llbase.cellbase-core.5.8.4.source-code.configuration.yml Maven / Gradle / Ivy
version: "${CELLBASE.VERSION}"
apiVersion: "${project.version}"
wiki: https://github.com/opencb/cellbase/wiki
maintenanceFlagFile: "/tmp/maintenance"
maintainerContact: [email protected]
logLevel: "INFO"
logDir: "./logs"
# where to output the logs
# can be "console" or "file", defaults to console
logOutput: "file"
# For testing
secretKey: "xPacig89igHSieEnveJEi4KCfdEslhmssC3vui1JJQGgDQ0y8v"
databases:
mongodb:
host: "${CELLBASE.DB.MONGODB.HOST}"
user: "${CELLBASE.DB.USER}"
password: "${CELLBASE.DB.PASSWORD}"
shards:
- id: "${CELLBASE.DB.MONGODB.REPLICASET.0.NAME}"
nodes: "${CELLBASE.DB.MONGODB.REPLICASET.0}"
- id: "${CELLBASE.DB.MONGODB.REPLICASET.1.NAME}"
nodes: "${CELLBASE.DB.MONGODB.REPLICASET.1}"
options:
authenticationDatabase: "${CELLBASE.DB.MONGODB.AUTHENTICATIONDATABASE}"
authenticationMechanism: "${CELLBASE.DB.MONGODB.AUTHENTICATION_MECHANISM}"
readPreference: "${CELLBASE.DB.MONGODB.READPREFERENCE}"
replicaSet: "${CELLBASE.DB.MONGODB.REPLICASET}"
connectionsPerHost: 20
sslEnabled: false
# sslInvalidCertificatesAllowed: true
# sslInvalidHostnameAllowed: true
enableSharding: false
neo4j:
hsapiens:
host: "${CELLBASE.DB.NEO4J.HOST}"
user: "${CELLBASE.DB.USER}"
password: "${CELLBASE.DB.PASSWORD}"
mmusculus:
host: "${CELLBASE.DB.NEO4J.HOST}"
user: "${CELLBASE.DB.USER}"
password: "${CELLBASE.DB.PASSWORD}"
server:
rest:
port: "${CELLBASE.SERVER.REST.PORT}"
defaultOutdir: "/tmp"
download:
ensembl:
database:
host: ensembldb.ensembl.org:3306
user: anonymous
password: ''
libs: "${CELLBASE.ENSEMBL.LIBS}"
url:
host: ftp://ftp.ensembl.org/pub
ensemblGenomes:
database:
host: mysql-eg-publicsql.ebi.ac.uk:4157
user: anonymous
password: ''
libs: "${CELLBASE.ENSEMBL.LIBS}"
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
refSeq:
host: https://ftp.ncbi.nih.gov/refseq/H_sapiens/annotation/GRCh38_latest/refseq_identifiers/GRCh38_latest_genomic.gtf.gz
refSeqFasta:
host: https://ftp.ncbi.nih.gov/refseq/H_sapiens/annotation/GRCh38_latest/refseq_identifiers/GRCh38_latest_genomic.fna.gz
refSeqProteinFasta:
host: https://ftp.ncbi.nih.gov/refseq/H_sapiens/annotation/GRCh38_latest/refseq_identifiers/GRCh38_latest_protein.faa.gz
refSeqCdna:
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
lrg:
host: http://ftp.ebi.ac.uk/pub/databases/lrgex/list_LRGs_transcripts_xrefs.txt
geneUniprotXref:
host: http://ftp.uniprot.org/pub/databases/uniprot/current_release/knowledgebase/idmapping/by_organism/
geneExpressionAtlas:
host: ftp://ftp.ebi.ac.uk/pub/databases/microarray/data/gxa/allgenes_updown_in_organism_part_2.0.14.tab.gz
mirbase:
host: ftp://mirbase.org/pub/mirbase/CURRENT/miRNA.xls.gz
mirbaseReadme:
host: ftp://mirbase.org/pub/mirbase/CURRENT/README
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
uniprot:
host: ftp://ftp.uniprot.org/pub/databases/uniprot/current_release/knowledgebase/complete/uniprot_sprot.xml.gz
uniprotRelNotes:
host: ftp://ftp.uniprot.org/pub/databases/uniprot/relnotes.txt
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
interproRelNotes:
host: ftp://ftp.ebi.ac.uk/pub/databases/interpro/current/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
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/RCV_xml_old_format/ClinVarFullRelease_2024-05.xml.gz
version: 2024-05
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/VCV_xml_old_format/ClinVarVariationRelease_2024-05.xml.gz
version: 2024-05
clinvarSummary:
host: http://ftp.ncbi.nlm.nih.gov/pub/clinvar/tab_delimited/variant_summary.txt.gz
clinvarVariationAllele:
host: http://ftp.ncbi.nlm.nih.gov/pub/clinvar/tab_delimited/variation_allele.txt.gz
clinvarEfoTerms:
