supertrees using tree ranking

Constructing supertrees from trees with ranks

General background on how the supertree relates to the Open Tree of Life project

The overall goal of the project is to summarize what is known about phylogenetic relationships in a transparent manner with a clear connection to analyses and the published studies that different clads. Comprehensive coverage of published phylogenetic statements is a very long term goal which would require work from a large community of biologists. The short-term goal for the supertree presented on the tree browser is to summarize a small set of well-curated inputs in a clear manner.

We hope that this tree (and its many deficiencies) will inspire biologists to help fix the tree by alerting us to relevant phylogenetic statements that are missing in our set of inputs or contributing to the Open Tree effort in other ways. We have built tools to support 2 types of feedback:

1. Our phylogenetic study curator tool for adding new input trees.
2. Clade-linked issue reports. Click on "See comments" on a node of the tree, such as the Tetrapoda node to see a conversation that is tied to a particular phylogenetic grouping. Full issue-tracking features for that conversation (e.g. notifications, authenticated comments, tagging, closing of the issue) are supported using GitHub issues system. This issue on GitHub contains the thread about the groupings shown at the Tetrapoda node linked above.

Of course, emailing comments is another way to provide feedback (see the opentreeofliife group for general comments and this list for software issues).

Requirements of the supertree operation

Note: MTH put together this list. #3 may not be agreed upon by all participants.

1. produce a full supertree: The set of taxa associated with leaves of the supertree is the set of "terminals" in the taxonomic input.
2. the tree should display (in the sense described here) as many edges from the curated input trees as possible . If the scoring criterion was a count of the number of groupings, then this would surely by an NP-hard problem REF needed. 1. ranked trees property: For simplicity of implementation, we use a tree-based ranking system to resolve conflicts for simplicity. 2. taxonomy last property: the taxonomy is the lowest ranked input
3. Display no unsupported groupings sensu the definition describe in this doc.
4. this one definitely needs some work: we might want some statement about the placement of groups that appear in only one source tree. Perhaps: when there is a large set of attachment points to which a group can be attached and still display the same set of input tree edges, then the group should be attached at the point closest to the root that is consistent with these displayed set. This placement is in the spirit of the Adams consensus of all of the attachment points.

technical links

See this page for one discussion of a scoring criterion which MTH thinks could be viewed as a criterion that the automated "tree synthesis" attempts to solve. This page will discuss some algorithmic considerations about this style of synthesis.