GO Rules: Add example passing/failing data to go rules

[dougli1sqrd]

Now that we have a yaml schema in place for the go rules to add example passing and failing data (https://github.com/geneontology/go-site/issues/956) we can add the data to the go rules. This data can be read by software to perform the comparison test.

[dougli1sqrd]

@pgaudet has created many examples already here: https://drive.google.com/drive/folders/12H6Kz9BsFEhtllx7CMy9OW29FeiC8GwQ

[pgaudet]

File is here https://github.com/geneontology/go-site/blob/master/docs/gorules_test_errors.gaf

[kltm]

@pgaudet @mugitty What test suite is this part of? The file should be in /tests/, not user documentation. If the test suite is not in go-site, that will complicate testing and this should probably be in the appropriate repository.

[mugitty]

@kltm, as the rules are implemented, they will be included as part of the standard suite of tests in the pipeline

[kltm]

@mugitty What pipeline are we talking about here? The main "pipeline" pipeline has no test suite. Is there operating code in go-site/pipeline that is under consideration that is separate from ontobio code?

[pgaudet]

Tests for implemented GORULES:

• [x] GORULE:0000001 - GAF lines are parsed according to GAF 2.2 specifications
• Need more tests
• Need to decide which checks are warnings, which are errors, see https://github.com/geneontology/go-site/issues/2079 - no gene symbol (col3); should lines be filtered because of this?
• [x] GORULE:0000002 - No 'NOT' annotations to 'protein binding ; GO:0005515'
• Needs to be updated - see https://github.com/geneontology/go-site/issues/2079
• All tests are already in the test GAF
• [x] GORULE:0000003 - DEPRECTATED
• [x] GORULE:0000005 - OK
• [x] GORULE:0000006 - IEP and HEP usage is restricted to terms from the Biological Process ontology, except when assigned by GOC
• possible bug: see https://github.com/geneontology/go-site/issues/2081
• [x] GORULE:0000007 - No IPI with 'catalytic activity' or descendants
• [x] GORULE:0000008 - Do not annotate terms
• [x] GORULE:0000011 - ND should not be used for root term annotations
• [ ] GORULE:0000013 - Taxon constraints - approved but cannot be tested since it is not implemented in onto-bio
• [x] GORULE:0000015 - Dual species check
• false positives in http://release.geneontology.org/2023-07-27/reports/assigned-by-UniProt-report.html#gorule-0000015; added a 'should pass test in the test GAF, see https://github.com/geneontology/go-site/issues/2089
• [x] GORULE:0000016 - IC required a with/from value
• [x] GORULE:0000017 - with/from not allowed with IDA
• [x] GORULE:0000018 - with/from mandatory with IPI
• Need to re-run the test errors; data had were not formatted correctly
• [x] GORULE:0000019 - Generic Reasoner Validation Check currently not reported
• [x] GORULE:0000020 - Repair or filter of obsolete terms
• Need to re-run the test errors; 2 new tests for extensions (currently, GO term and 'with' are being tested)
• [x] [GORULE:0000022 - Filter retracted papers
• [x] GORULE:0000026 - IBA annotations should come from the PAINT pipeline
• ok not reported in snapshot/release
• what is the difference between gorule-0000026 and gorule-0000037?
• [x] GORULE:0000027 - Each identifer in GAF is valid
• [x] GORULE:0000028 - GO aspect (column 9) should match GO term
• To rerun: added a test for missing value (test 4)
• [x] GORULE:0000029 - IEAs should be less than one year old
• To rerun: added a test older annotations (should be filtered)
• [x] GORULE:0000030 - Deprecated GOREFs are not allowed
• test not failing, see https://github.com/geneontology/go-site/issues/2093
• see also https://github.com/geneontology/go-site/issues/1423
• [ ] ~GORULE:0000031 - Annotation relations are replaced when not provided by source - not reported; where is this implemented? redundant with gorule-0000059 and gorule-000061?? NOW DEPRECATED~
• [x] GORULE:0000037 - IBA annotations should ONLY be assigned_by GO_Central and have PMID:21873635 as a reference -- isn't this the implementation of GORULE:0000026? Are both active? No error on snapshot; test errors not failing
• [x] GORULE:0000039 - Protein complexes can not be annotated to GO:0032991 (protein-containing complex) or its descendants
• [x] GORULE:0000042 - IKR requires a NOT qualifier
• [x] ~[GORULE:0000043 Check for valid combination of evidence code and GO_REF]~(https://github.com/geneontology/go-site/tree/master/metadata/rules#gorule0000043)
• marked as proposed but being reported in reports, but no results; also need to make tests, see https://github.com/geneontology/go-site/issues/1177 FIXED - not reported anymore
• [x] GORULE:0000046 - The ‘with’ field (GAF column 8) must be the same as the gene product (GAF column 2) when annotating to ‘self-binding’ terms
• rerun TEST2, had format problem
• [x] GORULE:0000050 - ISS , ISA and ISO annotations should not be self-referential
• [x] GORULE:0000051 - Some GO terms require a value in the Annotation Extension field
• [x] GORULE:0000055 - References should have only one ID per ID space
• [x] GORULE:0000057 Group specific filter rules should be applied to annotations -> marked as proposed, present in the REPORTS file, with no errors found anywhere. No tests REMOVED FROM REPORTS - relevant for Noctua load pipeline
• [x] GORULE:0000058 Object extensions should conform to the extensions-patterns.yaml file in metadata -> marked as proposed, present in the REPORTS file, with no errors found anywhere. No tests
• [x] GORULE:0000059 - GAF Version 2.0 and 2.1 are converted into GAF Version 2.2
• close https://github.com/geneontology/go-site/issues/1558 ? CANNOT TEST IN GAF1.2
• [x] GORULE:0000061 - Allowed gene product to term relations (gp2term)
• Some tests not failing see https://github.com/geneontology/go-site/issues/1658#issuecomment-1735449135

[pgaudet]

• GAF tests
• GPAD tests

[pgaudet]

Test summary is here: https://docs.google.com/spreadsheets/d/1rCc2UL1IDDpyqc3Lp-Wuu9onzbWCeySEBfq59o4_0KI/edit?gid=552959064#gid=552959064

GAFs are passing tests GPAD is not yet tested because the pipeline doesn't produce any files that adhere to GPAD2.0 specs. To do when the GPAD output is fixed.