host: ftp://ftp.ebi.ac.uk/pub/databases/eva/ClinVar/2015/ClinVar_Traits_EFO_Names_260615.csv
dbSNP:
host: https://ftp.ncbi.nih.gov/snp/latest_release/VCF/GCF_000001405.40.gz
version: "156"
iarctp53:
host: http://p53.iarc.fr/ajax/Zipper.ashx
docm:
host: http://docm.info/api/
docmVersion:
host: http://docm.info
dgv:
host: http://dgv.tcag.ca/v106/docs
simpleRepeats:
host: http://hgdownload.cse.ucsc.edu/goldenPath
windowMasker:
host: http://hgdownload.cse.ucsc.edu/goldenPath
genomicSuperDups:
host: http://hgdownload.cse.ucsc.edu/goldenPath
gwasCatalog:
#host: http://resources.opencb.org/opencb/cellbase/data/gwas/gwas_catalog_v1.0.2-associations_e106_r2022-05-17.tsv
host: "https://ftp.ebi.ac.uk/pub/databases/gwas/releases/2024/05/20/gwas-catalog-associations_ontology-annotated.tsv"
#version: "1.0.2 associations_e106_r2022-05-17"
version: "2024-05-20"
hpo:
host: https://ci.monarchinitiative.org/view/hpo/job/hpo.annotations/lastSuccessfulBuild/artifact/rare-diseases/util/annotation/phenotype_to_genes.txt
disgenet:
host: https://www.disgenet.org/static/disgenet_ap1/files/downloads
files:
- all_gene_disease_associations.tsv.gz
- readme.txt
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
reactome:
host: http://www.reactome.org/download/current/biopax.zip
gnomadConstraints:
host: https://storage.googleapis.com/gcp-public-data--gnomad/release/2.1.1/constraint/gnomad.v2.1.1.lof_metrics.by_transcript.txt.bgz
version: 2.1.1
hpoObo:
host: http://purl.obolibrary.org/obo/hp.obo
goObo:
host: http://purl.obolibrary.org/obo/go/go-basic.obo
doidObo:
host: http://purl.obolibrary.org/obo/doid.obo
goAnnotation:
host: http://geneontology.org/gene-associations/goa_human.gaf.gz
revel:
host: https://zenodo.org/record/7072866/files/revel-v1.3_all_chromosomes.zip
pubmed:
host: https://ftp.ncbi.nlm.nih.gov/pubmed/baseline/
files:
- pubmed22n[1..1114..4].xml.gz
pharmGKB:
host: https://www.pharmgkb.org/downloads
version: v1
files:
- https://api.pharmgkb.org/v1/download/file/data/genes.zip
- https://api.pharmgkb.org/v1/download/file/data/chemicals.zip
- https://api.pharmgkb.org/v1/download/file/data/variants.zip
- https://api.pharmgkb.org/v1/download/file/data/guidelineAnnotations.json.zip
- https://api.pharmgkb.org/v1/download/file/data/variantAnnotations.zip
- https://api.pharmgkb.org/v1/download/file/data/clinicalAnnotations.zip
- https://api.pharmgkb.org/v1/download/file/data/clinicalVariants.zip
- https://api.pharmgkb.org/v1/download/file/data/drugLabels.zip
- https://api.pharmgkb.org/v1/download/file/data/relationships.zip
species:
vertebrates:
- id: hsapiens
scientificName: Homo sapiens
assemblies:
- ensemblVersion: '104_38'
name: GRCh38
- ensemblVersion: '82_37'
name: GRCh37
data:
- clinical_variants
- conservation
- gene
- genome
- missense_variation_functional_score
- ontology
- protein
- refseq
- regulation
- repeats
- variation
- variation_functional_score
- splice_score
shards:
- collection: "variation"
key:
- chromosome
- start
- end
rangeKey: "chromosome"
zones:
- name: "zone0"
shardRanges:
- minimum: "1"
maximum: "10"
- minimum: "2"
maximum: "20"
- minimum: "3"
maximum: "9"
- name: "zone1"
shardRanges:
- minimum: "10"
maximum: "2"
- minimum: "20"
maximum: "3"
- minimum: "9"
maximum: "Z"
- id: mmusculus
scientificName: Mus musculus
assemblies:
- ensemblVersion: '82_38'
name: GRCm38
data:
- genome
- genome_info
- gene
- variation
- regulation
- protein
- conservation
- id: drerio
scientificName: Danio rerio
assemblies:
- ensemblVersion: '82_10'
name: GRCz10
data:
- genome
- genome_info
- gene
- variation
- protein
- id: rnorvegicus
scientificName: Rattus norvegicus
assemblies:
- ensemblVersion: '82_6'
name: Rnor_6.0
data:
- genome
