| Abstract Detail
Systematics Section / ASPT Simmons, Mark P. [1], Müller, Kai [2]. The Relative Performance of Indel-Coding Methods in Simulations. WE used simulations to compare the performance of ten approaches that have been applied for treating unambiguously-aligned gaps in phylogenetic analyses. These approaches were implemented in SeqState, which reads and outputs NEXUS-formatted data matrices. We examined how these approaches performed under the ideal conditions of correct alignments, as well as how robust they were to errors caused by the use of inferred alignments. Simulations were performed using different indel rates, lengths, and distributions. Performance was quantified using the averaged overall success of resolution of parsimony jackknife trees relative to the tree on which the characters were simulated. All of the approaches to incorporating phylogenetic signal from gap characters outperformed treating gaps as missing data or excluding gapped positions, using both correct and inferred alignments. Treating gaps as fifth states outperformed the other approaches, though further analyses in which weights were controlled indicated that this was an artifact caused by overweighting multi-position gaps. Simple indel coding and modified complex indel coding generally outperformed the remaining approaches examined.
Log in to add this item to your schedule
1 - Colorado State University, Department of Biology and Program in Molecular Plant Biology, Fort Collins, Colorado, 80523-1878, USA
2 - Rheinische Friedrich-Wilhelms-Universität Bonn, Nees-Institut für Biodiversität der Pflanzen, Meckenheimer Allee 170, Bonn, D-53115, Germany
Keywords:
indel coding
gap characters
indel characters
rare genomic changes
simulation.
Presentation Type: Oral Paper:Papers for Sections
Session: 24-1
Location: Auditorium/Laxson
Date: Monday, July 31st, 2006
Time: 1:45 PM
Abstract ID:86 |
