[InterMine Dev] OBO files and annotation
julie at intermine.org
Wed Jul 25 14:04:48 BST 2018
Maybe you still can use the same parser!
I got that error before, here is how I fixed it:
In FlyMine's FlyAnatomy project entry, I set:
"<property name="ontologyPrefix" value="FBbt"/>"
This parameter tells the parser to ONLY create anatomy terms and not
create any other terms, e.g. GO.
The OBO parser in InterMine can only parse ONE ontology at a time. If
there is more than one ontology in a single file, you get the error you
posted below. Which ontology is loaded is determined in the source's
property file -- the one Joel linked to.
In my case, the fly anatomy OBO file contains anatomy terms AND GO
terms. The parser can't handle this and creates these GO terms as
Does that make sense? Does your OBO file have non-anatomy terms? Maybe
try to add the prefix for your desired ontology term and see if that
fixes the issue.
PS. The "correct" way for InterMine to handle this would be to create
anatomy and go terms as the correct type, but the OBO file does not
contain associated ontology -- we have to set that in the parser's
property file. So that's not possible.
On 25/07/18 13:26, Paulo Nuin wrote:
> Thanks. Our implementation of Anatomy Ontology is basically identical of MouseMine’s DO, which I think you created. I compared both sources and there are probably one or two differences, apart from the different name. You can check it here
> and this works fine. My questions regarding OBO files might not be able to be answered, because InterMine might not provide the solution our curators want.
> My main problem is with queries after loading the Anatomy GAF file, which I am doing with the same loader as GO-Annotation and the error is below. I guess I need to extend and adapt this module to accommodate some differences in our input file. Would that be a correct assumption? My previous message is below for reference.
>> If you notice from my previous email, the Anatomy Annotation GAF file doesn’t annotate genes and/or proteins like the GO-Annotation module. Is it possible to change the module to annotate Anatomy Term elements instead (WBbt: IDs)?
>> I can load the GAF file fine, but not surprisingly I am getting an error when I try searches. I can see on catalina.out this error:
>> query: <query model="genomic" view="AnatomyOntology.synonyms.name AnatomyOntology.identifier AnatomyOntology.name AnatomyOntology.description" sortOrder="AnatomyOntology.synonyms.name ASC" ></query>
>> format: json
>> java.lang.IllegalArgumentException: No simple name for class: AnatomyOntology,GOTerm
>> Our latest code is on the branch im-265 on Github.
>> On Jul 25, 2018, at 1:29 AM, Julie Sullivan <julie at intermine.org> wrote:
>> Joel answered your first question. As for the error, that should be fixed by implementing Joel's suggestion. You are currently creating anatomy terms as GO terms which generates the error.
>> On 23/07/18 20:52, Paulo Nuin wrote:
>>> If you notice from my previous email, the Anatomy Annotation GAF file doesn’t annotate genes and/or proteins like the GO-Annotation module. Is it possible to change the module to annotate Anatomy Term elements instead (WBbt: IDs)?
>>> I can load the GAF file fine, but not surprisingly I am getting an error when I try searches. I can see on catalina.out this error:
>>> query: <query model="genomic" view="AnatomyOntology.synonyms.name AnatomyOntology.identifier AnatomyOntology.name AnatomyOntology.description" sortOrder="AnatomyOntology.synonyms.name ASC" ></query>
>>> format: json
>>> java.lang.IllegalArgumentException: No simple name for class: AnatomyOntology,GOTerm
>>> Our latest code is on the branch im-265 on Github.
>>>> On Jul. 9, 2018, at 11:40 a.m., Paulo Nuin <paulo.nuin at gmail.com> wrote:
>>>> Thanks Julie and Sam.
>>>> Indeed our WB files are GAF2.0 format, so I am guessing just adapt the code for the regular GO Annotation step should be enough.
>>>> !gaf-version: 2.0
>>>> !Project_name: WormBase
>>>> !Contact Email: help at wormbase.org
>>>> WB WBGene00000001 aap-1 Certain WBbt:0003679 WB_REF:WBPaper00005614 IDA WB:Expr2275 A Y110A7A.10 gene taxon:6239 20180416
>>>> WB WBGene00000001 aap-1 Enriched WBbt:0003679 WB_REF:WBPaper00030839 IDA WB:WBPaper00030839:Larval_Pan_Neuronal A Y110A7A.10 gene
>>>> taxon:6239 20180416 WB
>>>> Probably the most important change would be the _keys.properties file. I am assuming the keys in the file are connected to each column in the file, is that correct?
>>>> Thanks again
>>>>> On Jun 29, 2018, at 6:28 AM, Sam Hokin <shokin at ncgr.org> wrote:
>>>>> Hi, Paulo! Just FYI, I do my GO annotations in a post-processor, so in case you decide to do that (load the OBO and WB independently during integrate and then link them up in postprocess) here's an example:
>>>>> I do it this way because I load genes from all over the place and felt that annotation is more of a post-process, linking GO terms listed in the gene descriptions with GO terms. (Yes, funky, but that's how we roll.)
>>>>> On 06/26/2018 10:13 PM, Paulo Nuin wrote:
>>>>>> Hi everyone
>>>>>> We are trying to substitute on XML file that loads anatomy terms by an OBO file connected to an annotation one (extension WB for
>>>>>> us). I was able to easily create code necessary to load the OBO file, but the annotation is turning a little bit trickier.
>>>>>> Documentation is quite slim for this, is there any resource that would help me with this?
>>>>>> Do we need to generate our own class or we can use a generic annotation loader? What is the preferred configuration strategy?
>>>>>> Any help appreciated
>>>>> dev mailing list
>>>>> dev at lists.intermine.org
>>> dev mailing list
>>> dev at lists.intermine.org
More information about the dev