Medication Response

Why your genes affect medication response

Medication response varies between people because the proteins your body uses to absorb, transport, metabolize, and respond to a drug are all encoded in your DNA. Inherited variation in those genes changes how fast your liver clears a drug, how much reaches the target tissue, and how strongly the target responds. Standard prescribing assumes a baseline genotype that fits most people, not all. When the assumption misses, the consequence ranges from "the drug does nothing" to organ-system toxicity.

Three mechanisms drive most clinically actionable variation.

Metabolism

The cytochrome P450 enzymes (CYP2C9, CYP2C19, CYP2D6, CYP3A4) process the majority of prescription drugs. Variant alleles produce enzymes that work slower than baseline (poor metabolizer), faster (rapid or ultrarapid), or with altered substrate preference. A poor metabolizer can accumulate toxic levels at a standard dose; an ultrarapid metabolizer can clear a drug so fast the therapeutic window is never reached.

Transport

Drugs move into and out of tissues via membrane transporters. SLCO1B1 governs liver uptake of statins. A reduced-function variant means more statin remains in the bloodstream and muscle tissue, raising myopathy risk at any given dose.

Drug target sensitivity

Some genes encode the protein the drug acts on. VKORC1 is the enzyme warfarin inhibits to thin blood. Variants change how much warfarin produces the same anticoagulation level. Historical practice dosed by a weight-based estimate then titrated by INR, which routinely under- or over-anticoagulated patients during the dose-finding window.

Pharmacogenomics is the area of clinical genetics where variant-level knowledge has the highest immediate stakes. Knowing your variant status before a prescription is written shifts the conversation from trial-and-error to informed dose selection. The entries below are drawn from CPIC consensus guidelines, the standing authority on actionable gene-drug pairs.

What Varia reports

The table below summarizes the gene-drug pairs Varia checks for in your scan, drawn from CPIC consensus guidelines (Level A actionable). Your scan tells you which variants you carry; this table tells you which medications those variants intersect with.

Each row above corresponds to a pharmacogenomic finding in your scan if Varia detects the relevant variant in your file. The Physician section of the Varia Genomic Brief formats your specific genotype results into a one-page-per-variant document for your prescriber, including the rsID, your genotype, and 2 to 4 supporting citations. The brief names your variant status and the relevant gene-drug pair; it does not make prescribing recommendations.

What Varia cannot detect with consumer data

Consumer genetic data (SNP arrays from 23andMe/AncestryDNA, variants-only VCF exports from clinical WGS labs) was not designed for pharmacogenomic profiling. It captures common single-nucleotide variants reliably and misses three classes of pharmacogenomic information that matter clinically.

CYP2D6 copy number variation. CYP2D6 metabolizes codeine, tramadol, oxycodone, paroxetine, fluoxetine, venlafaxine, metoprolol, tamoxifen, and roughly a quarter of all prescribed drugs. Resolving CYP2D6 metabolizer status requires counting gene copies (deletions produce poor metabolizers; duplications produce ultrarapid metabolizers) and disambiguating CYP2D6 from the nearby CYP2D7 pseudogene. SNP arrays cannot count copies; standard WGS variant calling does not separate CYP2D6 from CYP2D7. Specialized PGx panels (long-range PCR or targeted assays) are the standard. If CYP2D6-driven drugs are on the table for you, ask your prescriber whether a clinical PGx panel is indicated. Varia cannot substitute.

Full HLA typing. Hypersensitivity reactions to abacavir (HLA-B*5701), carbamazepine (HLA-B*1502, HLA-A*3101), and allopurinol (HLA-B*5801) are CPIC-actionable when the relevant HLA allele is present, and serious enough that pre-prescription typing is standard of care in some populations. The HLA region is too polymorphic for SNP-array coverage to call alleles reliably; clinical-grade HLA typing uses targeted sequencing pipelines. Varia does not attempt HLA allele calls.

Gene-level deletions and other structural variants. Whole-gene deletions (CYP2D6 *5 deletion, GSTM1 null, UGT2B17 null), large insertions, and other structural variation typically do not appear in SNP-array data or in standard VCF variant calls. WGS can detect them with specialized pipelines beyond what most clinical labs return.

If your prescriber raises any of these gene-drug pairs, treat Varia as a starting point and a complement to clinical PGx testing, not a replacement.

What to do with these results

Pharmacogenomic variants matter at one specific moment: when a prescription decision is being made. The point of having your variant status in hand is not to act on it independently, it is to bring it into the conversation when a drug from one of the affected classes is being considered or is already in your regimen.

Generate a Variant Evidence Summary. After your scan, Varia produces a PDF literature and provenance summary for clinician review. Each pharmacogenomic finding gets one page with your observed genotype, rsID, gene, native-scale drug-response facets (CPIC level and metabolizer phenotype as cited information), authority chain, and 2 to 4 supporting citations to CPIC guidelines and primary literature with VEAS annotations. It reports what sources describe; it does not prescribe a dose or recommend a substitution.

Take it to the conversation. Bring the summary to the prescriber when any of the gene-drug pairs in Section 2 intersect your care plan: a new statin, an anticoagulant decision, an antiplatelet question after a cardiac event, an antidepressant selection, a PPI longer than the OTC label allows. The summary gives the prescriber the genotype data they would otherwise have to order a separate test to obtain, and the citation set they would otherwise have to assemble from CPIC themselves.

Do not self-adjust prescriptions based on Varia output. Varia does not have your full clinical picture: your other medications, your liver and kidney function, the indication being treated, your renal and hepatic dosing context. The variant is one input among many that informs a dose decision, and the dose decision belongs to the prescriber.

Ask about clinical PGx testing for the gaps. If the medication being considered hits one of the classes Section 3 named (CYP2D6-driven drugs like codeine, tramadol, many antidepressants and antipsychotics; HLA-flagged drugs like abacavir, carbamazepine, allopurinol), ask the prescriber whether a clinical-grade PGx panel is indicated. Varia covers what consumer data can resolve. Clinical PGx covers the rest.

References and guidelines

The gene-drug pairs above are drawn from CPIC consensus guidelines (Level A actionable), cross-referenced against the FDA Table of Pharmacogenomic Biomarkers in Drug Labeling and PharmGKB clinical annotations. Star-allele designations follow PharmVar nomenclature. The full reference list of authorities Varia cites across its database is on the Sources page.