How To Increase Size Of Branch Of Phylogenetic Trees

Question

[Pavid]

Hi everybody!

I trying to create a tree, since I have many samples, one of the steps I make is to group all the equal samples together.
For instance I have this file:

>1
ATGACGAGATAGACAGA
>2, 3, 45, 6 
ACATAGGTATATATGGG

well I think I've made my point.

So I would like to distinguish then into the tree, like increase the size for each branch depending on the number of equal samples.

If Anyone has any other idea how to do this automatically I would also appreciate?

Thanks guys

[Pierre Lindenbaum]

how are you building your tree (without thinking of the size of the branch) with only the information available from this file ?

[Larry_Parnell]

Is this meant to be a tree depicting similarity or parismony or something else? I find the question vague.

[Pavid]

@Pierre I would most likely build my tree with ETE, but I may use another software. @Larry this is meant to be a tree to show similarity, that's the reason why I group then if there are equal. And also the reason why I want to increase the size of the branch

Answer 1

[Dave Lunt]

This isn't an easy one Patricia, because of the sequence collapsing you've done at the start. If you instead kept all sequences included then many programs (e.g. FigTree) could display the tree with nodes as triangles and width proportional to number of sequences in that group. Unless you have very many thousand sequences you shouldn't see much of a slow down in building your tree with any modern algorithm (e.g. RAxML, FastTree).

If you really do want to collapse identical sequences to haplotypes as you show then you could try ScripTree or Treedyn. You need to create an annotation file with the number of sequences each name represents. You could then annotate that sequence with a graphic of size determined by the annotation file value. See example 6, tree symbol annotation, on the ScripTree page.

A third suggestion is that you just get a tree as you are doing know. Output a image-file of the tree from any program you like then just artificially stretch the group you are interested in proportionally to the number of individuals sharing that sequence. In terms of phylogenetics that axis has no meaning, so your changes don't alter the tree itself. This could be a lot of work though if you have many groups!

[Pavid]

Thank you for your response Dave. Well I did the step of group the equal sequences because I've lots of sequences, right now I'm trying only with 600samples, but I've more than 900. with these step I was able to decrease significantly the number of sequences which was very helpful since most softwares have a 500limit. And I needed something like ClustalW to change my fasta to phy or nw, so that I could build my tree.

[Pavid]

I gone try both your suggestions anyway :)