John
HUELSENBECK

 
Evolutionary biologists can learn about the pattern and process of molecular evolution by comparing DNA sequences sampled from different species. However, one known process — the insertion and deletion of DNA in a sequence — has defied rigorous statistical analysis because the computations required to carry out a rigorous analysis are so expensive. The reason: The probability calculations basic to any statistical analysis account for the infinite possible ways one sequence can evolve into another. In this project, Huelsenbeck intends to condition probability calculations on a “known” history of change, thereby making probability calculations trivial. He will use a method called Markov chain Monte Carlo to numerically explore alternative histories of insertion and deletion thereby accounting for uncertainty in the histories of change. In this way, he will be able to test ideas about the process of insertion and deletion. He will extend the methodology to understand the dynamics of human language evolution.