Kimmen Sjölander, Department of Bioengineering, UCB
Protein function and structure prediction
Because proteins are the workhorses of the cell, defining much of
who we are, our disease susceptibility, and more, scientists in both academia and
industry are devoting enormous resources and time to understanding protein function
and structure. Partly due to the cost and difficulty of experimental ("wet-bench")
methods, computational methods of predicting protein function and structure
are being increasingly relied upon.
In this talk, I will present some new methods for the prediction of protein
function and structure, using a combination of information-theoretic and
probabilistic tools. These statistical approaches have been shown to be very
powerful for the recognition of related family members, prediction of key
functional positions in proteins, construction of evolutionary trees, and
identification of functional subfamilies.