MBE Advance Access published online on February 26, 2009
Molecular Biology and Evolution, doi:10.1093/molbev/msp036
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Article |
Gene sampling can bias multi-gene phylogenetic inferences: the relationship between red algae and green plants as a case study
1 Center for Computational Sciences
2 Institute for Biological Sciences
3 Graduate School of Systems and Information Engineering, University of Tsukuba, Tsukuba, Ibaraki 305-8577, Japan
* To whom corresponding addressed to: YI, +81 29 853 6483, yuji{at}ccs.tsukuba.ac.jp.
Received for publication September 15, 2008. Revision received January 30, 2009. Accepted for publication February 17, 2009.
The monophyly of Plantae including glaucophytes, red algae, and green plants (green algae plus land plants) has been recovered in recent phylogenetic analyses of large multi-gene data sets (e.g., those including >30,000 amino acid (aa) positions). On the other hand, Plantae monophyly has not been stably reconstructed in inferences from multi-gene data sets with fewer than 10,000 aa positions. An analysis of 5,216 aa positions in Nozaki et al. (2007) strongly rejected the monophyly of Plantae, while Hackett et al. (2007) robustly recovered the Plantae clade in an analysis of 6,735 aa positions. We suspected that the significant incongruity observed between the two studies was attributable to a bias generally overlooked in multi-gene phylogenetic estimation, rather than data size, taxon sampling or methods for tree reconstruction. Although glaucophytes were excluded from our analyses due to a shortage of sequence data, we found that the recovery of a sister-group relationship between red algae and green plants primarily depends on gene sampling in phylogenetic inferences from <10,000 aa positions. Phylogenetic analyses of data sets with fewer than 10,000 aa positions, which can be prepared without large-scale sequencing (e.g., expressed sequence tag analyses), are practical in challenging various unresolved issues in eukaryotic evolution. However, our results indicate that severe biases can arise from gene sampling in multi-gene inferences from <10,000 aa positions. We also address the validity of fast-evolving gene exclusion in multi-gene phylogenetic analyses, in light of this gene sampling bias.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  P. Deschamps and D. Moreira Signal Conflicts in the Phylogeny of the Primary Photosynthetic Eukaryotes Mol. Biol. Evol., December 1, 2009; 26(12): 2745 - 2753. [Abstract] [Full Text] [PDF]
|
|