SGD import models 20240124

[dustine32]

For geneontology/project-management#34. New batch of 7075 SGD models from Protein2GO-sourced GPAD. Had to do a few things to this GPAD to make it Noctua import-able:

1. Map UniProtKB IDs in columns 1 (DB object), 7 (with/from), and 11 (extensions) to SGD IDs.
2. Map RNAcentral IDs in column 1 to SGD IDs.
3. Map some of the annotation_properties (col 12) contributor-id values to ORCIDs (e.g., contributor-id=GOC:Kara Dolinski -> contributor-id=https://orcid.org/0000-0002-7010-0264).

More details and scripts to come in comment(s) below.

[dustine32]

To clarify the changes we made to the upstream Protein2GO GPAD:

1. Map UniProtKB IDs in columns 1 (DB object), 7 (with/from), and 11 (extensions) to SGD IDs:

python3 scripts/remap_gpad_ids_from_gpi.py -g SGD_gpad4noctua.gpa -i source/gpi.sgd > SGD_gpad4noctua.gpa.remapped

2. Map RNAcentral IDs in column 1 to SGD IDs:

grep -e "^RNAcentral" SGD_gpad4noctua.gpa.remapped > SGD_gpad4noctua.gpa.rnacentral
join -1 2 -2 6 -t "    " -o 2.1,1.3,1.4,1.5,1.6,1.7,1.8,1.9,1.10,1.11,1.12,1.13 <(sed 's/RNAcentral:/RNAcentral:\t/g' SGD_gpad4noctua.gpa.rnacentral | sort -k2) <(sort -t "    " -k6 source/rnacentral_sgd_lkp.tsv) | sed 's/^/SGD:/g' > SGD_gpad4noctua.gpa.rnacentral_fixed
cat <(grep -v -e "^RNAcentral" SGD_gpad4noctua.gpa.remapped | grep -v -e "^UniProtKB") SGD_gpad4noctua.gpa.rnacentral_fixed > SGD_gpad4noctua.gpa.all_fixed_pre_valid

3. Map some of the annotation_properties (col 12) contributor-id values to ORCIDs:

python3 scripts/replace_gpad_properties.py -g SGD_gpad4noctua.gpa.all_fixed_pre_valid -n source/contrib_orcid_lkp.tsv > source/SGD_gpad4noctua.gpa.remapped

The final file output from this process that was then submitted for GO rule validation (to make go_cam_sgd_valid.gpad) and GO-CAM conversion (to make the TTL files in models/) is SGD_gpad4noctua.gpa.remapped.