Skip to content

Commit

Permalink
test: update test data (#222)
Browse files Browse the repository at this point in the history
  • Loading branch information
@jsstevenson
jsstevenson authored Jul 19, 2023
1 parent 59fed53 commit 2839deb
Show file tree
Hide file tree
Showing 7 changed files with 56 additions and 22 deletions.
0 tests/unit/data/etl_data/ensembl_109.gff3 → tests/unit/data/etl_data/ensembl_110.gff3
View file
File renamed without changes.
4 changes: 2 additions & 2 deletions tests/unit/data/etl_data/hgnc_20210810.json
View file
Original file line number Diff line number Diff line change
Expand Up @@ -550,14 +550,14 @@
"refseq_accession": [
"NR_046439"
],
"locus_type": "pseudogene",
"locus_type": "unknown",
"hgnc_id": "HGNC:31102",
"ensembl_gene_id": "ENSG00000274847",
"entrez_id": "727764",
"symbol": "MAFIP",
"date_name_changed": "2012-03-26",
"location": "14 unplaced",
"name": "MAFF interacting protein (pseudogene)",
"name": "MAFF interacting protein",
"date_modified": "2017-09-21",
"uniprot_ids": [
"Q8WZ33"
Expand Down
2 changes: 1 addition & 1 deletion tests/unit/data/etl_data/ncbi_info_20210813.tsv
View file
Original file line number Diff line number Diff line change
Expand Up @@ -17,7 +17,7 @@
9606 30849 PIK3R4 - VPS15|p150 MIM:602610|HGNC:HGNC:8982|Ensembl:ENSG00000196455 3 3q22.1 phosphoinositide-3-kinase regulatory subunit 4 protein-coding PIK3R4 phosphoinositide-3-kinase regulatory subunit 4 O phosphoinositide 3-kinase regulatory subunit 4|PI3-kinase p150 subunit|PI3-kinase regulatory subunit 4|phosphatidylinositol 3-kinase-associated p150|phosphoinositide 3-kinase adaptor protein|phosphoinositide-3-kinase, regulatory subunit 4, p150 20210808 -
9606 731716 PRKRAP1 - - HGNC:HGNC:33447 6 6p21.3 alternate reference locus protein activator of interferon induced protein kinase EIF2AK2 pseudogene 1 pseudo PRKRAP1 protein activator of interferon induced protein kinase EIF2AK2 pseudogene 1 O protein kinase, interferon-inducible double stranded RNA dependent activator pseudogene 1 20210611 -
9606 25782 RAB3GAP2 - MARTS1|RAB3-GAP150|RAB3GAP150|SPG69|WARBM2|p150 MIM:609275|HGNC:HGNC:17168|Ensembl:ENSG00000118873 1 1q41 RAB3 GTPase activating non-catalytic protein subunit 2 protein-coding RAB3GAP2 RAB3 GTPase activating non-catalytic protein subunit 2 O rab3 GTPase-activating protein non-catalytic subunit|RAB3 GTPase activating protein subunit 2 (non-catalytic)|RGAP-iso|rab3 GTPase-activating protein 150 kDa subunit|rab3-GAP p150|rab3-GAP regulatory subunit 20210709 -
9606 293 SLC25A6 - AAC3|ANT|ANT 2|ANT 3|ANT3|ANT3Y MIM:300151|MIM:403000|HGNC:HGNC:10992|Ensembl:ENSG00000169100 X|Y X;Y solute carrier family 25 member 6 protein-coding SLC25A6 solute carrier family 25 member 6 O ADP/ATP translocase 3|ADP,ATP carrier protein 3|ADP,ATP carrier protein, liver|ADP/ATP translocator of liver|adenine nucleotide translocator 3|epididymis secretory sperm binding protein|solute carrier family 25 (mitochondrial carrier; adenine nucleotide translocator), member 6 20210708 -
9606 293 SLC25A6 - AAC3|ANT|ANT 2|ANT 3|ANT3|ANT3Y MIM:300151|MIM:403000|HGNC:HGNC:10992|Ensembl:ENSG00000169100|Ensembl:ENSG00000292334|AllianceGenome:HGNC:10992 X|Y Xp22.33 and Yp11.2 solute carrier family 25 member 6 protein-coding SLC25A6 solute carrier family 25 member 6 O ADP/ATP translocase 3|ADP,ATP carrier protein 3|ADP,ATP carrier protein, liver|ADP/ATP translocator of liver|adenine nucleotide translocator 3|epididymis secretory sperm binding protein|solute carrier family 25 (mitochondrial carrier; adenine nucleotide translocator), member 6 20230718 -
9606 100049159 SPG37 - - MIM:611945 8 8p21.2-q13.3 spastic paraplegia 37 (autosomal dominant) unknown - - - - 20191002 -
9606 10251 SPRY3 - spry-3 MIM:300531|HGNC:HGNC:11271|Ensembl:ENSG00000168939|AllianceGenome:HGNC:11271 X|Y Xq28 and Yq12 sprouty RTK signaling antagonist 3 protein-coding SPRY3 sprouty RTK signaling antagonist 3 O protein sprouty homolog 3|antagonist of FGF signaling|sprouty homolog 3|sprouty3 20220805 -
9606 7637 ZNF84 - HPF2 MIM:618554|HGNC:HGNC:13159|Ensembl:ENSG00000198040|AllianceGenome:HGNC:13159 12 12q24.33|map from Rosati ref via FISH [AFS] zinc finger protein 84 protein-coding ZNF84 zinc finger protein 84 O zinc finger protein 84|zinc finger protein HPF2 20220906 -
Expand Down
2 changes: 1 addition & 1 deletion tests/unit/test_database_and_etl.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -103,7 +103,7 @@ def test_ensembl_etl(test_get_seqrepo, processed_ids, db_fixture, etl_data_path)
shutil.rmtree(e.src_data_dir)

