search

geneontology / web-gocam

GO-CAM site as an Angular 2+ Application
http://geneontology.org/go-cam
BSD 3-Clause "New" or "Revised" License
1 stars 0 forks source link

Gene product display on GO-CAM page not ideal ? #14

Open pgaudet opened 3 years ago

pgaudet commented 3 years ago

Hello,

On the GO-CAM site https://geneontology.cloud/browse gene products show up as IRIs/links - is this intended ?

I thought the gene name was shown

image

Thanks, Pascale

lpalbou commented 3 years ago

Yes, it was showing the gene labels before - at least when present in NEO.

I looked around and I think this is due to the changes of NEO made by Ben sometime ago. I fixed the query and we do get most gene names again.

There are however still a few genes that won't resolve: Screen Shot 2020-12-15 at 12 16 40 AM

Interestingly, they do resolve in the graph editor: http://noctua.geneontology.org/editor/graph/gomodel:5e72450500001653 so this would indicate the triple store in production (rdf.geneontology.org) doesn't have the same NEO/gene info than minerva.

I think this is a sufficient fix as most gene names do resolve now, if that's the case, feel free to close that ticket.

pgaudet commented 3 years ago

Great, thanks for the quick response !

lpalbou commented 3 years ago

After discussion with bbop and to keep a reference of that xenbase gene that doesn't resolve with the blazegraph in production but does resolve within minerva (from noctua graph): https://api.triplydb.com/s/oSw3hwZ2l

And as a counter example when it works: https://api.triplydb.com/s/zJub7DDmy

balhoff commented 3 years ago

@lpalbou how did you change the query to make it work? The problem is that NEO stopped being loaded into the graph store. I will fix that. But I'm curious what your change was, since as far as I know there should be no gene labels loaded in the current configuration.

lpalbou commented 3 years ago

@balhoff Ah Jim, sometimes I wonder what we actually have in our triple/quad store ;) Like our friend Forrest Gump and its box of chocolate, you never know what you are gonna get with a schema less database.

This was my fix. Instead of looking for the names in the Graph http://purl.obolibrary.org/obo/go/extensions/go-graphstore.owl, I started looking at the rool level. I didn't do some archeology to find out from which graph the data came from.

Thanks for https://github.com/geneontology/go-site/issues/1599.