The Biostatistics & Data Science Methods Research Program focuses on the development of innovative study designs and analysis methods for application to environmental exposures of importance in the Health Outcomes Research Programs. Methodological challenges include how to integrate high dimensional, often correlated, exposure mixtures and other data from new measurement platforms, longitudinal data identifying key windows of susceptibility to environmental exposures over the lifecourse, multiple phenotypes, and physiological and biomarker measurements. This research program is focusing on developing systems approaches, data integration including network analyses for omics data with environmental perturbations, and pathways approaches for breathomics, metabolomics, microbiome, genomics and transcriptomic data.

Recent Highlights

• A new method called LUCID (Latent Unknown Clustering Integrating Multi-omics Data) integrates observed exposures/genotypes, high dimensional omics data, and observed health outcomes using joint modeling via a latent cluster. LUCID has been utilized by Center investigators in analyses of PFAS and mercury effects on metabolic pathways.

• As part of the NIH Pediatric Research using Integrated Sensor Monitoring Systems (PRISMS) program, a statistical analysis pipeline was developed for mHealth data and novel statistical methods.

• Hierarchical Bayesian models were developed to estimate effects of air pollution exposures on airway and alveolar inflammation.