TR_Mock-up
Discussion of "proof of concept" resolution service
The mock-up to be presented at the 1kp meeting will not be functioning software but a vehicle to solicit ideas, requirements, and design input from the meeting attendees. We would use the information to help us deliver our prototype in early 2010.
TreeBEST is a possible candidate. It accepts as input aligned sequences and a species tree and produces maximally resolved gene tree. It is distributed as C source code which is appealing for integration.
Prime-GTR is also to be evaluated.
See opportunity to start with sequences. Not sure alignment support in TreeBest but we can add that support if needed.
Cecile – it would be a good idea to integrate alignment. Value added in comparing reconciliation among gene families.
Pre-compute alignment stored as HMM. Question whether to use existing database of alignment or do plant wide transcript alignment?
Discussion of using "canned" data for demo purposes from Plant GDB (http://www.plantgdb.org/)
Comments
Sorry, these are my anonymous comments
I think we should focus on tree reconciliation. Others can circumscribe gene families and provide alignments (e.g. PlantTribes).
While 1kp is really interested in estimating species trees from gene trees, we should make it clear that we are focused on tree reconciliation, although primerGSR could be used to estimate gene trees given gene family alignments and species trees.
Given our focus on tree reconciliation, I am thinking that 1kp folks may be most interested in having us provide methods for querying reconciled trees and providing views of the following:
the number/proportion of gene families showing duplication events at each node (or specific nodes) on the species tree,
duplications on each node broken down by GO categorization,
show hotspots on the species tree for duplications for a gene family or GO category
Test for association between duplication events across all gene families or GO categories and speciation rates.