Is there any way to quantify differences between neighbour-joining phylogenetic trees?

Question

[matt.ward93]

I am working on a project and I am currently only using the SNP data to make my trees. I would like to compare the trees made using SNP data to the trees made using indel data to ensure they are similar, however I am unsure as to how to go about doing this?

Answer 1

[Brice Sarver]

There are many established ways to quantify differences between and among trees.

Perhaps the simplest is the Robinson-Foulds distance, which looks for groups present in one tree that are not present in another. The Kuhner-Felsenstein difference can also quantify differences taking branch lengths into account.

If you want a simple metric, you can look for differences among trees based on the number of nearest-neighbor interchanges (NNI), subtree pruning and regrafting (SPR), or tree bisection and reconnection (TBR) distances between them.

These metrics are relatively simple to calculate and are implemented in many R libraries and phylogenetic packages.

A word of caution: be careful making trees using model-based approaches from SNPs. There aren't any sophisticated models of evolution for SNP data, and a dataset constructed just of SNPs violates many of the assumptions of the methodologies.