Ontology Annotator - The GO annotation interface
Mapping of GO OA fields to Postgres table names
- OA field -- Postgres Table Name
- Paper -- gop_paper
- WBGene -- gop_wbgene
- Project -- gop_project
- Curator -- gop_curator
- GO -- gop_goontology
- GO Term -- gop_goid
- EC -- gop_goinference
- Accession Number -- gop_accession
- With WBGene -- gop_with_wbgene
- With WB Variation -- gop_with_wbvariation
- With Other -- gop_with
- Qualifier -- gop_qualifier
- Xref to -- gop_xrefto
- Object -- gop_dbtype
- Gene Product -- gop_protein
- Comment -- gop_comment
- Last Updated -- gop_lastupdate
- pgid -- not a table, column used by all tables
Requests for interface change
Change requested: October 2010
Addition of new 'Accession Number' field to the GO interface: This field will be used to enter accession numbers from external databases.
- Name of field: Accession Number
- Position: Between 'EC' and 'With WBGene'
- Will be a multi-value drop-down (works similar to the 'Project field')
Do you want a multidropdown, or a multiontology ? I've made it a multidropdown, but it doesn't have all the functionality since we're not set on what it will be. The Project field is a multiontology field, but we could also make that a multidropdown if you want.
We can leave it as is, though I thought it was a multiontology dropdown.
They're different things. Multiontology fields have autocomplete and nowhere to click to see all values. Multidropdown let you click to see all values but don't show term Info. If you want me to change it, it'd easier to do now. Right now it's multiontology (autocomplete).
The multiontology means that it autocompletes (nothing to click for dropdown list), and you can see Term Info about it ; also we need postgres tables for it which we can update with a cronjob. A multidropdown lets you click the V button to see the possible values (instead of having to type) but you don't get to see Term Info ; also the values are hardcoded in. Let me know which you'd like, how often the values would change.
I think we should leave it as a multi-dropdown. My reasoning is, this is not like say, a GO term where you know the term, like 'reproduction' and then the multiontology completes it for you. These are anonymous accessions and you are not going to remember them, so better to have a multi-dropdown that shows you what they are and the choices you have. Maybe its not good reasoning, but it seems like making it a multiontology is bit of a overkill? Also I don't think the accessions are going to change that much.
Sounds good, so it seems okay as is ? Make extra sure that it works like you'd expect it to, since it's changed how multidropdowns work in general, which affect transgene and antibody (I think)
Yes, lets stick to what we decided--leave the Project field as is, and the new accession field is a multi-dropdown.
- Values are (Detailed explanations are here: http://www.geneontology.org/cgi-bin/references.cgi):
- GO_REF:0000011 Hidden Markov Models
- GO_REf:0000012 Pairwise alignment
- GO_REF:0000015 Use of the ND evidence code for GO terms
- GO_REF:0000033 Annotation inferences using phylogenetic trees
- GO_REF:0000024 Curator sequence analysis for ISS
Added those 5 values, didn't add all values from the GO site. Also, if you want the info for each entry in the Term Info, let me know, and it would have to be multiontology.
Thats fine. The link for the GO site was for documentation purposes.
- values can show in the drop-down as:
'Hidden Markov Models (GO_REF:0000011)' if we want to conform to display of other values, eg, GO Term.
Done. but will change if we make it multiontology.
As of now we don't need to work on modifying the dumping script, but would like to start recording this data during curation for storing in Postgres. Will be dumping in the very near future.
Sounds good, but right now it's not recording the data in postgres, until we work out which of the two types we want. All changes are on the sandbox OA.
Okay, will test once all the changes as above are in.
Sorry, should have updated this after changing it later that night. It should work now.
Have created postgres tables for accession data in sandbox, need to do in live. /home/postgres/work/pgpopulation/go/go_curation/20101011_gop_accession/make_tables_gop_accession
made code live
Back to Gene Ontology