|
Activity Number:
|
313
|
|
Type:
|
Invited
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #307957 |
|
Title:
|
Better Stochastic Mapping and Effective Tests for Evolutionary Innovation
|
|
Author(s):
|
Marc A. Suchard*+ and Vladimir N. Minin
|
|
Companies:
|
University of California, Los Angeles and University of California, Los Angeles
|
|
Address:
|
Human Genetics, Los Angeles, CA, 90095-1788,
|
|
Keywords:
|
Stochastic processes ; Phylogenetics ; MCMC ; Evolution
|
|
Abstract:
|
Stochastic mapping is an invaluable tool to study evolutionary innovation. Previously successful examples include identifying developmental pathway differences between human and chimps and measuring the forces of selection on HIV over the course of infection. Current stochastic mapping tools derive from highly inefficient simulation techniques and are often misleading. Coaching evolutionary innovation as a novel Bayesian Markov Arrival Process, we are able to provide analytic expressions for the mean and variance of the expected number of events in an evolutionary history and recursive solutions to determine their probability mass function. From these expressions, we formulate a test for evolutionary innovation. Unlike simulation-based methods, test quality does not suffer as the number of traits and their complexity increases.
|
