Inter-individual variability in drug response represents a major challenge in clinical medicine and public health. The same medication show different response across individuals, leaving some patients with suboptimal benefit or severe side effects. Such variability is influenced by genetic variation, levels of drug-related key proteins, metabolites, hormones, sex, age, and lifestyle factors. Although prior studies have revealed certain signals of variability, they are often limited to a single drug and lack a systematic integration at the population scale.
This project aims to systematically investigate variability and underlying mechanisms of drug response using the large-scale multimodal resources of UK Biobank. The specific aims are to: (1) Quantify real-world variability in drug response across prescribed drug classes using class-specific, clinically interpretable endpoints (e.g., pre-post change in continuous biomarkers and hard outcomes). For example, for antihypertensives, the primary endpoint is (mean SBP change)/(mean DBP change) from a pre-treatment baseline to an on-treatment window; secondary endpoints include cardiovascular events and mortality; (2) Explore the relationships between genetic variants of drug targets and ADME genes and plasma protein levels, thereby elucidating molecular mechanisms regulating drug response; (3) Clarify the influence of sex, age, and metabolic environment on drug response; (4) Construct integrated prediction models based on mechanistic networks of drug response variability.
UK Biobank provides prescription and health records, genetic data, proteomics, metabolomics, and long-term follow-up, creating a unique opportunity to study drug response variability and its mechanisms at scale. By leveraging causal inference , machine learning, and systems modeling, this project will integrate molecular, population, and clinical evidence to identify determinants of drug response variability, advancing precision medicine.