The potential role of large language models in addressing phenoconversion in pharmacogenomics
- 1Division of Rheumatology, University of California San Francisco , , ,
- 2University Institute of Molecular Pathology Biomarkers, Universidad de Extremadura , ,
Abstract
Pharmacogenomics (PGx) is a cornerstone of precision medicine, enabling individualized drug therapy by identifying genetic factors that influence drug disposition and response. However, genotype-based predictions frequently fail to reflect actual drug-metabolizing capacity because enzyme activity can be modified by non-genetic factors, including concomitant medications, inflammation, organ dysfunction, aging, smoking, pregnancy, and other environmental influences. Phenoconversion can generate substantial discrepancies between genotype-predicted and observed metabolic phenotypes, contributing to adverse drug reactions, therapeutic failure, and reduced clinical utility of pharmacogenomic testing. The identification of phenoconversion remains challenging because the relevant determinants are often dispersed across multiple sources of information, including medication records, laboratory data, diagnostic information, and unstructured clinical notes within electronic health records. Recent advances in large language models (LLMs) offer new opportunities to address this limitation by retrieving, organizing, and contextualizing heterogeneous clinical and pharmacogenomic information. Beyond identifying drug-drug-gene and drug-gene-environment interactions, LLM-enabled systems may support dynamic phenotype assessment by modeling metabolic capacity as a continuously evolving characteristic rather than a static genotype-derived category and by identifying relevant modifiers even for drugs without actionable pharmacogenomic recommendations. This review examines the biological basis and clinical significance of phenoconversion and discusses the potential role of LLMs in bridging the gap between genetic predisposition and real-world drug response. We highlight the emerging concept of dynamic phenotype models that combine genetic, clinical, and environmental determinants of drug metabolism and may support more precise, context-aware pharmacotherapy.
Related
- Evaluating large language models in patient education: a comparative analysis addressing frequently asked questions in peri-acetabular osteotomy
- Addressing benchmarking gaps in large language models for health and medicine with dynamic red-teaming
- Comparative Evaluation of Large Language Model Interfaces in Third Molar Surgery Complication Scenarios: Response Quality, Clinical Content, Potential Clinical Risk, Readability, and Externally Observable Response Latency—A Cross-Sectional Comparative Benchmark Study
- The Potential for Large Language Models in Eating Disorder Care: Opportunities for Education, Detection and Assessment, and Intervention
- Guideline adherence and expert-rated potential safety concerns in AI-generated dietary advice for patients with cirrhosis: a comparative evaluation of large language models.
BibTeX
@article{garciaagundez2026potential,
title = {{The potential role of large language models in addressing phenoconversion in pharmacogenomics}},
author = {Augusto García-Agúndez and José A. G. Agúndez},
journal = {Frontiers in genetics},
year = {2026},
month = oct,
doi = {10.3389/fgene.2026.1987469},
url = {https://europepmc.org/article/PMC/PMC13730465},
note = {CC BY},
}
RIS
TY - JOUR
TI - The potential role of large language models in addressing phenoconversion in pharmacogenomics
AU - Augusto García-Agúndez
AU - José A. G. Agúndez
PY - 2026
DA - 2026/10/21
JO - Frontiers in genetics
DO - 10.3389/fgene.2026.1987469
UR - https://europepmc.org/article/PMC/PMC13730465
AB - Pharmacogenomics (PGx) is a cornerstone of precision medicine, enabling individualized drug therapy by identifying genetic factors that influence drug disposition and response. However, genotype-based predictions frequently fail to reflect actual drug-metabolizing capacity because enzyme activity can be modified by non-genetic factors, including concomitant medications, inflammation, organ dysfunction, aging, smoking, pregnancy, and other environmental influences. Phenoconversion can generate substantial discrepancies between genotype-predicted and observed metabolic phenotypes, contributing to adverse drug reactions, therapeutic failure, and reduced clinical utility of pharmacogenomic testing. The identification of phenoconversion remains challenging because the relevant determinants are often dispersed across multiple sources of information, including medication records, laboratory data, diagnostic information, and unstructured clinical notes within electronic health records. Recent advances in large language models (LLMs) offer new opportunities to address this limitation by retrieving, organizing, and contextualizing heterogeneous clinical and pharmacogenomic information. Beyond identifying drug-drug-gene and drug-gene-environment interactions, LLM-enabled systems may support dynamic phenotype assessment by modeling metabolic capacity as a continuously evolving characteristic rather than a static genotype-derived category and by identifying relevant modifiers even for drugs without actionable pharmacogenomic recommendations. This review examines the biological basis and clinical significance of phenoconversion and discusses the potential role of LLMs in bridging the gap between genetic predisposition and real-world drug response. We highlight the emerging concept of dynamic phenotype models that combine genetic, clinical, and environmental determinants of drug metabolism and may support more precise, context-aware pharmacotherapy.
SN - 1664-8021
ER -