Nazia Riasat

I investigate whether introgression leaves detectable geometric signatures in genealogical tree sequences. This work combines coalescent simulation, tree distances, representative-tree methods, and statistical testing to characterize when evolutionary signals can be distinguished from background variation.

I study how large language models behave when scientific evidence is incomplete, confounded, or semantically misleading. My work evaluates stability, causal reasoning, ranking behavior, and failure modes of reasoning-oriented models in controlled biomedical decision-making settings.

I develop and evaluate statistical simulation methods for transcriptomic data, with emphasis on preserving marginal distributions, dependence structure, co-expression networks, stability, and differential-expression behavior. My dissertation includes a task-specific benchmark of widely used bulk RNA-seq simulators and the development of a new dependence-aware generative framework for biologically structured synthetic data.

I use topological and functional data methods to study resting-state fMRI connectivity in traumatic brain injury. Current work focuses on Euler characteristic curves, functional principal components, permutation methods, and other approaches for comparing network topology across clinical groups.

I conducted statistical analyses integrating evidence-based tobacco and nicotine dependence treatment into advanced nursing curricula. This project validated significant improvements in nurse practitioner students’ knowledge, confidence, and counseling ability.

During the COVID-19 pandemic, I led statistical analyses of the survey of healthcare students in addressing vaccine hesitancy. Results showed statistically significant gains in knowledge, confidence, and a measurable reduction in vaccine-hesitant attitudes. Importantly, 92% of participants reported intent to apply these strategies with patients.

I analyzed over 230,000 pediatric health visits in an NIH-funded study of obesity prevalence before, during, and after the pandemic. Results revealed significant, persistent increases in obesity, particularly among adolescents, rural populations, and American Indian/Alaska Native children even after return to pre-pandemic activities. My analyses and visualizations highlighted the role of social determinants of health in exacerbating disparities, providing critical evidence for designing targeted federal and state interventions.

I led the statistical analysis and writing on the safety outcomes of 15-year-old novice drivers under Graduated Driver Licensing policy. Using logistic regression on over 15,000 records, I demonstrated that public school driver education programs reduced crash risk by 32%. This evidence supports policy reforms aligned with federal “Vision Zero” goals to eliminate roadway fatalities.