You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Track CGD GPI files as per-species dataset metadata
Replace the six GPI stanzas added to cgd.yaml with one dataset file per
species, following the xenbase_xlae / xenbase_xtro precedent.
- Unwrap Proofpoint URL Defense from all six source URLs. Two of the
unwrapped URLs (C_albicans_SC5314_with_B_alleles.gpi, C_auris_B8441.gpi)
404: CGD serves only gzipped GPIs, so all six now point at .gpi.gz.
This also drops the pasted Outlook HTML attributes that had been
captured into two of the source values.
- Give each GPI its own "dataset" name. All six previously shared
"dataset: cgd" with the existing GAF, which collides on the
<dataset>-src.<type> download path and, in
download-annotation-sources.py, lets gpi (rank 2) displace the CGD GAF
(rank 1) outright.
- Drop the "url" fields. They pointed at UniProt-generated GOA mirror
files under annotations/gpi/ (two of which do not exist) rather than
at anything CGD produces. Per the xenbase precedent, per-species GPI
entries carry no url.
- Use go-site 4-character species codes (Calb, Caur, Cdub, Cgla, Cpar,
Ctro) rather than UniProt mnemonics.
cgd.yaml is returned to its state on master; the combined GAF entry is
untouched.
Refs #2725
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
0 commit comments