configuration: update most of the data source versions

opencb · Dec 27, 2023 · 8ce81c6 · 8ce81c6
1 parent bae792f
commit 8ce81c6
Showing 1 changed file with 25 additions and 11 deletions.
diff --git a/cellbase-core/src/main/resources/configuration.yml b/cellbase-core/src/main/resources/configuration.yml
@@ -62,7 +62,7 @@ download:
     url:
       host: ftp://ftp.ensemblgenomes.org/pub
   hgnc:
-    host: https://ftp.ebi.ac.uk/pub/databases/genenames/hgnc/archive/monthly/tsv/hgnc_complete_set_2022-01-01.txt
+    host: https://ftp.ebi.ac.uk/pub/databases/genenames/hgnc/archive/monthly/tsv/hgnc_complete_set_2023-11-01.txt
   refSeq:
     host: https://ftp.ncbi.nih.gov/refseq/H_sapiens/annotation/GRCh38_latest/refseq_identifiers/GRCh38_latest_genomic.gtf.gz
   refSeqFasta:
@@ -73,12 +73,15 @@ download:
     host: https://ftp.ncbi.nih.gov/refseq/H_sapiens/annotation/GRCh38_latest/refseq_identifiers/GRCh38_latest_rna.fna.gz
   maneSelect:
 #    host: https://ftp.ncbi.nlm.nih.gov/refseq/MANE/MANE_human/release_0.93/MANE.GRCh38.v0.93.summary.txt.gz
-    host: https://ftp.ncbi.nlm.nih.gov/refseq/MANE/MANE_human/release_1.0/MANE.GRCh38.v1.0.summary.txt.gz
-    version: 0.93
+#    host: https://ftp.ncbi.nlm.nih.gov/refseq/MANE/MANE_human/release_1.0/MANE.GRCh38.v1.0.summary.txt.gz
+    host: https://ftp.ncbi.nlm.nih.gov/refseq/MANE/MANE_human/release_1.1/MANE.GRCh38.v1.1.summary.txt.gz
+    version: "1.1"
   lrg:
     host: http://ftp.ebi.ac.uk/pub/databases/lrgex/list_LRGs_transcripts_xrefs.txt
+    version: "2021-03-30"
   geneUniprotXref:
     host: http://ftp.uniprot.org/pub/databases/uniprot/current_release/knowledgebase/idmapping/by_organism/
+    version: "2023-11-08"
   geneExpressionAtlas:
     host: ftp://ftp.ebi.ac.uk/pub/databases/microarray/data/gxa/allgenes_updown_in_organism_part_2.0.14.tab.gz
   mirbase:
@@ -88,33 +91,43 @@ download:
   targetScan:
     host: http://hgdownload.cse.ucsc.edu/goldenPath/
   miRTarBase:
-    host: https://mirtarbase.cuhk.edu.cn/~miRTarBase/miRTarBase_2022/cache/download/8.0/hsa_MTI.xlsx
+    host: https://mirtarbase.cuhk.edu.cn/~miRTarBase/miRTarBase_2022/cache/download/9.0/hsa_MTI.xlsx
+    version: "9.0"
   uniprot:
     host: ftp://ftp.uniprot.org/pub/databases/uniprot/current_release/knowledgebase/complete/uniprot_sprot.xml.gz
+    version: "2023-11-08"
   uniprotRelNotes:
     host: ftp://ftp.uniprot.org/pub/databases/uniprot/relnotes.txt
+    version: "2023-11-08"
   intact:
     host: ftp://ftp.ebi.ac.uk/pub/databases/intact/current/psimitab/intact.txt
   interpro:
-    host: ftp://ftp.ebi.ac.uk/pub/databases/interpro/current/protein2ipr.dat.gz
+#    host: ftp://ftp.ebi.ac.uk/pub/databases/interpro/current/protein2ipr.dat.gz
+    host: https://ftp.ebi.ac.uk/pub/databases/interpro/current_release/protein2ipr.dat.gz
   interproRelNotes:
-    host: ftp://ftp.ebi.ac.uk/pub/databases/interpro/current/release_notes.txt
