How can you explain the high rate of substitution for single branch of BEAST tree?

Question

[konstantinkul]

Hi everybody!

I found a branch on the phylogenetic tree (constructed using BEAST) which shows an abnormally high rate of substitutions in comparison with others branches. How this can be explained? Whether it can be as a result of abnormal environmental conditions in which organism (bacteria) existed?

Thanks in advance for any response.

[5heikki]

The hand of God. This is not a bioinformatics question. I recommend you post it at http://biology.stackexchange.com/

Be sure to include many more details in your post like e.g. what gene/s the tree is based on, how the alignment is, etc.

[konstantinkul]

Thanks for suggestion!

Answer 1

[DG]

This is really (likely) more of a biology question than bioinformatics but, especially when dealing with bacteria, or really microbes of any sort, having one (or a few) highly divergent taxa in a dataset isn't unheard of. Look at any phylogenetic tree for microbial eukaryotes that includes microsporidians for instance and you often see branch lengths that are 3-10x longer than anything else in the tree. They are highly reduced parasites and that creates a lot of divergence, particularly when looking at just single genes or a small subset of genes.

But there are some things to consider from a bioinformatics perspective as well. When you have extremely long branches (or short branches) it often results in incorrect lineage placement within the tree. So you need to be wary when interpreting the tree. Particularly if there is more than one of these long branching species you can run into artefacts caused by long branch attraction. There are plenty of papers published on this that may be of interest. These issues also mean that selection of your substitution matrix, gene sampling approach, taxon selection, etc are all very important. So you'll want to look very carefully at your data and analytical methods to see what may need closer attention and adjustment.

[konstantinkul]

Ok! I get it. Thank you, I'll revise all data.

[DG]

I'm not saying you need to revise the data, just saying that you want to look at your data, methods, parameter choices, model choices, etc carefully. Also take into account the biology. What species is this? What are the other species in the tree? It may very well make sense for it to be a very long branch.