Skip to content

Commit

Permalink
Merge 357461f into 3e82f4b
Browse files Browse the repository at this point in the history
  • Loading branch information
@justaddcoffee
justaddcoffee committed Jul 6, 2020
2 parents 3e82f4b + 357461f commit 2a029d3
Show file tree
Hide file tree
Showing 3 changed files with 80 additions and 68 deletions.
106 changes: 53 additions & 53 deletions curated/ORFs/uniprot_sars-cov-2.gpi
View file
Original file line number Diff line number Diff line change
@@ -1,53 +1,53 @@
!gpi-version: 1.2
!
!This file contains additional information for proteins in the UniProt KnowledgeBase (UniProtKB).
!Protein accessions are represented in this file even if there is no associated GO annotation.
!
!Columns:
!
! name required? cardinality GAF column # Example content
! DB required 1 1 UniProtKB
! DB_Object_ID required 1 2/17 Q4VCS5-1
! DB_Object_Symbol required 1 3 AMOT
! DB_Object_Name optional 0 or greater 10 Angiomotin
! DB_Object_Synonym(s) optional 0 or greater 11 AMOT|KIAA1071
! DB_Object_Type required 1 12 protein
! Taxon required 1 13 taxon:9606
! Parent_Object_ID optional 0 or 1 - UniProtKB:Q4VCS5
! DB_Xref(s) optional 0 or greater - WB:WBGene00000035
! Properties optional 0 or greater - "db_subset=Swiss-Prot|target_set=KRUK,BHFL"
!
!Generated: 2020-06-10 13:25
!
UniProtKB P0DTC1 pp1a Replicase polyprotein 1a ORF1a|1a|pp1a protein taxon:2697049
UniProtKB P0DTC1-PRO_0000449645 nsp11 Non-structural protein 11 nsp11|P0DTC1(4393-4405) protein taxon:2697049 UniProtKB:P0DTC1
UniProtKB P0DTC2 S protein Spike glycoprotein S|2|S protein protein taxon:2697049
UniProtKB P0DTC2-PRO_0000449647 S(13-685) Spike protein S1 P0DTC2(13-685)|S|2 protein taxon:2697049 UniProtKB:P0DTC2
UniProtKB P0DTC2-PRO_0000449648 S(686-1273) Spike protein S2 P0DTC2(686-1273)|S|2 protein taxon:2697049 UniProtKB:P0DTC2
UniProtKB P0DTC2-PRO_0000449649 S(816-1273) Spike protein S2' P0DTC2(816-1273)|S|2 protein taxon:2697049 UniProtKB:P0DTC2
UniProtKB P0DTC3 ORF3a ORF3a protein ORF3a|3a protein taxon:2697049
UniProtKB P0DTC4 E protein Envelope small membrane protein E|4|E protein protein taxon:2697049
UniProtKB P0DTC5 M protein Membrane protein protein taxon:2697049
UniProtKB P0DTC6 ORF6 ORF6 ORF6|accessory protein 6|ns6|X3 protein taxon:2697049
UniProtKB P0DTC7 ORF7a ORF7a protein ORF7a|7a protein taxon:2697049
UniProtKB P0DTC8 ORF8 ORF8 ns8 protein taxon:2697049
UniProtKB P0DTC9 N protein Nucleoprotein NC|Protein N|NPRBD|P0DTC9(1-419)|N protein taxon:2697049
UniProtKB P0DTD1 pp1ab Replicase polyprotein 1ab rep|1a-1b|ORF1ab protein taxon:2697049
UniProtKB P0DTD1-PRO_0000449619 nsp1 Host translation inhibitor nsp1 nsp1|P0DTD1(1-180) protein taxon:2697049 UniProtKB:P0DTD1
UniProtKB P0DTD1-PRO_0000449620 nsp2 Non-structural protein 2 nsp2|P0DTD1(181-818) protein taxon:2697049 UniProtKB:P0DTD1
UniProtKB P0DTD1-PRO_0000449621 nsp3 Non-structural protein 3 nsp3|ADRP|PL-PRO|PL2-PRO|Papain-like proteinase|P0DTD1(819-2763) protein taxon:2697049 UniProtKB:P0DTD1
UniProtKB P0DTD1-PRO_0000449622 nsp4 Non-structural protein 4 nsp4|P0DTD1(2764-3263) protein taxon:2697049 UniProtKB:P0DTD1
UniProtKB P0DTD1-PRO_0000449623 nsp5 3C-like proteinase nsp5|3CL-PRO|3CLPro|3CLp|Mpro|P0DTD1(3264-3569) protein taxon:2697049 UniProtKB:P0DTD1
