Deep mutational scanning of CYP2C9, CYP2C19, and NUDT15 shows that pharmacogene variant interpretation requires assay-specific functional data
- 1Faculty of Engineering, Department of Bioengineering, Bilecik Şeyh Edebali University, TR, Bilecik 11100, Türkiye
Abstract
Pharmacogene missense variants can disrupt protein stability, catalytic competence, or substrate handling through distinct mechanisms. General-purpose predictors estimate clinical pathogenicity as a single scalar, whereas pharmacogene interpretation requires knowing which biochemical dimension a variant perturbs, since that determines whether reduced function is substrate-dependent. Five deep mutational scanning datasets comprising 26,198 missense variants across CYP2C9, CYP2C19, and NUDT15 were assembled from MaveDB. Paired assays showed that this dimensionality dominates the data: 28% of CYP2C9 variants (1,236 of 4,421) decoupled catalytic activity from abundance, and 48% of NUDT15 variants (1,364 of 2,844) decoupled thiopurine sensitivity from stability, with CYP2C9 discordance concentrating at substrate-channel residues. AlphaMissense, a representative general-purpose pathogenicity predictor, scored these classes in line with its clinical training objective rather than the assayed biochemistry, assigning likely-benign scores to 38 of 195 stable-but-dead CYP2C9 variants and likely-pathogenic scores to 140 of 222 destabilized but thiopurine-resistant NUDT15 variants. To test whether this dimensionality is recoverable, a supervised ESM-2 sequence baseline was benchmarked against the ESM1v zero-shot ensemble and AlphaMissense under position-based 5-fold cross-validation, together with 3 architectural extensions: AlphaFold structural features, multitask learning across paired assays, and contact-graph neural networks. The baseline reached Pearson r of 0.54–0.72, matching or marginally exceeding both comparators, and no extension improved upon it. Trained directly on each assay, it nonetheless recovered the paired-assay difference at r = 0.28 for CYP2C9 and 0.43 for NUDT15, separating discordant variants at AUROC 0.60 and 0.51. Pharmacogene interpretation therefore requires assay-specific, substrate-aware functional measurements rather than a single generic score.
Related
- Unraveling the EMT Spectrum in Aggressive Breast Cancer: Bridging Molecular Hybrid States with Functional Assay Interpretation
- 1078 Mutational signatures as mechanistic biomarkers of tumor immunity: associations with TMB, PD-L1 and immune phenotypes across cancer types
- Germline structural variation involving gynecologic and breast cancers reveals polygenic signatures of tumor mutational load
- Mutational profiles of hepatocellular carcinoma that arose in outbred mice after heavy-ion particle and γ-ray radiation
- Novel biallelic PATL2 variants causing predominant germinal vesicle arrest and recurrent ART failure: expanded mutational spectrum and evidence-based reproductive counseling guidance
BibTeX
@article{kırboga2026deep,
title = {{Deep mutational scanning of CYP2C9, CYP2C19, and NUDT15 shows that pharmacogene variant interpretation requires assay-specific functional data}},
author = {Kevser Kübra Kırboğa},
journal = {G3: Genes | Genomes | Genetics},
year = {2026},
month = aug,
doi = {10.1093/g3journal/jkag248},
url = {https://doi.org/10.1093/g3journal/jkag248},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Deep mutational scanning of CYP2C9, CYP2C19, and NUDT15 shows that pharmacogene variant interpretation requires assay-specific functional data
AU - Kevser Kübra Kırboğa
PY - 2026
DA - 2026/08/29
JO - G3: Genes | Genomes | Genetics
DO - 10.1093/g3journal/jkag248
UR - https://doi.org/10.1093/g3journal/jkag248
AB - Pharmacogene missense variants can disrupt protein stability, catalytic competence, or substrate handling through distinct mechanisms. General-purpose predictors estimate clinical pathogenicity as a single scalar, whereas pharmacogene interpretation requires knowing which biochemical dimension a variant perturbs, since that determines whether reduced function is substrate-dependent. Five deep mutational scanning datasets comprising 26,198 missense variants across CYP2C9, CYP2C19, and NUDT15 were assembled from MaveDB. Paired assays showed that this dimensionality dominates the data: 28% of CYP2C9 variants (1,236 of 4,421) decoupled catalytic activity from abundance, and 48% of NUDT15 variants (1,364 of 2,844) decoupled thiopurine sensitivity from stability, with CYP2C9 discordance concentrating at substrate-channel residues. AlphaMissense, a representative general-purpose pathogenicity predictor, scored these classes in line with its clinical training objective rather than the assayed biochemistry, assigning likely-benign scores to 38 of 195 stable-but-dead CYP2C9 variants and likely-pathogenic scores to 140 of 222 destabilized but thiopurine-resistant NUDT15 variants. To test whether this dimensionality is recoverable, a supervised ESM-2 sequence baseline was benchmarked against the ESM1v zero-shot ensemble and AlphaMissense under position-based 5-fold cross-validation, together with 3 architectural extensions: AlphaFold structural features, multitask learning across paired assays, and contact-graph neural networks. The baseline reached Pearson r of 0.54–0.72, matching or marginally exceeding both comparators, and no extension improved upon it. Trained directly on each assay, it nonetheless recovered the paired-assay difference at r = 0.28 for CYP2C9 and 0.43 for NUDT15, separating discordant variants at AUROC 0.60 and 0.51. Pharmacogene interpretation therefore requires assay-specific, substrate-aware functional measurements rather than a single generic score.
SN - 2160-1836
ER -