Russell Schwartz,
Departments of Biological and Computer Sciences, Carnegie Mellon University

While the sequencing of a consensus human genome has provided us with a valuable tool for understanding humanity's common genetic heritage, the relatively rare variations in the genome from one person to another are a key source of information on the causes of and potential treatments for many human diseases. By locating variations found with unusually high or low frequency in people with a given disease, we can predict who is at particular risk for that disease, learn about its molecular basis, and possibly suggest avenues for treatment. The masses of data becoming available on human genetic variations thus present a great opportunity for biology and medicine. The interpretation and exploitation of this information also, however, poses significant computational challenges.

To take full advantage of data on genetic variations, we will need a deeper understanding of their underlying regularities and how to exploit them. I will discuss how the computer science community can address these data analysis problems. Doing so will require computational methods for understanding how our genomes differ from one another, what common structures the patterns of genetic variation possess, and how this knowledge can be put to practical use. I plan to discuss a variety of problems involved in interpreting data on genetic variations and applying it to the search for variations related to human disease. For each of these problems, I will describe its biological significance, discuss how we can formulate it as a solvable computational problem, and present methods that have proven effective in my own work or in the field as a whole. Finally, I will consider possible future directions for computational research in understanding human genetic variability.

Return to homepage.

The second and fourth images in the header are courtesy the BIODIDAC website.