UniProtKB P0DTD1-PRO_0000449624 nsp6 Non-structural protein 6 nsp6|P0DTD1(3570-3859) protein taxon:2697049 UniProtKB:P0DTD1
UniProtKB P0DTD1-PRO_0000449625 nsp7 Non-structural protein 7 nsp7|P0DTD1(3860-3942) protein taxon:2697049 UniProtKB:P0DTD1
UniProtKB P0DTD1-PRO_0000449626 nsp8 Non-structural protein 8 nsp8|P0DTD1(3943-4140) protein taxon:2697049 UniProtKB:P0DTD1
UniProtKB P0DTD1-PRO_0000449627 nsp9 Non-structural protein 9 nsp9|P0DTD1(4141-4253) protein taxon:2697049 UniProtKB:P0DTD1
UniProtKB P0DTD1-PRO_0000449628 nsp10 Non-structural protein 10 nsp10|GFL|P0DTD1(4254-4392) protein taxon:2697049 UniProtKB:P0DTD1
UniProtKB P0DTD1-PRO_0000449629 nsp12 RNA-directed RNA polymerase nsp12|Pol|RdRp|P0DTD1(4393-5324) protein taxon:2697049 UniProtKB:P0DTD1
UniProtKB P0DTD1-PRO_0000449630 nsp13 Helicase nsp13|helicase|nsp13 helicase|Hel|P0DTD1(5325-5925) protein taxon:2697049 UniProtKB:P0DTD1
UniProtKB P0DTD1-PRO_0000449631 nsp14 Proofreading exoribonuclease nsp14|ExoN|P0DTD1(5926-6452) protein taxon:2697049 UniProtKB:P0DTD1
UniProtKB P0DTD1-PRO_0000449632 nsp15 Uridylate-specific endoribonuclease nsp15|P0DTD1(6453-6798) protein taxon:2697049 UniProtKB:P0DTD1
UniProtKB P0DTD1-PRO_0000449633 nsp16 2'-O-methyltransferase nsp16|P0DTD1(6799-7096) protein taxon:2697049 UniProtKB:P0DTD1
UniProtKB P0DTD2 ORF9b accessory protein 9b ORF9b|ORF-9b protein taxon:2697049
UniProtKB P0DTD3 ORF14 Uncharacterized protein 14 ORF14 protein taxon:2697049
UniProtKB P0DTD8 ORF7b Accessory protein 7b ORF7b|ns7b protein taxon:2697049
!gpi-version: 1.2
!
!This file contains additional information for proteins in the UniProt KnowledgeBase (UniProtKB).
!Protein accessions are represented in this file even if there is no associated GO annotation.
!
!Columns:
!
! name required? cardinality GAF column # Example content
! DB required 1 1 UniProtKB
! DB_Object_ID required 1 2/17 Q4VCS5-1
! DB_Object_Symbol required 1 3 AMOT
! DB_Object_Name optional 0 or greater 10 Angiomotin
! DB_Object_Synonym(s) optional 0 or greater 11 AMOT|KIAA1071
! DB_Object_Type required 1 12 protein
! Taxon required 1 13 taxon:9606
! Parent_Object_ID optional 0 or 1 - UniProtKB:Q4VCS5
! DB_Xref(s) optional 0 or greater - WB:WBGene00000035
! Properties optional 0 or greater - "db_subset=Swiss-Prot|target_set=KRUK,BHFL"
!
!Generated: 2020-06-10 13:25
!
UniProtKB P0DTC1 pp1a Replicase polyprotein 1a ORF1a|1a|pp1a protein taxon:2697049 PR:P0DTC1-1
UniProtKB P0DTC1-PRO_0000449645 nsp11 Non-structural protein 11 nsp11|P0DTC1(4393-4405) protein taxon:2697049 UniProtKB:P0DTC1 PR:000050280
UniProtKB P0DTC2 S protein Spike glycoprotein S|2|S protein protein taxon:2697049 PR:P0DTC2
UniProtKB P0DTC2-PRO_0000449647 S(13-685) Spike protein S1 P0DTC2(13-685)|S|2 protein taxon:2697049 UniProtKB:P0DTC2 PR:000050267
UniProtKB P0DTC2-PRO_0000449648 S(686-1273) Spike protein S2 P0DTC2(686-1273)|S|2 protein taxon:2697049 UniProtKB:P0DTC2 PR:000050268
UniProtKB P0DTC2-PRO_0000449649 S(816-1273) Spike protein S2' P0DTC2(816-1273)|S|2 protein taxon:2697049 UniProtKB:P0DTC2 PR:000050269
UniProtKB P0DTC3 ORF3a ORF3a protein ORF3a|3a protein taxon:2697049 PR:P0DTC3