+#    host: ftp://ftp.ebi.ac.uk/pub/databases/interpro/current/release_notes.txt
+    host: https://ftp.ebi.ac.uk/pub/databases/interpro/current_release/release_notes.txt
   conservation:
     host: https://hgdownload.cse.ucsc.edu/goldenPath/
   gerp:
-    host: http://ftp.ensembl.org/pub/release-104/compara/conservation_scores/90_mammals.gerp_conservation_score/gerp_conservation_scores.homo_sapiens.GRCh38.bw
+    host: http://ftp.ensembl.org/pub/release-110/compara/conservation_scores/91_mammals.gerp_conservation_score/gerp_conservation_scores.homo_sapiens.GRCh38.bw
+    version: "2023-04-22"
   clinvar:
 #    host: https://ftp.ncbi.nlm.nih.gov/pub/clinvar/xml/ClinVarFullRelease_2021-07.xml.gz
 #    host: https://ftp.ncbi.nlm.nih.gov/pub/clinvar/xml/ClinVarFullRelease_2022-02.xml.gz
-    host: https://ftp.ncbi.nlm.nih.gov/pub/clinvar/xml/ClinVarFullRelease_2022-11.xml.gz
+#    host: https://ftp.ncbi.nlm.nih.gov/pub/clinvar/xml/ClinVarFullRelease_2022-11.xml.gz
+    host: https://ftp.ncbi.nlm.nih.gov/pub/clinvar/xml/ClinVarFullRelease_2023-12.xml.gz
   clinvarVariation:
 #    host: https://ftp.ncbi.nlm.nih.gov/pub/clinvar/xml/clinvar_variation/ClinVarVariationRelease_2021-07.xml.gz
 #    host: https://ftp.ncbi.nlm.nih.gov/pub/clinvar/xml/clinvar_variation/ClinVarVariationRelease_2022-02.xml.gz
-    host: https://ftp.ncbi.nlm.nih.gov/pub/clinvar/xml/clinvar_variation/ClinVarVariationRelease_2022-11.xml.gz
+#    host: https://ftp.ncbi.nlm.nih.gov/pub/clinvar/xml/clinvar_variation/ClinVarVariationRelease_2022-11.xml.gz
+    host: https://ftp.ncbi.nlm.nih.gov/pub/clinvar/xml/clinvar_variation/ClinVarVariationRelease_2023-12.xml.gz
   clinvarSummary:
     host: http://ftp.ncbi.nlm.nih.gov/pub/clinvar/tab_delimited/variant_summary.txt.gz
+    version: "2023-12-17"
   clinvarVariationAllele:
     host: http://ftp.ncbi.nlm.nih.gov/pub/clinvar/tab_delimited/variation_allele.txt.gz
+    version: "2023-12-17"
   clinvarEfoTerms:
     host: ftp://ftp.ebi.ac.uk/pub/databases/eva/ClinVar/2015/ClinVar_Traits_EFO_Names_260615.csv
   iarctp53:
@@ -144,7 +157,8 @@ download:
   dgidb:
     host: https://dgidb.org/data/monthly_tsvs/2021-Jan/interactions.tsv
   cadd:
-    host: https://krishna.gs.washington.edu/download/CADD/v1.6/GRCh38/whole_genome_SNVs.tsv.gz
+#    host: https://krishna.gs.washington.edu/download/CADD/v1.6/GRCh38/whole_genome_SNVs.tsv.gz
+    host: https://krishna.gs.washington.edu/download/CADD/v1.7-pre/GRCh38/whole_genome_SNVs.tsv.gz
   reactome:
     host: http://www.reactome.org/download/current/biopax.zip
   gnomadConstraints:
@@ -182,7 +196,7 @@ species:
     - id: hsapiens
       scientificName: Homo sapiens
       assemblies:
-        - ensemblVersion: '104_38'
+        - ensemblVersion: '110_38'
           name: GRCh38
         - ensemblVersion: '82_37'
           name: GRCh37