e._sequence_location.get_aliases = _get_aliases
e._data_src = etl_data_path / 'ensembl_109.gff3'
e._data_src = etl_data_path / 'ensembl_110.gff3'
e._add_meta()
e._transform_data()
db_fixture.db.complete_write_transaction()
Expand Down
4 changes: 2 additions & 2 deletions tests/unit/test_ensembl_source.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -304,9 +304,9 @@ def test_meta_info(ensembl):
assert resp.source_meta_.data_license == "custom"
assert resp.source_meta_.data_license_url == \
"https://useast.ensembl.org/info/about/legal/disclaimer.html"
assert resp.source_meta_.version == "109"
assert resp.source_meta_.version == "110"
assert resp.source_meta_.data_url == \
"ftp://ftp.ensembl.org/pub/current_gff3/homo_sapiens/Homo_sapiens.GRCh38.109.gff3.gz" # noqa: E501
"ftp://ftp.ensembl.org/pub/current_gff3/homo_sapiens/Homo_sapiens.GRCh38.110.gff3.gz" # noqa: E501
assert resp.source_meta_.rdp_url is None
assert resp.source_meta_.genome_assemblies == ["GRCh38"]
assert resp.source_meta_.data_license_attributes == {
Expand Down
4 changes: 2 additions & 2 deletions tests/unit/test_hgnc_source.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -345,7 +345,7 @@ def mafip():
"""Create a MAFIP gene fixture."""
params = {
"match_type": MatchType.NO_MATCH,
"label": "MAFF interacting protein (pseudogene)",
"label": "MAFF interacting protein",
"concept_id": "hgnc:31102",
"symbol": "MAFIP",
"location_annotations": ["unplaced", "14"],
Expand Down Expand Up @@ -374,7 +374,7 @@ def mafip():
"pubmed:16549056",
"pubmed:15881666"
],
"gene_type": "pseudogene"
"gene_type": "unknown"
}
return Gene(**params)

Expand Down
62 changes: 48 additions & 14 deletions tests/unit/test_ncbi_source.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -315,32 +315,66 @@ def slc25a6():
"concept_id": "ncbigene:293",
"symbol": "SLC25A6",
"aliases": ["AAC3", "ANT", "ANT 2", "ANT 3", "ANT3", "ANT3Y"],
"xrefs": ["hgnc:10992", "ensembl:ENSG00000169100"],
"xrefs": ["hgnc:10992", "ensembl:ENSG00000169100", "ensembl:ENSG00000292334"],
"previous_symbols": ["ANT3Y"],
"associated_with": ["omim:300151", "omim:403000"],
"symbol_status": None,
"location_annotations": ["X", "Y"],
"location_annotations": [],
"strand": "-",
"locations": [
{
"_id": "ga4gh:VSL.HG0bXHwmZoxZzU2ckz4T6lvxIswXhLQZ",
"_id": "ga4gh:VCL.CfEpX1j-bA7gs8k-BrgRA-7yk5ccHEOG",
"type": "ChromosomeLocation",
"species_id": "taxonomy:9606",
"chr": "X",
"interval": {
"end": {"value": 1392113, "type": "Number"},
"start": {"value": 1386151, "type": "Number"},
"type": "SequenceInterval"
},
"type": "CytobandInterval",
"start": "p22.33",
"end": "p22.33"
}
},
{
"_id": "ga4gh:VCL.TxAQEgJyJOrHy8GLUltsUvmeA_z2FlpA",
"type": "ChromosomeLocation",
"species_id": "taxonomy:9606",
"chr": "Y",
"interval": {
"type": "CytobandInterval",
"start": "p11.2",
"end": "p11.2"
}
},
{
"_id": "ga4gh:VSL.HG0bXHwmZoxZzU2ckz4T6lvxIswXhLQZ",
"type": "SequenceLocation",
"sequence_id": "ga4gh:SQ.w0WZEvgJF0zf_P4yyTzjjv9oW1z61HHP",
"type": "SequenceLocation"
"interval": {
"type": "SequenceInterval",
"start": {
"type": "Number",
"value": 1386151
},
"end": {
"type": "Number",
"value": 1392113
}
}
},
{
"_id": "ga4gh:VSL.1J-MNAWJ9hvZtIM_90lqLbxEt707zL_A",
"interval": {
"end": {"value": 1392113, "type": "Number"},
"start": {"value": 1386151, "type": "Number"},
"type": "SequenceInterval"
},
"type": "SequenceLocation",
"sequence_id": "ga4gh:SQ.8_liLu1aycC0tPQPFmUaGXJLDs5SbPZ5",
"type": "SequenceLocation"
"interval": {
"type": "SequenceInterval",
"start": {
"type": "Number",
"value": 1386151
},
"end": {
"type": "Number",
"value": 1392113
}
}
}
],
"gene_type": "protein-coding"
Expand Down

0 comments on commit 2839deb

Please sign in to comment.