UniProtKB P0DTC4 E protein Envelope small membrane protein E|4|E protein protein taxon:2697049 PR:P0DTC4
UniProtKB P0DTC5 M protein Membrane protein protein taxon:2697049 PR:P0DTC5
UniProtKB P0DTC6 ORF6 ORF6 ORF6|accessory protein 6|ns6|X3 protein taxon:2697049 PR:P0DTC6
UniProtKB P0DTC7 ORF7a ORF7a protein ORF7a|7a protein taxon:2697049 PR:P0DTC7
UniProtKB P0DTC8 ORF8 ORF8 ns8 protein taxon:2697049 PR:P0DTC8
UniProtKB P0DTC9 N protein Nucleoprotein NC|Protein N|NPRBD|P0DTC9(1-419)|N protein taxon:2697049 PR:P0DTC9
UniProtKB P0DTD1 pp1ab Replicase polyprotein 1ab rep|1a-1b|ORF1ab protein taxon:2697049 PR:P0DTD1-1
UniProtKB P0DTD1-PRO_0000449619 nsp1 Host translation inhibitor nsp1 nsp1|P0DTD1(1-180) protein taxon:2697049 UniProtKB:P0DTD1 PR:000050270
UniProtKB P0DTD1-PRO_0000449620 nsp2 Non-structural protein 2 nsp2|P0DTD1(181-818) protein taxon:2697049 UniProtKB:P0DTD1 PR:000050271
UniProtKB P0DTD1-PRO_0000449621 nsp3 Non-structural protein 3 nsp3|ADRP|PL-PRO|PL2-PRO|Papain-like proteinase|P0DTD1(819-2763) protein taxon:2697049 UniProtKB:P0DTD1 PR:000050272
UniProtKB P0DTD1-PRO_0000449622 nsp4 Non-structural protein 4 nsp4|P0DTD1(2764-3263) protein taxon:2697049 UniProtKB:P0DTD1 PR:000050273
UniProtKB P0DTD1-PRO_0000449623 nsp5 3C-like proteinase nsp5|3CL-PRO|3CLPro|3CLp|Mpro|P0DTD1(3264-3569) protein taxon:2697049 UniProtKB:P0DTD1 PR:000050274
UniProtKB P0DTD1-PRO_0000449624 nsp6 Non-structural protein 6 nsp6|P0DTD1(3570-3859) protein taxon:2697049 UniProtKB:P0DTD1 PR:000050275
UniProtKB P0DTD1-PRO_0000449625 nsp7 Non-structural protein 7 nsp7|P0DTD1(3860-3942) protein taxon:2697049 UniProtKB:P0DTD1 PR:000050276
UniProtKB P0DTD1-PRO_0000449626 nsp8 Non-structural protein 8 nsp8|P0DTD1(3943-4140) protein taxon:2697049 UniProtKB:P0DTD1 PR:000050277
UniProtKB P0DTD1-PRO_0000449627 nsp9 Non-structural protein 9 nsp9|P0DTD1(4141-4253) protein taxon:2697049 UniProtKB:P0DTD1 PR:000050278
UniProtKB P0DTD1-PRO_0000449628 nsp10 Non-structural protein 10 nsp10|GFL|P0DTD1(4254-4392) protein taxon:2697049 UniProtKB:P0DTD1 PR:000050279
UniProtKB P0DTD1-PRO_0000449629 nsp12 RNA-directed RNA polymerase nsp12|Pol|RdRp|P0DTD1(4393-5324) protein taxon:2697049 UniProtKB:P0DTD1 PR:000050284
UniProtKB P0DTD1-PRO_0000449630 nsp13 Helicase nsp13|helicase|nsp13 helicase|Hel|P0DTD1(5325-5925) protein taxon:2697049 UniProtKB:P0DTD1 PR:000050285
UniProtKB P0DTD1-PRO_0000449631 nsp14 Proofreading exoribonuclease nsp14|ExoN|P0DTD1(5926-6452) protein taxon:2697049 UniProtKB:P0DTD1 PR:000050286
UniProtKB P0DTD1-PRO_0000449632 nsp15 Uridylate-specific endoribonuclease nsp15|P0DTD1(6453-6798) protein taxon:2697049 UniProtKB:P0DTD1 PR:000050287
UniProtKB P0DTD1-PRO_0000449633 nsp16 2'-O-methyltransferase nsp16|P0DTD1(6799-7096) protein taxon:2697049 UniProtKB:P0DTD1 PR:000050288
UniProtKB P0DTD2 ORF9b accessory protein 9b ORF9b|ORF-9b protein taxon:2697049 PR:P0DTD2
UniProtKB P0DTD3 ORF14 Uncharacterized protein 14 ORF14 protein taxon:2697049 PR:P0DTD3
UniProtKB P0DTD8 ORF7b Accessory protein 7b ORF7b|ns7b protein taxon:2697049 PR:P0DTD8
10 changes: 7 additions & 3 deletions download.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -75,11 +75,15 @@
local_name: HUMAN_9606_idmapping.dat.gz

