Ha Nguyen: Advocating for Better-Fitting Women’s Underwear in the Military
In collaboration with Dr. Fatma Baytar from the Fashion & Body Tech Lab, we study anatomical variability in the female crotch curve to inform better-fitting underwear designs for women in the military. Right now, women’s underwear in the military is mass-produced with only a few size and shape options, which often do not fit well. This can cause discomfort, especially because service members have to wear them for long hours. We systematically characterize crotch curvature variability across individuals, aiming to provide quantitative evidence for the need for more tailored and inclusive designs.
Sanghee Kim: Evaluating Uncertainty of Financial Markets via Microstructure Measures
Jointly with Dr. David Easley, we aim to evaluate the market uncertainty by analyzing large-scale high-frequency trade data. In recent years, trades are made in nano-seconds via human or an algorithm. As this vast data tells important aspects of price formation and behaviors of market participants, we construct microstructure variables and apply data-driven machine learning/time-series models. By measuring market illiquidity, volatility, and information-based trades, we would like to understand the future market behavior.
Hao Xue: Integrative Analysis of Differentially Expressed Genes in Time-Course Multi-Omics Data with MINT-DE
We propose a framework called MINT-DE (Multi-omics INtegration of Time-course for Differential Expression analysis) that is capable of selecting genes according to the magnitude of change and the temporal dynamic consistency across two modalities based on summarized. We compare the selection obtained from MINT-DE with those obtained from other existing methods. The analysis reveals that MINT-DE is able to identify differentially expressed time-course pairs with the highest correlations. Their corresponding genes are significantly enriched for known biological functions, as measured by gene-gene interaction networks and the Gene Ontology enrichment.
Automated annotation of cell types in embryonic heart development via spatial transcriptomics
In collaboration with Iwijn Vlaminck’s lab, we study spatial transcriptomic data from fetal chicken hearts using machine learning to construct a spatio-temporal map of heart development. Our focus is on jointly identifying and annotating cell types that are shared across, or specific to, different stages of fetal heart growth. Chicken embryos provide an excellent model for cardiogenesis, as their heart anatomy closely mirrors many aspects of human cardiac structure. While single-cell transcriptomic data have been instrumental in studying the cellular mechanisms underlying heart development, their lack of spatial information has limited prior investigations of the complex biomechanical processes driving cardiogenesis. Our ultimate goal is to develop an end-to-end ML pipeline capable of integrating chicken heart transcriptomic data across developmental stages and performing automatic, biologically meaningful cell type annotation, thereby advancing research in cardiovascular biology.
Navonil Deb: Weight prediction of rhinoceroses in Namibia
In collaboration with Robin Radcliffe, Cornell College of Veterinary Medicine, we are developing a statistical tool to predict rhinoceros weights using easily collected body measurements and features (e.g., spine length, shoulder height, half-girth, sex). The project is motivated by challenges faced by our collaborators during wildlife translocation, where weighing scales are often unavailable and the use of tranquilizers is constrained by dosage sensitivity and limited effective duration. Based on easy to interpret generalized linear models, our prediction method aims to provide accurate weight estimates from minimal data, supporting safer and more efficient decision-making in the field.
