CodeML stop codon

[audr...@umich.edu]

When I run CodeML, with all protein coding data, I get stop codons. But, when I take the sequence file and translate it in Geneious there are no stop codons. I am sure there are no stop codons in my sequence file, has anyone had this problem before?

Thank you in advance.

[Andrey Yurchenko]

Hello! I am using codeml with nucleotide data and bacterial code(10). I am getting all the time the same error: 

stop codon TAAM in seq # 1

What it is?

[Ziheng]

Not sure why the program is printing out taam rather than taa as a stop codon.
Make sure you use the right value for icode.
It is likely that you still have stop codons. Use verbose = 1 and look at the listed sequences in the output file, to see whether you have a taa codon.
Ziheng

[Janice Valencia]

Hello,

I have the same problem in a Mac:

stop codon TAAM in seq # 1

I'm working with plants, so icode: 0

:/

I run it also in windows and it worked well... so, I don't know what could it be. But I'd love to know if there is any solution. 

Thanks.

Janice

[Ziheng]

i haven't seen this before, i think.  does it make any difference if you use a different or newer version, is that is relevant.

i notice you are using a mac.  does it make any difference if you compile codeml using different flags.  

cd src/

here are some possibilities:

cc -o codeml -O3 codeml.c tools.c -lm

cc -o codeml -O2 codeml.c tools.c -lm

cc -o codeml -O1 codeml.c tools.c -lm

i have the impression that there are sometimes problems with the clang compiler that apple now uses.

ziheng

[Nathanael Walker-Hale]

Hi Ziheng,

I just wanted to inquire if you knew about how this error might also crop up during ASR with baseml? I am reconstructing coding sequence data with GC in the data file, model=4 and mgene=4 as suggested in the manual, using paml4.9h. It appears to run quite well but then quits with this error:

NG86: Stop codon TGA

Does this mean the program has inferred a stop codon and quit?

Best wishes,

Nathanael

[Ziheng]

if the program runs, you can ignore the error.  the NG86 calculation is unimportant.

the baseml reconstruction of coding sequences does not rely on a codon model, and may thus generate stop codons in ancestral sequences (although this should be rare), and the stop codons may cause some trouble when the program counts differences using the NG86 algorithm.  

if the dataset is small, you can use codon models in codeml to do the ancestral reconstruction, and compare with the results from baseml.

ziheng