Advancing statistical genetics tools for reliable drug target discovery and treatment optimization

The discovery of novel drug targets and the optimization of treatment regimens remain critical challenges in addressing diseases with unmet medical needs. Despite substantial progress in genetic and multi-omics research, existing methodologies often lack generalizability across diverse populations and fail to capture the complex, potentially non-linear relationships underlying drug responses. Specifically, most human genetics- facilitated drug target discovery efforts using Mendelian randomization (MR) rely on genetic resources from populations of predominantly European ancestry, potentially exacerbating health and healthcare disparities that affect underrepresented populations. Additionally, the dose-dependent and combinatorial effects of drug targets remain poorly understood, hampering the development of tailored treatments and precision medicine approaches. Several methodological challenges persist in achieving reliable drug target discovery and treatment optimization. For instance, confounding due to linkage disequilibrium (LD) undermines the validity of MR analyses, particularly in diverse populations with varying genetic architectures. Moreover, the absence of robust causal inference frameworks for populations with pre-existing health conditions complicates drug target discovery. In addition, biases in existing non-linear MR approaches hinder accurate evaluation of dose-response relationships, especially for therapeutic targets with saturation kinetics or other non-linear effects. To address these challenges, this program will develop and implement advanced statistical genetics tools over the next five years. From a theoretical perspective, we will focus on improving the robustness of MR to LD-induced confounding, ensuring more reliable causal inferences across diverse populations. We will design novel MR methodologies incorporating structural equation models for drug target discovery tailored to individuals with pre- existing health conditions, providing actionable insights for patient-specific therapies. Additionally, we will refine non-linear MR approaches to address known biases, thereby enabling more accurate and clinically relevant dose-response analyses. With these innovations, this program will also deliver transformative applications. We will conduct systematic drug target discovery by incorporating multi-omics and genetic resources from populations of diverse genetic ancestries, as well as populations with different combinations of pre-existing health conditions. We will elucidate the dose-dependent effects of the identified drug targets to optimize therapeutic dosing regimens for clinical application. Furthermore, we will extend the factorial MR framework to explore the effects of combined therapies, identifying potential synergistic interactions and supporting the development of precision treatment strategies. By integrating applied and theoretical advances, this program seeks to establish a robust, generalizable framework for identifying and optimizing drug targets. Ultimately, our goal is to bridge the gap between genetic discoveries and their clinical translation, advancing personalized medicine to improve health outcomes for diverse populations in the United States and worldwide. Project Number: 1R35GM162188-01 | Fiscal Year: 2026 | NIH Institute/Center: National Institute of General Medical Sciences (NIGMS) | Principal Investigator: Tianyuan Lu | Institution: UNIVERSITY OF WISCONSIN-MADISON, MADISON, WI | Award Amount: $414,305 | Activity Code: R35 | Study Section: Maximizing Investigators' Research Award - F Study Section[MRAF] View on NIH RePORTER: https://reporter.nih.gov/project-details/11268381