Methodology

Overview

DecodeMyBio transforms consumer DNA raw data into pharmacogenomic reports through a multi-step pipeline. Each step follows established standards in the pharmacogenomics field. This page describes the technical process in detail.

Step 1: File Parsing and Variant Extraction

Raw data files from 23andMe, AncestryDNA, MyHeritage, and FamilyTreeDNA follow different formats but share a common structure: each line maps a SNP identifier (rsID) to a genotype call. DecodeMyBio parses these files and extracts genotype calls for pharmacogenomically relevant positions.

We maintain a curated list of clinically relevant rsIDs drawn from PharmVar allele definitions and CPIC gene-specific tables. Only positions with established pharmacogenomic significance are extracted.

Step 2: Star Allele Mapping

Pharmacogenes are described using a star allele nomenclature system (e.g., CYP2C19*2, CYP2D6*4). Each star allele is defined by a specific combination of variant positions. DecodeMyBio maps extracted genotypes to star alleles using allele definition tables from PharmVar.

Consumer genotyping arrays cover a subset of defining variants for each star allele. When a defining variant is not genotyped, it is assumed to carry the reference (wild-type) allele. This is a standard assumption in consumer PGx analysis, but it means some rare alleles may not be detected. See our Limitations page for details.

Step 3: Diplotype and Phenotype Assignment

Once star alleles are identified for each gene, DecodeMyBio determines the diplotype (the combination of alleles on each chromosome) and translates it to a metabolizer phenotype. Phenotype categories follow CPIC standardized terminology:

• Ultrarapid Metabolizer (UM)
• Normal Metabolizer (NM)
• Intermediate Metabolizer (IM)
• Poor Metabolizer (PM)

Activity scores are used where applicable (e.g., CYP2D6), following the CPIC-standardized activity score system.

Step 4: Drug-Gene Interaction Lookup

Each phenotype is matched against CPIC and DPWG drug-gene interaction guidelines. For example, the clopidogrel–CYP2C19 interaction has CPIC Level A evidence. Only interactions with CPIC evidence Level A (strong) or Level B (moderate) are included in reports. Each interaction entry cites the specific guideline, gene, phenotype, and therapeutic recommendation.

Step 5: Report Generation

The final Medication Safety Report presents findings organized by gene, with each gene section showing the detected diplotype, metabolizer phenotype, and any associated drug-gene interactions. All findings link to the underlying variant, gene, and source guideline.

CPIC Evidence Levels

CPIC classifies pharmacogenomic evidence into levels that indicate the strength of the gene-drug association:

• Level A: Strong evidence — genetic information should be used to change prescribing of the affected drug.
• Level B: Moderate evidence — genetic information could be used to change prescribing of the affected drug.
• Level C/D: Limited or insufficient evidence — not included in DecodeMyBio reports.

Limitations

• Consumer genotyping arrays are not equivalent to full sequencing. They test a predefined set of positions and may miss novel or rare variants.
• Structural variants — gene deletions, duplications, and hybrid rearrangements — may not be detected. This is particularly relevant for the CYP2D6 gene, where structural variants are both common and clinically significant.
• Rare alleles not included in PharmVar allele definition tables may be missed, potentially resulting in a default *1 (normal-function) assignment.
• Clinical prescribing decisions should be made by a healthcare provider who can consider pharmacogenomic results alongside medical history, current medications, and other clinical factors.

For a complete discussion of analytical boundaries, see our Limitations page.

Updates and Evidence Review

DecodeMyBio's analysis pipeline and content are reviewed periodically and updated when CPIC, DPWG, or other major guideline bodies publish new or revised recommendations. We monitor PharmVar for new allele definitions and update our allele definition tables accordingly. See our editorial policy for details on our content standards and review process.

If you believe any content on this site contains an error, please contact us at support@decodemybio.com.