detection positive selection using multiple copy genes in orthologs

Question

[Mehmet]

Dear all,

I have done positive selection analysis using one-to-one copy ortholog groups via codeml tool. But I have seen some papers in which multiple copy gene ortholog families were also used to test positive selection. When I tried to do same analysis (multiple copy approach), ETE tool used only one sequence (one gene) from each species although each species has multiple copies in the same ortholog family.

Is there any way to deal with gene duplication event because ETE uses only the first sequence of each species?

Codeml in PAML says Seq occurs more than once in the alignment.

Any of you has done this kind of analysis (using multiple copy orthologs) before?

[h.mon]

Are the sequence names duplicated? Or are the names unique? Are there repeated sequences? ETE and codeml doesn't care about species, but they probably don't like duplicated names and / or duplicated sequences.

[Mehmet]

Sequence names are duplicated because the sequence names must match with the names in the tree. For instance;

in the sequence file; names are :

SpeciesA SpeciesA SpeciesA

SpeciesB SpeciesB

This is the name order in the tree as well. (SpeciesA, SpeciesB so on). When I use sequence name as different (SpeciesA_seq1, SpeciesA_seq2...so on), the names don't match with the names in the tree. This is another problem.

Answer 1

[h.mon]

Each gene needs a unique identifier, and alignment and tree gene names must match. So your genes should be named SpeciesA_seq1, SpeciesA_seq2, SpeciesB_seq1, and so on, at both the nucleotide alignment file and tree file.