DOES IMBALANCE IN PHYLOGENIES REFLECT ONLY BIAS?
Article first published online: 9 MAY 2007
Volume 56, Issue 6, pages 1292–1295, June 2002
How to Cite
Stam, E. (2002), DOES IMBALANCE IN PHYLOGENIES REFLECT ONLY BIAS?. Evolution, 56: 1292–1295. doi: 10.1111/j.0014-3820.2002.tb01440.x
- Issue published online: 9 MAY 2007
- Article first published online: 9 MAY 2007
- Received June 1, 2001. Accepted February 18, 2002.
- Colless's index;
- consistency index;
- misinformative characters;
- phylogenetic tree balance;
- retention index;
- tree estimation bias
Abstract Phylogenetic tree imbalance was originally believed to indicate differences in evolutionary rates within trees, but other sources of imbalance have been identified, such as tree incompleteness and low quality of the data. To examine the effect of data quality, I calculated Colless's index for 69 recent complete phylogenies. On average, these phylogenies were more unbalanced than phylogenies generated by the equal rates Markov (ERM) model. I tried Mooers's (1995) method to correct for tree size, but his measure appeared to become dependent on tree size when there are large trees (i.e., 14 tips) in a collection. Instead I corrected for tree size by taking the difference between Colless's index of observed trees and the ERM model expectation for a tree of the same size. The balance measure thus obtained did not correlate significantly to consistency and retention indices as indicators of data quality. It was also independent of the factors kingdom (plants and animals) and taxon level at the tips and type of data (molecular, morphological, and combined).