- genome_info
- gene
- variation
- protein
- id: ptroglodytes
scientificName: Pan troglodytes
assemblies:
- ensemblVersion: '82_214'
name: CHIMP2.1.4
data:
- genome
- genome_info
- gene
- variation
- id: ggorilla
scientificName: Gorilla gorilla
assemblies:
- ensemblVersion: '82_31'
name: gorGor3.1
data:
- genome
- genome_info
- gene
- id: csabaeus
scientificName: Chlorocebus sabaeus
assemblies:
- ensemblVersion: '82_1'
name: ChlSab1.1
data:
- genome
- genome_info
- gene
- id: sscrofa
scientificName: Sus scrofa
assemblies:
- ensemblVersion: '82_102'
name: Sscrofa10.2
data:
- genome
- genome_info
- gene
- variation
- protein
- id: cfamiliaris
scientificName: Canis familiaris
assemblies:
- ensemblVersion: '82_31'
name: CanFam3.1
data:
- genome
- genome_info
- gene
- variation
metazoa:
- id: celegans
scientificName: Caenorhabditis elegans
assemblies:
- ensemblVersion: '29_82_245'
name: WBcel235
data:
- genome
- genome_info
- gene
- variation
- protein
- id: dmelanogaster
scientificName: Drosophila melanogaster
assemblies:
- ensemblVersion: '29_82_6'
name: BDGP6
data:
- genome
- genome_info
- gene
- variation
- protein
- id: agambiae
scientificName: Anopheles gambiae
assemblies:
- ensemblVersion: '29_82_4'
name: AgamP4
data:
- genome
- genome_info
- gene
- variation
- protein
fungi:
- id: scerevisiae
scientificName: Saccharomyces cerevisiae
assemblies:
- ensemblVersion: '29_82_4'
name: R64-1-1
data:
- genome
- genome_info
- gene
- variation
- protein
- id: afumigatus
scientificName: Aspergillus fumigatus
assemblies:
- ensemblVersion: '29_82_2'
name: CADRE
data:
- genome
- genome_info
- gene
- id: anidulans
scientificName: Aspergillus nidulans
assemblies:
- ensemblVersion: '29_82_6'
name: ASM1142v1
data:
- genome
- genome_info
- gene
protist:
- id: pfalciparum
scientificName: Plasmodium falciparum
assemblies:
- ensemblVersion: '29_82_3'
name: ASM276v1
data:
- genome
- genome_info
- gene
- variation
- protein
- id: lmajor
scientificName: Leishmania major
assemblies:
- ensemblVersion: '29_82_2'
name: ASM272v2
data:
- genome
- genome_info
- gene
plants:
- id: athaliana
scientificName: Arabidopsis thaliana
assemblies:
- ensemblVersion: '29_82_10'
name: TAIR10
data:
- genome
- genome_info
- gene
- variation
- protein
- id: osativa
scientificName: Oryza sativa
assemblies:
- ensemblVersion: '29_82_7'
name: IRGSP-1.0
data:
- genome
- genome_info
- gene
- variation
- id: gmax
scientificName: Glycine max
assemblies:
- ensemblVersion: '29_82_1'
name: V1.0
data:
- genome
- genome_info
- gene
- id: vvinifera
scientificName: Vitis vinifera
assemblies:
- ensemblVersion: '29_82_3'
name: IGGP_12x
data:
- genome
- genome_info
- gene
- variation
- id: zmays
scientificName: Zea mays
assemblies:
- ensemblVersion: '29_82_6'
name: AGPv3
data:
- genome
- genome_info
- gene
- variation
- id: sbicolor
scientificName: Sorghum bicolor
assemblies:
- ensemblVersion: '29_82_14'
name: Sorbi1
data:
- genome
- genome_info
- gene
- variation
- id: slycopersicum
scientificName: Solanum lycopersicum
assemblies:
- ensemblVersion: '29_82_250'
name: GCA_000188115.2
data:
- genome
- genome_info
- gene
- variation
virus:
- id: evirus
scientificName: Ebola virus
assemblies:
- name: KM034562v1
data:
- genome
- gene
bacteria:
- id: mpneumoniae
scientificName: Mycoplasma pneumoniae M129
assemblies:
- ensemblVersion: '29_82_1'
ensemblCollection: bacteria_0_collection
name: ASM2734v1
data:
- genome
- genome_info
- gene
- protein
- id: ecoli
scientificName: Escherichia coli
assemblies:
- ensemblVersion: '29_82_1'
ensemblCollection: bacteria_91_collection
name: HUSEC2011CHR1
data:
- genome
- genome_info
- gene
- protein
© 2015 - 2024 Weber Informatics LLC | Privacy Policy