#
# SciBite CORD-19 annotations v1.4
# SciBite CORD-19 annotations v1.5
#
-
url: https://media.githubusercontent.com/media/SciBiteLabs/CORD19/master/annotated-CORD-19/1.4/CORD-19_1_4.zip
local_name: CORD-19_1_4.zip
url: https://media.githubusercontent.com/media/SciBiteLabs/CORD19/master/annotated-CORD-19/1.5/pdf_json.zip
local_name: pdf_json.zip

-
url: https://media.githubusercontent.com/media/SciBiteLabs/CORD19/master/annotated-CORD-19/1.5/pmc_json.zip
local_name: pmc_json.zip

# SciBite CORD-19 entity co-occurrences v1.2
-
Expand Down
32 changes: 20 additions & 12 deletions kg_covid_19/transform_utils/scibite_cord/scibite_cord.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -41,49 +41,59 @@ def run(self, data_file: Optional[str] = None) -> None:
Args:
data_file: data file to parse
Should be:
[pdf_json.zip, pmc_json.zip, cv19_scc_1_2.zip]
Returns:
None.
"""
data_files = list()
if not data_file:
data_files.append(os.path.join(self.input_base_dir, "CORD-19_1_4.zip"))
data_files.append(os.path.join(self.input_base_dir, "pdf_json.zip"))
data_files.append(os.path.join(self.input_base_dir, "pmc_json.zip"))

data_files.append(os.path.join(self.input_base_dir, "cv19_scc_1_2.zip"))
else:
data_files.append(data_file)
data_files.extend(data_files)

self.node_header = ['id', 'name', 'category', 'description', 'provided_by']
self.edge_header = ['subject', 'edge_label', 'object', 'relation', 'provided_by']
node_handle = open(self.output_node_file, 'w')
edge_handle = open(self.output_edge_file, 'w')
node_handle.write("\t".join(self.node_header) + "\n")
edge_handle.write("\t".join(self.edge_header) + "\n")
self.parse_annotations(node_handle, edge_handle, data_files[0])
self.parse_annotations(node_handle, edge_handle, data_files[0], data_files[1])

node_handle = open(os.path.join(self.output_dir, "entity_cooccurrence_nodes.tsv"), 'w')
edge_handle = open(os.path.join(self.output_dir, "entity_cooccurrence_edges.tsv"), 'w')
node_handle.write("\t".join(self.node_header) + "\n")
edge_handle.write("\t".join(self.edge_header) + "\n")
self.parse_cooccurrence(node_handle, edge_handle, data_files[1])
self.parse_cooccurrence(node_handle, edge_handle, data_files[2])

def parse_annotations(self, node_handle: Any, edge_handle: Any, data_file: str) -> None:
"""Parse annotations from CORD-19_1_2.zip.
def parse_annotations(self, node_handle: Any, edge_handle: Any,
data_file1: str,
data_file2: str) -> None:
"""Parse annotations from CORD-19_1_5.zip.
Args:
node_handle: File handle for nodes.csv.
edge_handle: File handle for edges.csv.
data_file: Path to CORD-19_1_2.zip.
data_file1: Path to first CORD-19_1_5.zip.
data_file2: Path to second CORD-19_1_5.zip.
Returns:
None.
"""
with ZipFile(data_file, 'r') as ZF:
with ZipFile(data_file1, 'r') as ZF:
ZF.extractall(path=self.input_base_dir)
with ZipFile(data_file2, 'r') as ZF:
ZF.extractall(path=self.input_base_dir)

subsets = ['biorxiv_medrxiv', 'comm_use_subset', 'noncomm_use_subset', 'custom_license']
subsets = ['pmc_json', 'pdf_json']
for subset in subsets:
subset_dir = os.path.join(self.input_base_dir, 'CORD19', subset, subset)
subset_dir = os.path.join(self.input_base_dir, subset)
for data_dir in os.listdir(subset_dir):
if os.path.isdir(os.path.join(subset_dir, data_dir)):
for filename in os.listdir(os.path.join(subset_dir, data_dir)):
Expand Down Expand Up @@ -126,8 +136,6 @@ def parse_annotation_doc(self, node_handle, edge_handle, doc: Dict, subset: str
terms.update(self.extract_termite_hits(x))

provided_by = f"{self.source_name}"
if subset:
provided_by += f" {subset}"

# add a biolink:Publication for each paper
write_node_edge_item(
Expand Down

0 comments on commit 2a029d3

Please sign in to comment.