I am gonna do ssurRNA sequence alignment of Dicyemid. I want to know the relationship of different dicyemid found in different country. Which phylogenetic model should i use? or if you have some better and easier method to describe the ancestor relationship, i would appreciate that.
A small subunit ribosomal RNA (ssu-rRNA) alignment may not provide solid information on the phylogeny of Dicyemids. The Dicyemids are a diverse group of mesozoans which could have shared a common ancestor more than 500 million years ago, so the ribosomal RNAs might be quite diverse. However, the subset of Dicyemids that you study might be less diverse, so you should definitely look at your data and decide. It will almost certainly be helpful to use an alignment method that takes into account the stem-loop secondary and 3D structure of ribosomal RNA such as the ssu-Aligner at Janelia. Choosing a good outgroup, and interpretting your resulting tree, will be as or more important that which model of evolution you use. Start with GTR plus gamma, and/or run ModelTest. FindModel is a web service based on ModelTest.
http://www.bioone.org/doi/pdf/10.2108/zsj.20.519
http://selab.janelia.org/software/ssu-align/
http://www.molecularevolution.org/software/phylogenetics/modeltest
http://www.hiv.lanl.gov/content/sequence/findmodel/findmodel.html
2003 Zoological Society of Japan
ZOOLOGICAL SCIENCE 20: 519–532 (2003) [REVIEW]
Biology of Dicyemid Mesozoans
Hidetaka Furuya and Kazuhiko Tsuneki
Department of Biology, Graduate School of Science, Osaka University,
Toyonaka, Osaka 560-0043, Japan
And for a little more perspective of multicellular life overall, look at this new paper.
Chapter Timing the Origins of Multicellular Eukaryotes Through Phylo...
Regardless of how important it is, more models is included in jmodeltest2, which also run on a practical java GUI. The results are also printed to a nice html file and models can be chosen on basis on either BIC, AIC or AICc (or best case scenario, that they all agree). As mentioned, GTR is quite a safe bet, but running a model test is quite quick and easy as well, and probably wont hurt you.
http://code.google.com/p/jmodeltest2/
Thank you very much doctor Brian and
Øystein. It is a valuable insight for my research. I hope i can ask for your knowledge again in later date. Thanks again!
- Badges
- Science topic
- Similar topics
- Data Mining
- Classification