Current Bio3 Seminar Series

Abstract: Statistical models that predict disease incidence, disease recurrence, or mortality following disease onset have broad public health and clinical applications. Of great importance are models that predict absolute risk (also called ‘cumulative incidence’ or ‘crude risk’), the probability that an individual who is free of a given disease at an initial age, a, will develop that disease in the subsequent interval (a, t]. Absolute risk is reduced by mortality from competing risks. I will discuss approaches to building risk models and then address validating a risk model in an external cohort. Model validation is an important step before a risk prediction model can be recommended for applications. The specific problem I address is model validation when not all predictor variables in the model are available on all members in the validation cohort. Missingness can be random or by design (e.g. predictor information is available only in case-cohort and nested case-control samples).  I will present new methods to deal with such missingness when estimating performance characteristics of absolute risk models.    

Location: Building D, Warwick Evans Conference Room

Abstract: Generative AI has rapidly transformed the biomedical imaging field by enabling image synthesis, helping address challenges of limited data availability, privacy, and diversity in biomedical research. Yet, the adoption of AI-generated images in biomedical studies requires rigorous methods to ensure their reliability for downstream analysis. In this talk, I will introduce novel and rigorous nonparametric approaches that strengthen the trustworthiness and statistical validity of synthetic biomedical imaging data. We develop simultaneous confidence regions to rigorously quantify uncertainty and detect meaningful differences between synthetic and original imaging data. To further enhance fidelity and utility, we propose a transformation that aligns the mean and covariance structures of synthetic images with those of the originals. I will also discuss methods for imputing missing imaging phenotypes using generative models and demonstrate how joint analysis of observed and imputed traits enhances inference while accounting for imputation error. Extensive simulations and applications to brain imaging data validate the proposed framework, demonstrating how these methods empower rigorous statistical inference and promote trustworthy advances in biomedical imaging.

Location: Med-Dent Harvey Amphitheater North

Abstract: Are you interested in how biostatistics connects with the pharmaceutical industry, regulatory science, and health innovation? The ASA Biopharmaceutical Section (BIOP) offers a wide range of opportunities for students and faculty to grow professionally, connect with leaders in the field, and access resources to support their research. In this session, BIOP representatives will give an overview of awards, conferences, and our external funding program. We will then open the floor for an interactive Q&A to discuss career paths, networking, and ways to get involved with BIOP.

Location: Building D, Warwick Evans Conference Room

Abstract: Biomedical studies now generate vast amounts of molecular data, ranging from thousands of biomarkers to millions of genetic variants, creating major challenges in distinguishing meaningful signals from noise. Patient outcomes vary widely, and methods that focus only on averages can overlook biomarkers linked specifically to poor or favorable outcomes. One direction of this work applies quantile regression to study different parts of the outcome distribution, allowing the identification of markers relevant at the extremes. Another challenge arises in the high-dimensional setting, where some biomarkers may not appear important individually but become informative when considered jointly. To address this, I will introduce a conditional screening approach developed to uncover these hidden signals.

The presentation will also include an overview of the NIH Intramural Research Program, which supports postbaccalaureate, predoctoral, and postdoctoral fellows in biostatistics and related fields, working closely with investigators across diverse areas of biomedical research.

Location: Building D, Warwick Evans Conference Room

Abstract: Transformative AI models are powerful tools for integrating and mining large-scale biomedical and omics data, and have been revolutionizing research in precision medicine.  In this talk, I will present novel transformative AI models that we have developed to combine large language models (LLMs) with graph-based AI to integrate and analyze vast omics datasets for identifying disease targets, inferring signaling pathways, and predicting effective drugs and drug combinations. A key component of the novel AI models is the novel text-numeric graph (TNG) or text-omic signaling graphs (TOSGs), a novel structure in which graph entities and associations carry both textual and numeric attributes. I will also introduce an AI multi-agent system that we have developed to accelerate biomedical discovery by unifying omics data analysis, literature-based deep search, and reasoning to generate novel scientific hypotheses. I will then showcase the applications of these novel AI models with analysis of heterogeneous pharmacogenomics data for precision medicine.

Location: Building D, Warwick Evans Conference Room