Ordering of inference method in curator dropdown

[josephwb]

Perhaps nit-picky, but commonly used methods should be at the top of the list. Currently we have:

Maximum parsimony
Maximum likelihood
Neighbour-joining     <- don't forget your 'u's!
UPGMA
Bayesian inference
Least squares
Three-taxon analysis  <- I don't think this one need to be in the list at all
Other

I suggest we go with the following ordering:

Bayesian inference
Maximum likelihood
Species tree inference <- note: not in original list
Maximum parsimony
Supertree              <- note: not in original list
Neighbour-joining
Other

Since this might be currently on the table, perhaps we should come up with a controlled list of accepted names.

One thing I notice we don't have is a way to distinguish "molecules" from "morphology". This is one piece of information that we should be indexing. "Molecules" could of course be more narrowly defined as "DNA", "Amino acids", "Transcriptomes", etc. Erg. this is sounding like another issue...

[josephwb]

I did some minor reordering with 6138378a4ec3b2c3c2013668261d855e60ecf209, but have otherwise not altered the list.

[kcranston]

Being discussed in this thread on the opentreeoflife mailing list.

[rossmounce]

Why exclude known & used (albeit rare) methods from the list? It's not exactly a long list. I think we should be exhaustive as possible here.

e.g. I don't see why you need to specifically exclude Three-taxon analysis

I'd also like to add Minimum Evolution (ME) to the list http://www.megasoftware.net/mega4/WebHelp/part_iv___evolutionary_analysis/constructing_phylogenetic_trees/minimum_evolution_method/rh_minimum_evolution.htm

it's out there, it's implemented and it's in-use (albeit rare). All reasons for inclusion IMO.

[tjv]

The EDAM ontology [1] already has a breakdown, less granular in terms of methods (ie ML and Bayesian are combined) but more granular in terms of datatypes. Their breakdown may be useful to consider in places, and I'd suggest letting the ontology maintainers know what you decide in case they wish to revise their own classification. [1] http://bioportal.bioontology.org/ontologies/EDAM/?p=classes&conceptid=http%3A%2F%2Fedamontology.org%2Foperation_0323

[rvosa]

Although I usually favour this kind of spelling, I'm pretty sure neighbor-joining is consistently written without the u.

[rossmounce]

ah yes, 'neighbor-joining' it is. My apologies. Hard to override Englishness ;)

[jimallman]

Does the current list (below) satisfy this issue? If so, please close.

[josephwb]

I think this is now solid.