safety Report with gene-by-gene results and CPIC-backed drug-gene interactions. For step-by-step download and upload instructions, see how to upload your AncestryDNA or 23andMe raw data.