Pharmacogenomics¶

Pharmacogenomics (PGx) examines how genetic variation affects an individual's response to medications. AIVA's PGx analysis automatically assigns star alleles to 88 pharmacogenes, predicts metabolizer phenotypes, and generates clinical drug recommendations based on CPIC guidelines.

PGx results help clinicians identify patients who may need dose adjustments, alternative medications, or enhanced monitoring based on their genetic profile.

How it works¶

PGx analysis runs automatically as part of the secondary analysis pipeline when you upload FASTQ files. The pipeline proceeds through the following stages:

1. Variant calling: Variants are called from the BAM file across all 88 pharmacogenes.
2. Star allele assignment: Each gene receives a diplotype (a pair of star alleles, e.g., *1/*4) based on detected SNVs.
3. Phenotype prediction: Metabolizer phenotypes are predicted from the diplotype (e.g., Poor Metabolizer, Normal Metabolizer).
4. Activity scoring: An activity score is calculated by summing the functional scores of each allele.
5. Drug recommendations: CPIC guidelines are consulted to generate prescribing recommendations for each gene's phenotype.
6. Multi-gene enrichment: Drugs that require phenotypes from two genes (e.g., fluvastatin requires both CYP2C9 and SLCO1B1) are identified and enriched with combined recommendations.

Supported genes¶

AIVA analyzes 88 pharmacogenes. These genes are organized into the following categories:

CYP450 enzymes¶

Cytochrome P450 enzymes are responsible for metabolizing the majority of clinically used drugs.

Phase II enzymes¶

These enzymes catalyze conjugation reactions that facilitate drug elimination.

Transporters¶

Drug transporters affect absorption, distribution, and elimination of medications.

Other pharmacogenes¶

Understanding results¶

Each gene in your PGx results includes the following fields:

Diplotype¶

The diplotype represents the two star alleles identified for a gene. Star alleles are standardized nomenclature for common haplotypes. For example, CYP2D6 *1/*4 indicates one normal-function allele (*1) and one no-function allele (*4).

Phenotype¶

The predicted metabolizer status based on the diplotype:

Activity score¶

A numeric value representing total enzyme activity, calculated by summing the functional scores of each allele. Higher scores indicate greater enzyme activity. Activity scores provide finer granularity than phenotype categories alone.

Allele functions¶

Each allele receives a functional classification:

• Normal Function: Standard activity
• Increased Function: Above-normal activity
• Decreased Function: Reduced activity
• No Function: Absent activity
• Uncertain Function: Insufficient evidence to classify

CNV detection¶

For genes with known structural variation (e.g., CYP2D6, GSTM1, GSTT1, UGT2B17), AIVA performs a separate copy number variation (CNV) analysis. CNV results are informational only and do not override the SNV-based diplotype. Possible CNV calls include Normal, WholeDel1, WholeDup1, and Tandem duplications.

Drug recommendations¶

AIVA generates prescribing recommendations based on CPIC (Clinical Pharmacogenetics Implementation Consortium) guidelines.

Single-gene recommendations¶

Most drug recommendations are based on a single gene's phenotype. For example:

• CYP2D6 Poor Metabolizer + codeine: "Use alternative analgesic; increased risk of reduced efficacy"
• CYP2C19 Ultrarapid Metabolizer + clopidogrel: "Standard dosing; expected normal response"

Multi-gene recommendations¶

Some drugs require phenotype information from two genes for accurate guidance. AIVA automatically detects these cases and provides combined recommendations. For example:

• Fluvastatin: Requires both CYP2C9 and SLCO1B1 phenotypes
• Warfarin: Benefits from combined CYP2C9 and VKORC1 information

Multi-gene recommendations appear in the results for both contributing genes and are labeled accordingly.