I have few sequence alignments (ClustalW) in .nex format and I simply need to construct a phylogenetic tree using MCMC. Please note I am not a bioinformatic guy nor familiar with all the parameters and algorithms. I am trying to reproduce a tree published on a recent paper and get some experiences with BEAST. I need instructions to set the GTR+Γ4+I model and then run the analysis. Conditions mentioned on the paper are listed below,
1. Trees were constructed taking into consideration time of isolation and using a GTR + T4 + I model of nucleotide substitution with three codon positions and substitution, rate heterogeneity, and base frequencies unlinked across all codon positions.
2.Four independent MCMC samples of 10 million iterations each were performed for each set of sequences and model.
3.Samples were recorded every 1,000 generations and combined with 10% burn-in removal using the Log Combiner program (implemented in BEAST).
4.Convergence of the chain and sufficient effective sampling size (ESS; ESS > 200) were confirmed using the Tracer program.
5.Maximum clade credibility trees were generated using TreeAnnotator (part of BEAST) and visualized in FigTree
Please assist.
Hie Harindra,
I suggest you pay visit to "beast-users" google group. Open an account and join so that you can share your experiences when running BEAST analysis and also learn from other user's posts. But back to your query. You must be conscious of this fact, BEAST can only implement a number of models, note the word implement here. While J-Model test may specify that the best model is for example X, if this model can not be implemented in BEAST, then there is little merit in testing testing 88 or more models to find a "best" one that you can use with your data set! So to get you started please visit "beast2.org" there you will be able to get "tutorials" on discrete and continuous phylogeography (and others as well). If you really need to understand whats going on as BEAST runs, I strongly suggest you download the BEAST 1.4 manual (note, there are recent versions, 1.8 to 2.1.3 respectively). I found this one more explanatory. Anything else you need you'll most likely get it on beast-users (e.g. how to fix errors that you may encounter, etc). Otherwise you should be able to run your data (i.e. load your .nex or .fasta file, generate an xml file, load this into beast run it and proceed from there.) If you want, as others do, you can use model-test function in MEGA to test for the best model.
Cheers
Eugene
One thing that I will suggest is that you not use clustalW for alignment purposes. It produces poor quality alignments, compared with more modern alignment programs, and as noted by its authors since the introduction of the newer clustal omega.
The good thing is that you can use a webserver to use clustal omega:
http://www.ebi.ac.uk/Tools/msa/clustalo/
Save the output to a fasta file, open it in clustalX, and save it to .nxs again. A good alignment is critical for producing good phylogenetic trees.
I do not know what the T4 in your question stands for (it might be just ignorance of mine). The GTR model is easily implemented via Beauti. You can also implement models that do not appear listed in Beauti as options by opening the file produced by Beauti and edit it in a text editor (e.g., TextWrangler).
In addition to the excellent previous recommendations, you can review this tutorial:
https://beast.community/first_tutorial
Regards,
C
Dear Harindra Sathkumara,
I always refer to this thread: https://groups.google.com/forum/#!topic/beast-users/_zCoCTSkWRs in the beast-users group when I need to set a substitution model that is not in the default Beauti options.
Hope this can be helpful,
Cheers,
M
- Badges
- Science topic
- Similar topics
- Bioinformatics
- Bioinformatics and Computational Biology