Spring Seminar Series - April 6, 2006
University of Minnesota
School of Statistics
College of Liberal Arts
Statistical
Matching and Alignment in Protein Bioinformatics
Kanti V. Mardia
Senior Research Professor
University of Leeds, UK
Thursday, April 6, 2006
3:30 PM, 115
Ford Hall
Minneapolis, East Bank Campus
Social at 3:00 PM, 300 Ford Hall
Abstract
Among
the challenges for statistics posed by proteomics are various alignment
and matching problems. Here we consider matching protein gels in 2
dimensions,and aligning active sites of proteins
in 3 dimensions. In the latter case, we also want to use information
related to the grouping of the amino acids. We introduce
hierarchical Bayesian models for matching configurations of points in
space,
where the points are either unlabelled, or have at most a partial
labelling constraining the matching, and in which some points may only
appear in one of the configurations. We derive procedures for
simultaneous inference about the matching and the transformation, using
a
Bayesian approach (Green and Mardia, 2006). Our procedure is compared
to
other methods such as a graph theoretic method (Gold, 2003) and EM
algorithm (Taylor, Mardia and Kent, 2003; Kent ,Mardia and
Taylor,2004).
Implementation and performance of these methods on the proteomic
tasks
is described, and we discuss some open problems and suggest
directions for future work.