2024-09-08 snapshot stats and major changes

[pgaudet]

See also https://github.com/geneontology/go-releases/issues/91 - I had analyzed the changes at the July 14 snapshot, there were many changes that are expected again

https://docs.google.com/spreadsheets/d/16fdSZNLkil4zDVyotdEJEUYKowzrfOLqHjIplAcDzrE/edit?usp=drive_web&ouid=116162935390241629304

[pgaudet]

Major changes: LEGEND 2024-07-14|| 2024-09-08

BY ORGANISM Human: + 82870 ||  + 87655

• +18760 IDA || +19271 >> seems like a lot but the number was similar last time
• +60678 IPI || 61147 >> many new IntAct& ComplexPortal
• -9848 gene products || (same)
• Are these to isoforms?
  • 1466 "-PRO" in col 17
  • 45,000 have '-' (ie, isoforms or UniProt "chains")
  • New RNAs that are now not being converted to 'gene product':
    • premiRNA: +3257 || (same)_
    • lncRNA: +22082 >> main cause is that lncRNA are now properly labelled
    • rRNA: -6965 || (same)

AmiGO versus AmiGO-staging: (did not re-check)

• 707156 human annotations in AmiGO
• 769603 human annotations in AmiGO-staging
• diff is 62,447; not 82,870 as shown in the stats

Source files (did not re-check)

• 2024-06-17 release:
  • source: 705016 lines (note that this includes headers but they are the same)
  • GOC-produced file: 707209 lines
  • 2024-07-14 snapshot:
  • source: 767735 lines
  • GOC-produced file: 769659 lines

So, I am not sure why the stats are inconsistent. But the data in the files looks right relative to the upstream. In conclusion, Human data looks OK

---

Mouse : + 67644 || +89080

• +4768 IDA >> AmiGO versus staging: UniProt + ~ 1000 new mouse IDAs. But there is not 4700 new IDAs || + 5311
• +3595 IPI >> many new IntAct& ComplexPortal || +3636
• +37703 ISO >> pipeline moved to GOC || +37652 (& +35403 ISS) >> Not sure why we have so many more ISS; these are not new, according to the date. Maybe MGI had a redundancy filter?
• -13447 IEA || +7417

Mouse IEAs: AmiGO: 69359

• EC 1115
• InterPro 17944
• UniProtKB-KW 50300
• (all assigned_by MGI)

Staging: 55912

• Ensembl 26172
• UniProt 20516
• InterPro 5180
• GOC 2851
• RHEA 1193

Much better!

Drosophila: + 4427 || +5295

• 1542 IDA || +1646
• 2500 IMP || +2707

• 145 IGI || +160

  BY GROUP 2024-07-14 MGI : -138973 / 179719 | -77.33% HPA** : + 14103 / 24306 | 58.02%

2024-07-14 _GOCentral: +171390 / 4056952 | 4.22% Ensembl : 57340 / 371081 | 15.45% MGI | -138473 / 180219 | -76.84% TreeGrafter | 59619 / 70480 | 84.59% >> only Xenopus laevis HPA | 14103 / 24306 | 58.02% RHEA | 1432 / 10647 | 13.45%

-- | -- | --

[pgaudet]

Major changes:

• GOC now loads isoform data for human, dog, chicken, pig and cow. This results in about 45,000 new annotations for human
• A bug has been fixed with respect to how some non-coding RNA are imported, resulting in about 25,000 additional annotations for lncRNA and pre-miRNAs (that used to be converted to the more generic 'gene product')
• MGI pipeline now ran at GO Central - we now have annotations from Ensembl, and InterPro, EC and RHEA IEAs are now displayed using these sources as 'contributors'
• MGI internal references have been removed from GO products.
• ComplexPortal, IntAct and Human Protein Atlas (HPA) has made a new release, increasing the number of protein-protein interactions (in particular for human, +60678, and for mouse, + 3595), and increasing by 14000 the number of human CC.