Clustering program based on NJ phylogeny

Question

[kgbenn123]

Does anyone have a recommendation for a clustering program that uses neighbor-joining phylogeny to cluster a dataset of protein sequences and output a representative from each cluster? I'm thinking of something that works like cdhit, but uses NJ instead of sequence identity for clustering.

Any suggestions?

Answer 1

[leeandroid]

Check http://evolution.genetics.washington.edu/phylip.html.

[kgbenn123]

Phylip doesn't appear to have the function I'm looking for. I'm working with about 6000 seqs and want a program that can cluster through neighbor joining and output either a list of accessions that represent their respective clusters or some file format that allows me to extract them.

Answer 2

[gbl1]

Even if late NJ (neighbourth joining) clustering is pretty easy to do on R function nj()