search

monarch-initiative / monarch-legacy

Monarch web application and API
BSD 3-Clause "New" or "Revised" License
42 stars 37 forks source link

remove models without matches from getSimilarModels #407

Closed nlwashington closed 9 years ago

nlwashington commented 10 years ago

right now, for zebrafish matches to OMIM:255800, it gets a bunch of "matches" with a sim score of zero. these should be removed from the api call results.... they should at least be filtered from the monarch api call, but maybe should even be removed upstream from the owlsim calls.

cmungall commented 10 years ago

agreed. This would be in findMatches() in OwlSim?

nlwashington commented 10 years ago

yep. this could be done either with findMatches() in owlsim, or it could be in the monarch api simsearch which simply wraps the findMatches().

nlwashington commented 10 years ago

it's actually whatever the server call for searchByAttributeSet maps to.

nlwashington commented 10 years ago

when i tried to investigate this one, there are now MANY zebrafish matches (with the improved ZP), and so i'm having a hard time finding one with scores of zero.

so, this is not a good test case, and neither are many of the other diseases/phenotypes i've been trying. (even fever,HP_0001945, gives matches!)

but i got to thinking; i bet that in that previous case, the reason those genes were returned is that one of the scores passed the basic owlsim filters, it's just that the combined score > 0. therefore, i think it might be the responsibility of the monarch api to do the filtering, since it's what is setting the scoring metric.

jmcmurry commented 9 years ago

I'm not sure how best to investigate the status of this, so perhaps someone knows off the top of their head. Can this now be closed?

nlwashington commented 9 years ago

i think this might be fixed. FWIW, fever doesn't have any matches...(though i don't think that's correct behavior). closing for now; we'll reopen if it's not fixed.