Thursday March 2, 2023
Biostatistics PhD candidate
University of North Carolina Chapel Hill.
Hillary’s statistical focus has primarily been in cancer-related research, both through her graduate research assistant role in the Lineberger Comprehensive Cancer Center as well as her personal research applications. In cancer research as well as other biomedical areas, there are issues with the replicability of results between studies. Hillary will present her work on improving the replicable selection of gene signatures for the prediction of cancer subtypes by combining data across studies and performing variable selection on generalized linear mixed models. Hillary has been able to extend the feasible dimensionality of the application to hundreds of predictors by using a factor model decomposition on the random effects, which behaves as a dimension reduction technique. Hillary has developed software to perform this task and has published her ‘glmmPen’ R package on CRAN.