OrthoMCL: How to parse groups.txt for a phylogeny?

Question

[apelin20]

Hello,

I have 20 or so microsporidian genomes who are mostly assembled into a draft genome. I have extracted protein sequences for predicted genes in these genomes, and have clustered them with orthoMCL. I now want to construct a phylogeny, for which I need orthologous groups, prefereably consisting of single copy genes.

However, what I notice is that groups.txt also shows orthologous groups with multiples sequences per genome. How can I find a group in which all 20 genomes have only one sequence present, indicating that it is likely single copy?

Thanks,

Adrian

Answer 1

[h.mon]

Which version of OrthoMCL? I can not remember, but I believe both 1.4 and 2.0 output the clusters files named as "X_genes_Y_taxa", so you just have to search for files with same number of species and genes - something like:

find . -name "*20_genes_20_taxa*"

should work. If I remember the naming convention incorrectly, forgive me for the misleading answer.

[apelin20]

Hello, I am using V2 which I hear is very different than prior versions, the only output is group.txt, as well as 3 other files that show all pairwise comparisons (http://orthomcl.org/common/downloads/software/v2.0/UserGuide.txt). I figured out one way to do this is by greping iterativly all taxa and finding groups which have all taxa, then look for clusters that have a size equal to the total amount of taxa, so 20, which would show you single copy orthologs. Not sure if there is a cleaner way to do it.

[Lhl]

Hi Apelin20. I am doing the same thing to orthomcl2.0 output -- groups.txt. Have you, by any chance,, found any other way of processing the output for downstream analyses?