Extracting Subtree from OpenTree

[mwestneat]

How can a user extract a subtree from OpenTree by specifying specific clades and pruning the giant tree down to the taxa desired? a la PhyloMatic or PhyloTastic

[jar398]

This is a planned feature but isn't ready yet. Look for it in a month or two, or, if you prefer, help us finish the implementation... I think the code is there but there's no web service for it yet. See https://trello.com/c/fd2lr60r/48-extract-subtree-from-synthetic-tree-given-taxon-list

[bsidlauskas]

I like this idea and think that this capacity would be useful to many people.

[arlin]

Various versions of this idea would be useful. An obvious one is to submit a list of tokens (names), match this against the list of species in the OT reference taxonomy, and extract a tree containing only these species from the synthetic tree (or find any source trees that match M or more of the N species where M <= N). This would be useful to people who have comparative data on a set of N species and would like to get a phylogeny to reconstruct ancestral state or to test hypotheses about character evolution.

Often people want to find a phylogeny covering a particular clade, such as birds. If higher taxonomic categories are linked to the synthetic tree, then a clade for birds could be covered using the taxonomic term Aves. If the synthetic tree isn't linked to higher taxonomic terms, then maybe there is some way to use a phylocode-like definition (i.e., birds are the least inclusive clade with both archaeopteryx and sparrow, as in http://www.taxonsearch.org/dev/taxon_edit.php?Action=View&tax_id=43) to specify a root node, then extract the subtree from there. In many such cases, the user will not want the entire tree (TMI), but a representative sample. Often people want just the familiar or well known species. In principle, given some metric for familiarity (number of hits online, or in GenBank, or in NCBI genomes), you could prune down a larger tree to a smaller one with the most familiar species.

[bbanbury]

This feature would be very useful. I think adding metadata (for example which source trees contribute information) would also be important. Also, extracting subtrees from specific sources could be handy as well (direct comparisons across studies).

[temi]

I think this feature will be great. I am also interested in getting the pruned synthetic tree using taxons from a study that contributed to synthesis. The two trees then could be compared. Like @bbanbury suggested, question like what other trees contributed to the synthesis can be answered if the required metadata are provided. Also, it can be used find related studies.