How to determine sample clusters from phylogeny

Question

[rosies]

Hi, I have constructed a phylogenetic tree (Roary, FastTree) of outbreak bacterial isolates (staph) and am wondering how I can determine which isolates form a cluster. Also, do you have any advice on how to be consistent with cluster definitions when more isolates are added to the analysis.

Thank you for your help!

[Ron]

Do you want to see the clusters in hierarchical clustering?

[rosies]

Hi Ron, I'm a bit naive in the space of clustering algorithms, but, yes, hierarchical clustering looks good. I'm hoping the clustering algorithm can divide clades into clusters (and provide a statistic for where it calculates where a clade/cluster starts/ends). Do you have any recommendations for software that could determine the clusters? Thank you.

[Ron]

http://www.sigmath.es.osaka-u.ac.jp/shimo-lab/prog/pvclust/

Clusters with AU larger than 95% are highlighted by rectangles.Its similar to a p-value.

Answer 1

[Joseph Hughes]

Use ClusterPicker to define clusters based on branch support and pairwise distances of the sequences or if you are trying to delimit species use something like PTP/bPTP/GMYC.

[rosies]

Fantastic. Thank you Joseph. I tried it on my data (just the gui) but it doesn't look like it likes my dataset (112 samples at ~3MB each). I'll try via command line on the HPC to see if this works.