This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Is it possible to create a phylogenetic tree based on 10 proteins per organism

Hi,

I have 10 genes that have different functions between them and are present in 23 organisms. I would like to know if it is possible to create a tree based on organisms and not on genes that have the same function. Based on this, I have to create 10 trees because I haven't found a tool or program which allows me to align 10 genes vs 10 genes vs 10 genes etc

Normally, for example, people align one sequence per organism and perform muscle, and afterward, they create a phylogenetic tree through a statistical method like maximum likelihood. But the programs but programs do not discriminate on the basis of species.

I found that bcgtree is very useful because it compares proteome vs proteome vs proteome and creates a tree based on the organism, but I only need 6 genes.

Thanks

Hope to be clear.

phylogenetic-tree
• 1,670 views • link
updated by Ram Ram • written by Juliana • Juliana

1 answer

  • Create multiple sequence alignments for each gene separately but for all 23 organisms in one alignment
  • Once you got the multiple alignments, simply concatenate them into one long alignment
  • Do not edit the alignment any further after this step (unless you remove complete rows or columns), in particular, do not re-align the concatenated alignment
  • run a phylogenetic reconstruction (e.g. IQtree, RaxML, MrBayes) on the concatenated alignment as if it were from a single gene to create a single phylogenetic tree (start with IQtree, it is most easy to use) Edit: create partitions for each gene, see comments below.
• 0 views • link
written by Michael Dondrup Michael Dondrup

I agree, except for the last point. You should allow different sequence evolution models on different blocks (genes) of the concatenated MSA. I know for sure this can be done with MrBayes, and guessing IQTREE also has this option.

• 0 views • link
written by liorglic liorglic

It's a good point and it might make sense, however, I don't know how to achieve this in neither MrBayes nor iqtree. You could possibly allow for more rates of the gamma distribution.

• 0 views • link
written by Michael Dondrup Michael Dondrup

It's called partitioning. See this page for example for IQ-TREE.

• 0 views • link
written by liorglic liorglic

Log in to answer this question.

More posts like this