DiseaseOntology / HumanDiseaseOntology

Repository for the Human Disease Ontology.
Creative Commons Zero v1.0 Universal
329 stars 108 forks source link

New feature suggestions #1211

Open vjenkinsFB opened 1 year ago

vjenkinsFB commented 1 year ago

Hi DO,

I often want to search DO terms by associated "has_material_basis_in" a gene. There are a few things that get in the way of this:

1) I need to advanced search > search definition, which, okay, I can do. See below for a way to address this. 2) this search needs to be exact case to get any results. 3) Although the text makes it look like the gene is associated in some kind of logical way with the disease, ie that would show up in the relationships below the definition, it isn't. Compare DOID:0050012 chikungunya, which has a logical relationship link to NCBITaxon:37124. However, DOID:0060369 Parkinson's disease 6 has no gene link in the relationship section, despite having "has_material_basis_in mutations in the PINK1 gene" in the definition.

Are there plans to link individual genes to DOIDs that claim them as what the DOID has_material_basis_in? This would really help searching for these single-gene diseases, as you could search for them via advanced search > relationship. Also, you could make that case-independent easily, for people like me who forget that human genes are all-caps only. I imagine this is already on your roadmap, but I just wanted to check and see what plans are for this if so. Thank you!


If I'm spitballing more features I would like to ask Santa Claus for this year: it would be amazing if all searches ran in the standard format (name only?), as well as in all advanced versions, and then had a return which mentioned which type of result it was, like if there was a disease called "YYY-related disease" caused by the gene YYY, and the user searched for "YYY":

search type DOID result relationship type (for search type "relationship")
name DOID:XXX YYY-related disease
definition DOID:XXX YYY-related disease has_material_basis_in YYY gene.
relationship DOID:XXX HGNC:YYY has_material_basis_in

I bet a lot of people search for something by a string that would be in the definition, don't get any results in the default search, and assume DO dosen't have anything containing that string. Something like this would make sure that they don't get the no-results screen and then leave, when DO does actually have what they are looking for. Not searching the rest of your data by default can prevent an overly-long results list, but it also lets people overlook what the DO curators spend a lot of time on (definition, xref, etc.). Something like this could address those issues, as it lets people be aware of what results they're not getting w/the simple search, which still prioritizing the "name" results at the top of the results list.

Thank you!

allenbaron commented 1 year ago

Thank you Victoria for outlining improvements that would be helpful. @lschriml will probably need to address the specifics in this request.

Just a note, that we have recently been discussing how genes are listed in the DO. I also wanted to point out that I have done some work to standardize the use of gene symbols in definitions which I hope will be helpful (e.g. ea3e40ba).

vjenkinsFB commented 12 months ago

Thanks so much! I know these are somewhat long-term ideas, and at FlyBase we're very familiary with having more ideas than we have developer free time, I'm sure it's the same for you. Hopefully all the work you have put in to standardize gene name connections will make it relatively easy to codify w/logic and variables what you've fixed in the text, if you decide to do that in the future. Thanks!