Hi,

I only run CODEML in command line, and I always did prepare all the input files (control file, tree, and sequence file) by myself, not with pamlX - but I see some issues that might be a problem here.

1) The sequence names in .phy file are different that in .tree file - phy sequence format has a restriction on the number of character is the sequence heather/name. You can see this if you open these files in e.g., Notepad, Notepad++ or another simple text editor. I think you should first change names of the sequences in fasta file, so they have up to 10 letter, and only then make a phylogenetic tree and change fasta into phy format.

2) I am not sure if paml will run with sequences of different length – it might not be able to process “-“ symbols in the sequences! My advice would be to trim the sequences so all of them have equal length.

3) Another problem is, that generated control file – aves.ctl – should contain the names of your input and output files, but they differ from what you provided. Again, if you open aves.ctl in a text editor, you will see at the top:

seqfile = seqfile.phy * sequence data file name

treefile = tree.tree * tree structure file name

outfile = test.mlc * main result file name

Instead, seqfile should = aves.phy, and treefile should = aves.tree – only then paml can identify them. Maybe seqfile.phy and tree.tree where the names returned by the software along the control file in the first place, but you than changed them to aves.tree and aves.phy?

Also, if I understand correctly – pamlX main function is generation of the input files for paml, especially the control file.

From the paper introducing pamlX:

“The PAML programs use a simple command-line interface, and the task of preparing the control files in a text editor may appear daunting to beginning users. We have thus developed a graphical user interface (GUI) for the package, called PAMLX, which can also be used as a front-end to launch PAML analyses. The main function of the interface is to create, read, edit, save, or print the control files for the main programs in the PAML package, including BASEML, CODEML, and MCMCTREE.”

Once the files are generated, it can parse it (?) to the actual paml program to do the job, but you need to download and setup paml itself – I hope all these steps are also done?

I hope this can be helpful!

Best, Magda

Magda answer is excellent. Paml is very good but has a steep learning curve and your answer cannot be answered without the specifics of your attempted run.

An alternative is Hyphy - Data Monkey online- which has a variety of selection tests for molecular data. The developers - like Ziheng Yang with PAML- are brilliant with maximum likelihood methods so you should should get good results if you are doing the right test. The documentation on the Data Monkey site is quite good for identifying the particular test you need.

From your objectives, I would think you might use BUSTED in Data Monkey.I attach some papers that might help. And would be happy to try to answer any other questions.