Biostar Beta. Not for public use.
Comparing Phylogenies Built From Different Data Sources
4
Entering edit mode
9.9 years ago
Mara • 70
@Mara1276

I want to compare phylogenetic trees built using different aspects of evolution, i.e. genetic variation vs. structural variation. Certainly genetic variation is much larger than structural variation. Anyone have thoughts how I could make the two matrices comparable?

phylogenetics evolution • 1.7k views
ADD COMMENTlink
1
Entering edit mode

Mara, be careful with that word "signficant", bootstraps on phylogenies aren't significance tests, they're a measure of how consistently the data you have reflects the tree you've estimated. If the data is biased you can get large bootstraps for the wrong tree.

ADD REPLYlink
0
Entering edit mode

It is not entirely clear to me which two matrices you are talking about in the question. Is it the case that you have two distance or similarity matrices on which the phylogenetic trees are based, and you want to compare these matrices?

ADD REPLYlink
0
Entering edit mode

Hi Lars, I have one matrix that describes the distances between protein structures (25 structures) and another that describes the nucleotide variation in the data set (25 sequences). For each of these I generated a phylogenetic tree. The phylogenetic tree for the nucleotide data shows all branches are significant, but for the distance between structures, the variation is very small and the none of the branches are significant. I thought that there might be someway to weight the data so that the two data sets were more comparable.

ADD REPLYlink
0
Entering edit mode

did you try building a tree for the protein sequences (instead of the nucleotide seqs) after doing a multiple alignments->distance matrix ?

ADD REPLYlink
0
Entering edit mode

I am NOT certain if these kind of comparison meaningful. The nucleotide sequence evolving much faster than protein sequence, the protein sequences evolving also much faster than protein structure. That is why the tree is not significant from structure, but significant for nucleotide sequence.

ADD REPLYlink
1
Entering edit mode
9.9 years ago
@Jarretinha148

Hi Mara,

You can use a supertree approach using your datasets and some sort of "rarefaction" tree construction, i. e. adding data and tree construction in a stepwise manner By comparing supertrees with different amounts/types of data you can address the contribution of each character and (with a bit of luck) estimate heterotachy or similar effects. Supertrees can combine any kind of data.

ADD COMMENTlink
0
Entering edit mode

Thanks, this is a really interesting program.

ADD REPLYlink
0
Entering edit mode

@Jarretinha - I'm geting a "DOI Not Found" on your link. Would it be possible to update?

ADD REPLYlink
0
Entering edit mode

I've updated (and double checked) the link. It's working now!

ADD REPLYlink
0
Entering edit mode
9.9 years ago
@Casey Bergman314

You can try the pairwise tree comparison method of Nye et al. (2005) which is implemented as a java applet here. It takes 2 newick trees from the same OTUs and allows you to visualize where a clade in one tree is found in the other.

ADD COMMENTlink
0
Entering edit mode

Sweet program. Very easy to use with an appealing output. Thanks.

ADD REPLYlink

Login before adding your answer.

Similar Posts
Loading Similar Posts
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.3