We work in the broad research areas of high dimensional data,
measurement error and network models for metabolomics datasets.
Our work is motivated by ongoing collaborations in metabolomic
studies nested within the Women’s Health Initiative and
the Nurses’ Health Study cohorts, with a focus on cardiovascular disease, stroke
and mental health.
With application to data generated in metabolomic studies,
our recent work has centered on variable selection/prediction
in matched case-control studies, handling non-random censoring
in metabolomics and in network models.
Our work has involved developing statistical models and software
for the design and analysis of studies that collect self-reported
outcomes. Self-reports are attractive instruments for collecting
prevalent and incident disease outcomes in large-scale, epidemiologic
Early Detection of HIV Infection in Infants
We are currently involved in an international, multi-cohort collaboration
to evaluate the properties of diagnostic tests used in early infant
diagnosis of HIV infection, particularly in the context of maternal
regimens involving highly active antiretroviral therapy (HAART).