gorule-0000027 must check that DBs are in the db-xref file and use the 'name' field

[pgaudet]

gorule-0000027 states that all identifiers must be valid, but this is vague.

We will first check

Column 1 & 2: entity

• DB (GAF column 1) value should be in the "database" field db-xref = GORULE:0000027 - TEST 1
• database names or synonyms should be REPAIRED to value should be in the "database" field, and error reported = GORULE:0000027 - TEST 6 ==>> see question below
• id_syntax (GAF column 2) should be checked, if defined in db-xref = GORULE:0000027 - TEST 7 REPORTED AS A WARNING, should be an ERROR, see comment
• Column 1 and column 2 cardinality = 1; pipes are not allowed = GORULE:0000027 - TEST 8 REPORTED AS A ERROR

Column 8 "with" and Column 16 (extensions)

• cardinality 0-n; values are pipe-separated = GORULE_TEST:0000027-9 (PASS), GORULE_TEST:0000027-10 (PASS)
• values that contain DB:ID : same as column 1 & 2:
• DB (GAF column 1) value should be in the "database" field db-xref = GORULE:0000027 - TEST 12, GORULE:0000027 - TEST 13
• id_syntax (GAF column 2) should be checked, if defined in db-xref = GORULE:0000027 - TEST 14

References GAF - column 6, GPAD column 5

• cardinality 1-n; values are pipe-separated GORULE:0000027 - TEST 15 (PASS), GORULE:0000027 - TEST 16
• These have to be in db-xref and syntax has to be checked against syntax_id if possible. GORULE:0000027 - TEST 17**

Assigned_by field (GAF column 15; GPAD column 10)

• checked against groups.yaml = **GORULE:0000027 - TEST 18 (PASS), GORULE:0000027 - TEST 19
  Not failing, see comment

---

LATER

• errors should be FILTERED (to do later, once we are sure this is working correctly)

[kltm]

@pgaudet Looking at this and #1873 (comment), I think that we're pushing the bounds of what I would consider "low-hanging fruit"; there are some things here that require further discussion on the software side. I think we should roll these into a new project that is digging a little deeper in.

[pgaudet]

For id_syntax, @kltm asks that @mugitty does a test for speed, to make sure that this is not too intense computationally.

[pgaudet]

Correctly failing test

GORULE:0000027: SGDDB is not present in groups reference--`UniProtKB Q9HC96 CAPN10 involved_in GO:0006921 PMID:23072806 IDA P GORULE_TEST:0000027-2 Calpain-10 CAPN10,KIAA1845 protein taxon:9606 20140213 SGDDB

[pgaudet]

Need to add more tests to check all points mentioned in initial comment

[pgaudet]

Add test

• Assigned_by field (GAF column 15; GPAD column 10) should be checked against groups.yaml >> GO-OWL is not allowed, for example (currently we have 42 annotations in AmiGO)
• Added more tests for all cases in the initial comment
• Noting that FB:FBrf0193169 is incorrectly picked up as an error by gorule-0000027 (FB refs were recently added this to the dbxef file, this should be fixed once the change is incorporated in @mugitty 's local version for testing)

[mugitty]

@pgaudet , It is working on http://skyhook.berkeleybop.org/go-site-2210-gorule-0000027-id-syntax/reports/assigned-by-gorule-report.html

Update has to be tested on a build that is based on ontobio version 2.9.7 or greater. However, we have reverted to build 2.9.2 on Aug 28, 2024.

[kltm]

@mugitty You can continue with the newer (latest) ontobio. The temporary reversion was to see if we could get a snapshot through to support ontology and GAF (annotation) updates while we work on the ontobio fixes for release.

[mugitty]

Thanks @kltm! The updates for this ticket are already merged into the latest ontobio release.

[pgaudet]

@mugitty

For GORULE-0000027, I have a question about the behavior of the code when encountering this issue: "Database names or synonyms should be REPAIRED to value should be in the "database" field, and error reported (this is tested by GORULE_TEST:0000027-6)

GORULE:0000027: WormBase not found in list of database names in dbxrefs--`WormBase WBGene00000002 aat-1 part_of GO:1990184 WB_REF:WBPaper00006408|PMID:14668347 IPI WB:WBGene00000225 C GORULE_TEST:0000027-6 F27C8.1 gene taxon:6239 20131022 GO_Central

ie, I expect 'WormBase' in Col 1 to be repaired to WB. Usually the report prints the message 'but was repaired'; here is doesn't. What is done, is it indeed repaired?
There are 2 possible cases: either the label used is a synonym of the dbxref, in which case it can be replaced and should give a WARNING, or it's not present in the file, and should be an ERROR.

[pgaudet]

For GORULE-0000027: id_syntax (GAF column 2)** should be checked, if defined in db-xref (this is tested by GORULE_TEST:0000027-7), and reported as a WARNING; it should be an ERROR.

[pgaudet]

For GORULE-0000027: this test is not failing:
! FAILS GORULE:0000027 - TEST 19 - 'Assigned by' is in dbxref.yaml but not in groups.yaml (PO)
UniProtKB Q6UW15 REG3G involved_in GO:0006401 PMID:16931762 IDA P GORULE_TEST:0000027-19 Regenerating islet-derived protein 3-gamma REG3G|PAP1B|UNQ429/PRO162 protein taxon:9606 20220606 PO

[pgaudet]

That last test is now working. Yay!