Project 1 (Genomics, Abecasis): The students will be given access to a data set representing whole genome sequencing results for a set of 1,000 - 2,000 individuals. Students will explore the issues in large genetic studies, which range from being able to visualize the main patterns of variation in the data which allow researchers to distinguish data generated with different protocols and to estimate the ancestry of study participants. Students will search for more subtle factors describing connections between individual genetic variants and disease risk and work on deriving meaningful summaries for their presentation.


Project 2 (Imaging, Johnson): Students will be given an fMRI data set to analyze.  They will learn how the data are stored in the nifti (Neuroimaging Informatics Technology Initiative) file format, how to preprocess the data for statistical analysis, how to analyze single subject data, and how to analyze group level data.  Learning objectives will include: how data and header information are stored, how to set up the hemodynamic response function and its implications, how to filter the data, how to correct for magnet non-homogeneities, how to address effects of different experimental design, and how to correct for massive numbers of hypothesis tests.  


Project 3 (Health Records, Li): Students will be exposed to the analysis of large-scale registry database established for the evaluation of 6,000 dialysis facilities nationwide. Students will be given the opportunity to understand the Dialysis Facility Reports (DFR) created annually for each of the approximately 6,000 Medicare certified dialysis facilities in the country.  They will learn how to extract, based on the DFR, the facility-specific summaries of patient characteristics, treatment patterns, and patient outcomes for chronic dialysis patients treated during 2008-2011. They will learn how to visualize and identify trends in practices and outcomes over time, develop regional comparison statistics. They will study variability in clinical outcomes at the facility-level across the country to associate patient outcomes with practice patterns.