Center for Bioinformatics at KU

As part of our public Seminar Series
http://bioinformatics.ku.edu/seminars

Feb 22, Thu 2007
11:45am - 12:15pm, MRB Conference Room

Craig Helstowski

The University of Connecticut

Profiling Phylogenetic Informativeness in Angiosperms Amborella and Nymphaeales

A problem that is constantly faced today by evolutionary biologists is constructing evolutionary trees showing the divergence of one species from another.  Although there are many ways to construct trees, the most reliable way to construct one is from detecting changes in the genetic material and just how much different the DNA from two species are.  Current methods used by phylogeneticists to construct trees based on similarity between DNA sequences, such as Bootstrap values, or posterior probabilities only convey how much support should be given to the tree that was previously constructed.  Even though different character states among genes can allow phylogeneticists to construct clades based on these states, we still do not know exactly how and when two species diverged from each other and what the main cause (or more specifically which change in a particular gene) for the divergence was originally. 

What this research project entails is an inference of the evolutionary history of angiosperms, specifically the taxon Amborella and the group Nymphaeales.  A new analysis will be performed on a published dataset comprising genes: one will be from 16 taxa using 11 genes (18S rDNA, 26S rDNA, phyA, phyC, mtSSU, cox1, rps2, atpA, matR, rbcL, and atpB) and the other set will use a subset of genes just mentioned, however it will be done for 104 taxa.  By using these two different sets we can more closely scrutinize the evolution of Amborella and Nymphaeales and then compare it on a more broad scale by comparing their evolution to further taxa of angiosperms using set 2.   The sequences within each set will be aligned and then using a program called DNARates (by Gary Olsen), the rate of evolution amongst each gene can then be determined. Using these rates we can then determine exactly the power of the genes to infer the branching order and the evolution of the angiosperms being studied in comparison to others. 

What the profile of phylogenetic informativeness does is to determine the power for determining the branching order when a divergence occurred based on the rate of evolution of sites in each gene in a set.  The higher the rates of evolution of sites within a gene, the more likely a recent divergence may be correctly inferred, and the lower the rate, the more likely the divergence may be correctly inferred at a more ancient time.  The profile will take these rates and establish a relative informativeness based on the rates of evolution of sites within each gene using a summation of the poisson distributions of the rate of change of each character, or gene, based on some time T of a common ancestor.  Each gene will have its own distribution.  Then based on these profiles we can determine exactly in what order Amborella and Nymphaeales evolved, whether it be from each other or from other ancestors.  This way is a more quantitative alternative to other heuristic methods used to determine genes to use to create trees.  This method will give a more definitive answer as to the true history of flowering plant evolution based on genetic data sequences.