2026-10-08T02:21:05Z https://api.figshare.com/v2/oai

oai:figshare.com:article/33970168 2026-09-23T03:56:50Z category_24310 category_24361 category_24208 item_type_3 month_year_09_2026

Brucella database for mashID - 2026-09-22 Marc-Olivier Duceppe (11048067) Genomics Microbial genetics Bioinformatics and computational biology not elsewhere classified mashID Mash MinHash Brucella Brucella melitensis Brucella abortus Brucella suis Ochrobactrum brucellosis species identification bacterial identification whole genome sequencing sketch database reference database <p dir="ltr">Mash sketch database of the genus Brucella (NCBI taxon 234), including the species formerly classified as Ochrobactrum, for species-level identification with mashID (https://github.com/duceppemo/mashID).</p><p dir="ltr">Contents</p><p dir="ltr">- brucella_2026-09-22.msh: Mash 2.3 sketches, k-mer size 21, sketch size 10000, 387 reference genomes covering 29 Brucella species (31 MB). MD5: bf3336e18cf0418a3490b60a1b0d9483</p><p dir="ltr">- brucella_2026-09-22.metadata.tsv: one line per reference with accession, organism name, NCBI TaxID, genome length, sketch hash count, source file and original fasta header. mashID reads it automatically when it sits next to the .msh file.</p><p dir="ltr">Construction</p><p dir="ltr">All 1995 GenBank assemblies of taxon 234 available on 2026-09-22 were downloaded with NCBI Datasets (atypical assemblies excluded), binned by species using the organism name of the NCBI assembly report, and dereplicated within each species with Assembly-dereplicator 0.3.2 at a Mash distance of 0.0003. This distance is finer than the 0.001 used for other mashID databases because the classical Brucella species are only 0.2 to 0.5% apart; at 0.001 too few representatives survived to resolve them. Assemblies named only "Brucella sp." were excluded. Organism names and TaxIDs come from the NCBI assembly report. Reproducible with scripts/build_taxon_db.sh in the mashID repository (ASSEMBLY_SOURCE=GenBank, EXCLUDE_SP=1, DEREP_DISTANCE=0.0003, taxon 234).</p><p dir="ltr">Validation</p><p dir="ltr">On 226 GenBank genomes not included in the database (up to 30 per species), mashID reported the NCBI species for 93.4%. B. abortus, B. melitensis and B. suis were 29/30, 29/30 and 28/30; B. canis, B. ovis, B. ceti, B. pinnipedialis, B. microti and B. neotomae were 100%. The remaining discordances match a reference of another species at over 99.95% identity, including a set of assemblies deposited as B. intermedia that are identical to B. ciceri references.</p><p dir="ltr">Scope</p><p dir="ltr">The classical Brucella species are host-adapted lineages of one genomospecies, and B. suis biovars are polyphyletic, so species calls within the classical group are nearest-reference calls; biovars are not resolved. The former Ochrobactrum species are well separated from the classical group and from each other.</p> 2026-09-23T03:56:50Z Dataset Dataset 10.6084/m9.figshare.33970168.v1 https://figshare.com/articles/dataset/Brucella_database_for_mashID_-_2026-09-22/33970168 CC BY 4.0