CodeML Troubleshooting Help

Question

[p.juttyrajan]

Hi all,

This is my first post in the community, so first off, thanks for the time you spend in replying to people's queries. It's a great help and I've already learnt quite a bit from just reading through.

I'm trying to calculate the rates of evolution in my set of sequences and I've been learning to use the CodeML program that is a part of the PAML package. I have made my sequence file in the Phylip Interleaved format and I've made my phylogenetic tree in the Nexus format. Initially, I simply wanted to run the M0 model and play around with the parameters to get used to the system.

But I have one problem, the program appears to read the sequences and produces a results file in the right folder but the results file is empty (zero bytes). I ran some of the test sequences available online and the program works properly but I cannot for some reason get it to work on my sequences.

I tried using the Nexus file format but it produces the following error,

Error in sequence data file: X at 102 seq 1.
Make sure to separate the sequence from its name by 2 or more spaces.

I looked into the sequence file with a text editor and cannot see any errors as far as I can tell but that's where I think the error lies.

To summarise, when I use the Phylip interleaved format, PAML produces an empty results file. When using Nexus or sequential formats, it returns an error message.

Has anyone faced this issue before and if so, how did you manage to solve it ? Are there any reliable methods of producing these sequence files? So far, I've used MacVector, Ugene and the online sequence format converter, Readseq. In each case, the errors are the same. I would much appreciate any help.

Thanks in advance,
Prashanth

[marina.v.yurieva]

Did you see Phylogenetic Analysis By Maximum Likelihood (Paml) : Upon Phylip Input Yn00.Clt Program Not Running. post? Which editor did you use to look at your Nexus file? If you type :set list in vim it shows you tabs as "^I" characters. Make sure you have the correct delimiters.

[p.juttyrajan]

Hi, Thanks for your reply Marina.

I followed the other thread and tried implementing some of those suggestions. The program now reads my sequences but continues to produce an empty results file.

I've just installed Text Wrangler and I'm trying to look at the delimiters and to see if they might be the problem.

Thanks,
Prashanth

[Brice Sarver]

I think this reproduces an error that I solved a while back, but I want to be sure before I provide the answer. Could you provide an example taxon name?

"Error in sequence data file: X at 102 seq 1." suggests it may be '102 seq 1,' but I can't remember if 102 is a position. Is this right?

Answer 1

[Diana MoSa]

Hi there! I don't know if you fixed at the end. I had the exact same issue. I check the file in many programs to see if there were a wrong symbol but everything was fine.

After several minutes of anger I figure it out which was the problem! There were three gaps missing at the end of my sequence, therefore the length was not right, my sequences are 282 bp but without those three gaps that sequence was only 279 bp. So, even if I did have my two spaces after the ID, the actual reason was completely different.

I wrote this in case some one else have the same issue, before anything first check that all sequences have the right length.

[niconm89]

Thanks, you have saved me a lot of time!

[chenyangkang24]

Oh my god that's the point. I did orthologs codon-alignment which gives me the information of frame shift mutation in as sign "!", and I just convert the alignment to phylip fortmat and pass it to codeml. I think something's gonging wrong with the convertion, to deal with the "!", result in erro length info.

Answer 2

[k.balaji93]

I think the problem is with unintended whitespaces or tabs. I couldn't figure out where the extra spaces were, so I used write.dna() from library ape to write the output in phylip format. That fixed the problem. Make sure to remember to have two spaces after every sequence/species name.