[Bioperl-l] Codeml runtime error

Thomas Sharpton thomas.sharpton at gmail.com
Wed Mar 24 17:43:48 UTC 2010


Hi Bioperl gurus,

I'm trying to run PAML v4.3b on a series of orthologs, specifically by  
implementing codeml to detect signatures of positive selection between  
all orthologous pairs. In some of my files, I notice that I'm getting  
an EOF error that causes codeml to break. The weirdness is that I only  
get the EOF error under one hypothesis model (the null) and never on  
the alternative hypothesis model - even when run on the same initial  
data.

I've managed to track the problem down to the way BioPerl formats the  
temporary phylip alignment file that is fed into codeml. Apparently,  
PAML requires there to be at least two spaces between the sequence  
identifier and the start of the sequence. However, for some files -  
and I don't know if this is random or not - the temporary alignment  
file only contains one space after the sequence identifier. If I edit  
the phylip file accordingly and rerun codeml, the software compiles  
and processes the data correctly.

Has anyone run into this problem before and has someone figured a work  
around using the kaks_factory in  
Bio::Tools::Run::Phylo::PAML::Codeml.pm? If this is something others  
have not seen, I'll submit a full bug report.

Best regards,
Tom





More information about the Bioperl-l mailing list