Taxon-GO Checks and Commentary - Part 5
I have received feedback from David Hill at MGI and from Val Wood at The Pombe database. They have reviewed the gaffe file and I have removed any wrong rules that they found.
At MGI (David Hill):
- I have fixed or flagged 2 incorrect annotations
- The following rules should be rewritten:
- "negative regulation of photoreceptor cell differentiation" only_in NCBITaxon:33317 "Protostomia" This should be relaxed to contain all multicellular organisms.
- "nucleolar ribonuclease P complex" only_in NCBITaxon:2157 "Archaea". This should not be in Archaea, but in eukaryotes.
- "gamma-tubulin ring complex, centrosomal" only_in NCBITaxon:4751 "Fungi". This should also be relaxed to include all eukaryotes.
- I'm not sure what to do with the viral terms. I suspect that they are coming from sequences that are linked up to inactive retroviruses in the mouse genome. They are all coming in via IEA from SwissProt keywords or from Interpro domains. So we can either ignore them, modify the SP keyword/interpro annotation stream or rewrite the taxon rule. I'd be interested in hearing other peoples' thoughts on this and to see if anyone else is seeing this in their organism. I suspect that the human group is seeing it as well.
From Doug Howe: I had a look at the zebrafish lines in the taxon inconsistency file. There were only a very few lines, and they all appeared to related to bad mappings in IEA annotations. Once the bad mappings are removed..the annotations will be fixed when we do our next data load.
Jen: There are 9 annotations in the set he mentions.
From Rama Balakrishnan and Julie Park:
Here are our comments. 39 annotations to the following 5 terms were flagged.
1) GO:0012511 "monolayer-surrounded lipid storage body" only_in NCBITaxon:33090 "Viridiplantae" :: SGD This rule seems correct and we have a bunch of IEAs to this term all of which should be removed/fixed (no annotations with experimental evidence). We will look into posting this on the annotation tracker.
(21 IEA annottiosn removed.)
2) GO:0030543 "2-micrometer plasmid partitioning" only_in JD:0000004 "Prokaryota" :: SGD 3) GO:0030541 "plasmid partitioning" only_in JD:0000004 "Prokaryota" :: SGD
For #2 and #3, the rule is wrong (Midori also wrote in). We have experimental evidence to support our annotations.
(2 rules removed)
4) GO:0016032 "viral reproduction" only_in NCBITaxon:10239 "Viruses" :: SGD
We think this rule is wrong. Viral reproduction can be controlled by host proteins. However, you should consult other groups to come up with rules in this area.
(Checked with Jane Lomax and removed this rule)
5) GO:0005655 "nucleolar ribonuclease P complex" only_in NCBITaxon:2157 "Archaea" :: SGD This rule is wrong. This complex exists in yeast and we have experimental evidence to support our annotations.