Journal
MOLECULAR PHYLOGENETICS AND EVOLUTION
Volume 32, Issue 1, Pages 110-122Publisher
ACADEMIC PRESS INC ELSEVIER SCIENCE
DOI: 10.1016/j.ympev.2003.11.017
Keywords
allodapini; bee evolution; phylogenetics; long branch attraction; rate heterogeneity; AT bias
Ask authors/readers for more resources
Extreme AT bias in Hymenopteran mitochondrial genes have created difficulties for molecular phylogenetic analyses, especially for older divergences where multiple substitutions can erode signal. Heterogeneity in the evolutionary rates of different codon positions and different genes also appears to have been a major problem in resolving ancient divergences in allodapine bees. Here we examine the phylogeny of relatively recent divergences in the allodapine bee genus Braunsapis. We examined heterogeneity in nucleotide substitution parameters for one nuclear gene and codon positions in two mitochondrial genes, exploring various phylogenetic analyses for recovering relationships among species from Africa, Madagascar, southern Asia, and Australia. We explored maximum parsimony, maximum likelihood, Log determinant and Bayesian analyses. Broad topological features of best fit trees tended to be similar for equivalent data sets (e.g., total, or with 3rd mt positions excluded), regardless of the analytic method used (e.g., maximum likelihood or Bayesian). Analyses that used the total data set without modelling partitions separately gave unlikely results, indicating that the Malagasy species was most closely related to Australian species. However, analyses that excluded 3rd mitochondrial positions, or modelled partitions separately, suggested that the Malagasy species falls within the African clade. The unlikely topologies apparently result from long branch attraction, and this problem is ameliorated where modelling allows more realistic estimates of base composition and evolutionary rates for 3rd mitochondrial positions. However, we found that even when codon positions are modelled separately, estimated evolutionary rates for 3rd mitochondrial positions are likely to underestimate true rates. Long branch attraction and multiple substitutions are likely to be much more difficult to circumvent in analyses that explore older, generic-level, divergences in allodapine bees where overwriting is expected to be much more extreme. Our results indicate an African origin for Braunsapis, followed by a single, very early, dispersal event into Asia and then by a later dispersal event into Australia. The Malagasy species is derived from within the African clade. (C) 2003 Elsevier Inc. All rights reserved.
Authors
I am an author on this paper
Click your name to claim this paper and add it to your profile.
Reviews
Recommended
No Data Available