General Exam - Taek Son
Committee: Kwun Chuen Gary Chan (co-chair), Eardi Lila, (co-chair), Ting Ye, Abraham Flaxman, (GSR)
Presentation: Dimension Reduction Methods for Improved Estimation of Causal Parameters
Abstract: Drawing causal conclusions from observational data such as electronic health records (EHR) is of great scientific and clinical interest, as these data reflect real-world patient populations. They also enable researchers to investigate causal questions that are often infeasible or unethical to address through randomized controlled trials.
Examples of causal problems include the estimation of conditional average treatment effects (CATE), that is, quantifying the effect of a treatment conditional on specific values of individual-level characteristics. Closely related is the problem of estimating an individualized treatment regime (ITR), which defines a mapping from patient characteristics to a treatment decision rule with the goal of maximizing the expected clinical outcome. Another important class of problems is…
Event interval: Single day event. Campus location: Hans Rosling Center for Population Health (HRC). Online Meeting Link: https://washington.zoom.us/j/7580402955. Campus room: HRC 342. Accessibility Contact: Deb Nelson, nelsod6@uw.edu, 206-685-9323. Event Types: Academics.
Wednesday, March 11, 2026, 2:00 PM – 4:00 PM.
Final Exam - Yinxiang Wu
Committee: Ting Ye (co-chair), Andrea Rotnitzky (co-chair), Li Hsu, Marco Carone, Sara Lindstroem (GSR)
Presentation: Causal Inference in the Presence of Unmeasured Confounding: Advances in Mendelian Randomization and Proximal Causal Inference
Abstract: Observational data are indispensable for causal inference, particularly when randomized controlled trials are infeasible due to ethical, logistical, or economic constraints. However, observational data are subject to unmeasured confounding, where unobserved variables influence both the treatment and the outcome, potentially leading to biased estimates and spurious findings.
This dissertation aims to develop statistical methods for causal inference in the presence of unmeasured confounding, focusing on multivariable Mendelian randomization (MVMR) using summary-level data and proximal causal inference using individual-level data. MVMR uses genetic variants as instrumental variables (IVs) to infer the direct causal effects of multiple exposures on an outcome.…
Event interval: Single day event. Online Meeting Link: https://washington.zoom.us/j/6287851634. Campus room: HRC 342. Accessibility Contact: Deb Nelson, nelsod6@uw.edu, 206-685-9323. Event Types: Academics.
Thursday, March 12, 2026, 10:00 AM – 12:00 PM.
Biostatistics Seminar: Advancing Cancer Risk Prediction: Multi-Ancestry PRS, Model Recalibration, and Benefit-Burden Evaluation
Speaker: Li Hsu, PhD, Professor, Biostatistics Program, Fred Hutch Cancer Center; Affiliate Professor of Biostatistics, University of Washington
Presentation: Advancing Cancer Risk Prediction: Multi-Ancestry PRS, Model Recalibration, and Benefit-Burden Evaluation
Abstract: Precision cancer prevention depends on accurate and transportable risk prediction tools with equitable performance across populations, as well as rigorous of prevention strategies in real-world cohorts. In this talk, I will present several lines of recent methodological and applied work aimed at advancing these goals in colorectal cancer. I will discuss improvements in the development of polygenic risk scores (PRS) across population groups. While PRS show promise for identifying individuals at elevated risk who may benefit from targeted screening, most existing scores are derived primarily from European ancestry data and have sub-optimal performance in underrepresented populations. To address this limitation, we developed a functionally…
Event interval: Single day event. Campus location: Architecture Hall (ARC). Campus room: ARC G070. Accessibility Contact: Deb Nelson, nelsod6@uw.edu, 206-685-9323. Event Types: Academics.
Thursday, March 12, 2026, 3:30 PM – 4:30 PM.
General Exam - Jaewon Lim
Committee: Alex Luedtke (chair), Marco Carone, Ting Ye, Abraham Flaxman (GSR)
Presentation: Estimation and inference for function-valued parameters with data fusion and multimodal data
Abstract: There is a growing literature on estimating function-valued parameters by combining tools from semiparametric efficiency theory and statistical learning. Another trend is the increasing use of multiple data sources that share only partially overlapping information, estimation by integrating such sources through data fusion and multimodal fusion. In this dissertation, I develop theory and methodology at the intersection of these two directions.
In the first aim, I study data fusion for estimating causal dose-response functions (CDRFs). Estimating CDRFs is challenging because a single dataset typically provides limited support across the exposure space, leading to poor precision at many dose levels. I introduce a data fusion framework that leverages partially aligned sources, and I construct a Neyman-orthogonal loss…
Event interval: Single day event. Online Meeting Link: https://washington.zoom.us/j/93986181191. Campus room: Online. Accessibility Contact: Deb Nelson, nelsod6@uw.edu, 206-685-9323. Event Types: Academics.
Wednesday, March 18, 2026, 10:00 AM – 12:00 PM.
Biostatistics Seminar: Five lessons for target gene identification in the post-GWAS era
Speaker: Boxiang Liu, PhD, Presidential Young Professor,Department of Pharmacy and Pharmaceutical Sciences & Department of Biomedical Informatics, National University of Singapore; Principal Research Scientist, Genome Institute of Singapore, A*STAR
Title: Five lessons for target gene identification in the post-GWAS era
Abstract: Despite over a million risk variants identified by genome-wide association studies (GWAS), translating these associations into therapeutic insights remains a bottleneck in genomic medicine. In this talk, I present five critical lessons from post-GWAS research that collectively advance our ability to pinpoint disease-relevant target genes. Using large-scale single-cell RNA sequencing across diverse Asian populations and gene-by-intervention studies, we highlight the significance of cell-type-specific QTLs, ancestry-driven genetic insights, and context-dependent gene regulation. Our systematic benchmarking of over gene implication methods shows that ensemble approaches and adaptive…
Event interval: Single day event. Campus room: Virtual Seminar. Accessibility Contact: Deb Nelson, nelsod6@uw.edu, 206-685-9323. Event Types: Academics.
Thursday, March 19, 2026, 3:30 PM – 4:30 PM.