Philip S. Boonstra
I am a PhD Candidate in the Department of Biostatistics at the
University of Michigan, Ann Arbor, under the supervision of
Jeremy MG Taylor and
Bhramar Mukherjee.
I received my BA from Calvin College, studying Mathematics and Political Science.
My dissertation looks at the problem of efficiently predicting an outcome when covariates
are both high-dimensional and primarily measured with error. In my work, I focus on shrinkage estimators,
in which inefficient estimators are regulated by shrinking towards a target. I also consider how to adapatively shrink
multiple estimators to achieve a super-efficient estimator.
I study genetic anticipation, or a decrease in age-of-onset of disease between children and parents
due to genetic factors. That this phenomenon exists in Lynch syndrome, a familial colon cancer, has been posited,
but thus far no mechanistic cause has been found. Apart from genetic causes, decreases in age-of-onset can happen
from improvements in medical diagnostic techniques or ascertainment bias. The statistical challenge is then to
separate these spurious effects from true genetic anticipation, if it exists.
My other academic interests include
- Statistical applications in cancer research
- Data visualization
- Graphical models
- Nonparametric Bayesian methods
See my CV or contact me for further information:
Email: philbXumichYedu, where X=@ and Y=.