May 26 2010

12:00 LSB 2320

Michael J. Sanderson
Dept. of Ecology and Evolutionary Biology, University of Arizona

Building the tree of life one (small) tree at a time


A comprehensive phylogenetic tree of life will be built from smaller trees and data sets that partially overlap in the taxa they include. In this talk I describe some recent theoretical results that indicate the constraints imposed on the quality of this tree by the pattern of taxonomic overlap (or missing data). Results are described for both the case in which the pattern of overlap is already given (as in an already assembled collection of inputs) and in which sampling will continue according to some random sampling protocol (as in phylogenomics studies with partial genomic coverage).
In addition, an algorithm for optimal selection of a subset of data is described for a collection of trees and taxa that is negatively impacted by missing data.











































































































































































































































































































































































































































































































































































































this is idtest: