[Go] GO Web Services

[Chris Mungall]

On Apr 14, 2008, at 7:23 AM, Jim Hu wrote:

> Hi Suzi,
>
> I realize this is reviving a discussion from a couple of weeks ago,  
> but now that I'm in a position to want to actually use them:
>
> What web services do we offer from the GOC?
>
> For example, can I query a GOC tool with a UniProt or NCBI  
> identifier and get a list of all the gene associations as an XML  
> output?  I know I can page-scrape the Amigo page, but I don't think  
> that counts as a web service.

There are options in a1.5 to export in formats other than html, so  
you wouldn't need to screen scrape, but even so there are good  
reasons not to use amigo as a web service (the urls you use may be  
messy and unstable for one thing)

>  Is GOOSE a web service as it currently is set up?  Perhaps the  
> download as tab-delimited makes it one, but I'm not sure.

I don't see why not. We could add an xml option to make it more web- 
servicey, but this wouldn't really make sense for GOOSE

Even if we consider GOOSE a web service, it couldn't really be  
considered complete - one would expect something more API-oriented or  
service-oriented for web services.

We could extend AmiGO to offer something more akin to real web services

One problem here is that web services are ideal ways to launch killer  
queries against a database. An option we may consider is to have web  
services run against the sacrificial database server in berkeley.

> The example above is a real one; I want to prepopulate our  
> community annotation tables with what is already out there.  What  
> query do I use if I want all the existing annotations from all  
> sources for one gene?  This strikes me as running into the dreaded  
> problem of matching the gene identifiers across sources.

Indeed. We still don't have NCBI identifiers loaded, so even if we  
had a web service for you, it wouldn't deliver your requirements.

You could in theory compose web services - either manually or via a  
tool like taverna - using one web service for ID resolution and  
another for annotation retrieval. But this is maybe a slightly  
overwrought solution.

What is your time frame for this? I'm guessing yesterday..

If you're willing to be a guinea pig, we have been intending to  
deploy both web services and ID resolution, running of the berkeley  
mirror to begin with..

Did you manage to subscribe to the software group?

> Jim
>
>
> On Mar 30, 2008, at 12:52 PM, Suzanna Lewis wrote:
>> Yep, Jim has it right (including the blurry edges).
>>
>> And I believe that we do offer quite a few programmable ways to  
>> access GO data.
>>
>> -S
>>
>> On Mar 30, 2008, at 10:08 AM, Jim Hu wrote:
>>
>>> Hi Rama,
>>>
>>> I'm not sure what the others are talking about, but formally, web  
>>> services are not the same as web-based tools.
>>> Web services in my usage is stuff where a structured query  
>>> dynamically returns something like an XML file that is readable  
>>> by a computer.  For example, EcoliWiki is a web-based tool, and  
>>> I've been working on implementing a web service that takes a gene  
>>> identifier and returns a structured XML document with the GO  
>>> associations for that gene.  The distinction between web services  
>>> and frequently updated downloads is a bit blurry in my mind.
>>>
>>> Other examples include the E-Utils at PubMed, DAS for gbrowse,  
>>> and the stuff you don't see in the background that updates Amazon  
>>> pages from different publishers.   See
>>>
>>> http://en.wikipedia.org/wiki/Web_service
>>>
>>> Jim
>>>
>>>
>>> On Mar 30, 2008, at 11:49 AM, Rama Balakrishnan wrote:
>>>> Hi,
>>>>
>>>> May be I am missing the obvious here.
>>>>
>>>> We already have a page for GO tools-http://www.geneontology.org/
>>>> GO.tools.shtml
>>>> and this page is organized by what is web-based, downloadable,
>>>> provided by GO etc.
>>>>
>>>> What will the proposed GO web services page include?
>>>>
>>>> Thanks,
>>>>
>>>> Rama
>>>>
>>>>
>>>>
>>>> On Mar 28, 2008, at 5:01 AM, Suzanna Lewis wrote:
>>>>
>>>>> 2 pages perhaps? One for those services that we (already)  
>>>>> offer, and a
>>>>> second for those from others.
>>>>>
>>>>> -S
>>>>>
>>>>> On Mar 28, 2008, at 12:17 PM, Judith Blake wrote:
>>>>>
>>>>>> I think this is an excellent idea...
>>>>>>
>>>>>> I particularly like these services that would be useful to our  
>>>>>> users
>>>>>> but
>>>>>> that don't require that we set up toolshops.
>>>>>>
>>>>>> judy
>>>>>>
>>>>>>
>>>>>> Michael Ashburner wrote:
>>>>>>> Should we have a page on GO for GO web services ?
>>>>>>>
>>>>>>> GoPubMed
>>>>>>> Andreas Schlicker (MPI-INF) functional similarity is another
>>>>>>>
>>>>>>> M
>>>>>>> _______________________________________________
>>>>>>> Go mailing list
>>>>>>> Go at geneontology.org
>>>>>>> http://fafner.stanford.edu/mailman/listinfo/go
>>>>>>>
>>>>>>
>>>>>>
>>>>>> _______________________________________________
>>>>>> Go mailing list
>>>>>> Go at geneontology.org
>>>>>> http://fafner.stanford.edu/mailman/listinfo/go
>>>>>>
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Go mailing list
>>>>> Go at geneontology.org
>>>>> http://fafner.stanford.edu/mailman/listinfo/go
>>>>
>>>> _______________________________________________
>>>> Go mailing list
>>>> Go at geneontology.org
>>>> http://fafner.stanford.edu/mailman/listinfo/go
>>>
>>> =====================================
>>> Jim Hu
>>> Associate Professor
>>> Dept. of Biochemistry and Biophysics
>>> 2128 TAMU
>>> Texas A&M Univ.
>>> College Station, TX 77843-2128
>>> 979-862-4054
>>>
>>>
>>> _______________________________________________
>>> Go mailing list
>>> Go at geneontology.org
>>> http://fafner.stanford.edu/mailman/listinfo/go
>>
>>
>
> =====================================
> Jim Hu
> Associate Professor
> Dept. of Biochemistry and Biophysics
> 2128 TAMU
> Texas A&M Univ.
> College Station, TX 77843-2128
> 979-862-4054
>
>
> _______________________________________________
> Go mailing list
> Go at geneontology.org
> http://fafner.stanford.edu/mailman/listinfo/go