Remove underscores in root-level ontology terms

[ukemi]

There is a proposal to remove the underscores from the root-level terms in the ontology. Will this cause any major problems? Hopefully people are using identifiers.

molecular_function -> molecular function biological_process -> biological process cellular_component -> cellular component

https://github.com/geneontology/go-ontology/commit/6fb7e0e8a82ab003832db931fd001496f46d6f5c

[srengel]

Hi David, what is the expected timeframe for this change?

[ukemi]

Depends on how people respond. If people have the strings hard-coded, then we need to give some lead time. It is similar to any other term name change, but these are special since they are the root nodes, so I thought it best to check.

[tberardini]

I was able to dig up the last time this was proposed (June 2009) and a slew of objections at the time. Is it worth pasting that email thread here?

[ukemi]

Yes please. @balhoff

[tberardini]

Date: Thu, 04 Jun 2009 08:31:29 -0400 From: Barry Zeeberg zeebergb@mail.nih.gov To: Midori Harris midori@ebi.ac.uk Subject: Re: [Gofriends] Alert: proposal to remove underscores from root term names

Dear Midori,

I agree the underscores are inconsistent, but that inconsistency should not have been introduced in the first place. Given that it unfortunately was introduced, I think that you should not change it at this late date, as that could lead to a lot of effort, by us and other developers, to change existing software that relies on that format. We have lost most of our software team, and it would add to our problems to have to track down all the potential places for problems (or to have users run into these problems).

So I really hope you can re-consider this action.

Thanks, Barry

Note that the proposed change would affect only the 'name' tag in an OBO stanza, and the 'name' field of the 'term' table in the MySQL database; these fields are changed fairly frequently for non-root terms.

[tberardini]

RGD (back then): From: Stoddard, Alexander Sent: Thursday, June 04, 2009 9:40 AM To: Petri, Victoria; Bakir, Burcu; De Pons, Jeff Cc: Shimoyama, Mary; Patzer, Andrew Subject: RE: [Gofriends] Alert: proposal to remove underscores from rootterm names

A quick check of the code base shows that it does depend on the term name (NOT the stable GO id) to look up the root terms. In the ontology loading code the file "createDBMFiles.pl" uses a search for the terms "biological_process" etc to look up the root.

At least that file will need to be updated or our ontology stuff will break.

I have not checked the website code yet, just the ontology loading code.

Our ontology updates are currently on a manual schedule, not an automatic update, so I can coordinate testing of our ontology / Portal code with the change without the risk of the public site breaking.

Given the timing of the change (I may be the only RGD developer here) and the possibility that it won't happen at all I propose waiting to see if it actually happens. If so, I can then spend time to patch our code and test it.

Thank you, Alex

-----Original Message----- From: Petri, Victoria Sent: Thursday, June 04, 2009 8:53 AM To: Stoddard, Alexander; Bakir, Burcu; De Pons, Jeff Subject: FW: [Gofriends] Alert: proposal to remove underscores from rootterm names

Hello,

I don't know who is going to take care of uploading ontologies, but here is a proposal for changing the root terms of the three GO ontologies from underscore to space (and reaction to it). I know that in the past when such a change was introduced - I believe it was the opposite - we had problems loading the ontologies.

Please let me know and/or reply to this forwarded message.

Thanks,

Victoria

[tberardini]

from:	Benjamin Hitz hitz@genome.stanford.edu
to:	"Petri, Victoria" vpetri@mcw.edu
cc:	GO Friends gofriends@genome.stanford.edu
date:	Jun 4, 2009, 10:07 AM

I object too as I am sure we have this hard coded somewhere. I agree with the sentiment, however. It's a special case - but the root terms ARE special cases (especially since they now essentially mean "unknown")

Ben

On Jun 4, 2009, at 7:47 AM, Petri, Victoria wrote:

Hello,

I'm forwarding the e-mail I got from our developer(s) - on our side the change will create problems and we too are short at the time in available developers. So, I suggest waiting. I agree that it may be inconsistent but we 'lived' with it w/out much if any problems. If the change happens as dated, we will/may be unable to load newer version of the ontologies.

Thanks,

Victoria

Victoria Petri, Ph.D. Research Scientist Rat Genome Database (http://rgd.mcw.edu) Bioinformatics Program Human and Molecular Genetics Center Medical College of Wisconsin 8701 Watertown Plank Road, Milwaukee, WI 53226 (414) 456-7507 Fax (414) 456-6595 vpetri@mcw.edu

[tberardini]

hjd@informatics.jax.org via genome.stanford.edu | Jun 4, 2009, 10:18 AM Harold Drabkin hjd@informatics.jax.org via genome.stanford.edu

Our SE's took a look at a test file and yes it will brake a lot of stuff, so we object to the change also.

hjd

[tberardini]

Mark Wilkinson +1

Perhaps just change the rdf:label if people are finding the underscore objectionable...??

Mark

[balhoff]

Is nine years enough time? :smile:

Interesting to see reactions from last time.

[tberardini]

Sorin Draghici: Jun 4, 2009, 1:53 PM

Midori,

The Onto-Tools should be ok with the proposed changes. I bet that somebody, somewhere did hard code these :-), but it should not be a big deal to fix the source code since we know exactly what to look for.

However, if you decide to go ahead with it, I propose to make this change together with other necessary changes, next time GO needs a major overhaul. The overhead of switching somebody from whatever they are doing to fixing this is not justified for this issue alone.

All the best,

Sorin Draghici

[tberardini]

OK, that's all of the thread. I suspect that some places will still have the same code in place that would still break, but we'll see!

[ukemi]

Thanks @tberardini

[pgaudet]

out of date