Skip to content

Refreshed Shared BLAST Databases, Updated Monthly

The shared BLAST databases at /depot/datasets/bioinformatics/blast/latest have been refreshed and are now updated monthly. They are available on all RCAC clusters that mount /depot. latest is a symlink to the current snapshot, week-45-2026, published November 9, 2026.

Quick start

1
2
3
module load biocontainers blast/2.13.0
export BLASTDB=/depot/datasets/bioinformatics/blast/latest
column -t $BLASTDB/MANIFEST.tsv    # databases and their versions

With BLASTDB set, pass a database by name, for example -db core_nt. BLASTDB also lets BLAST find taxdb, so the sscinames and scomnames fields work in -outfmt. The databases are in BLAST v5 format and work with the blast/2.13.0 biocontainer module. For more on loading biocontainer modules, see Running Bioinformatics on RCAC.

Available databases

Type Databases
Nucleotide core_nt, nt, env_nt, refseq_rna, human_genome (GCF_000001405.39), mouse_genome (GCF_000001635.27), mito, Betacoronavirus, 16S_ribosomal_RNA, 18S_fungal_sequences, 28S_fungal_sequences, ITS_RefSeq_Fungi, ITS_eukaryote_sequences, LSU_eukaryote_rRNA, LSU_prokaryote_rRNA, SSU_eukaryote_rRNA, UniVec, UniVec_Core
Protein nr, env_nr, refseq_protein, swissprot (NCBI), pdbaa, landmark, uniprot_sprot, uniref90, uniref50
Taxonomy taxdb
  • core_nt is the recommended general nucleotide database: nt without most eukaryotic chromosome sequences. It holds 132,327,666 sequences (NCBI build October 24, 2026).
  • uniprot_sprot, uniref90, and uniref50 are from UniProt release 2026_04.

Pick a database that matches your program

nr is a protein database, for blastp and blastx. For blastn, use core_nt or nt.

What's new since the 2024 copy

  • core_nt, refseq_rna, refseq_protein, swissprot, and pdbaa
  • taxdb, so taxonomy names appear in tabular output
  • UniProt sets: uniprot_sprot, uniref90, and uniref50
  • UniVec and UniVec_Core
  • Monthly updates with version manifests

Example: taxonomic screening of contigs

This job screens assembled contigs, or any unknown sequences, against core_nt and keeps the top hit for each query.

Important

Run BLAST searches as Slurm jobs on compute nodes, not on login nodes. For job script basics, see HPC Orientation for Biologists.

blast_screen.sh
#!/bin/bash
#SBATCH --job-name=blast_screen
#SBATCH --account=<account-name>
#SBATCH --partition=<partition-name>
#SBATCH --nodes=1
#SBATCH --ntasks=1
#SBATCH --cpus-per-task=32          # illustrative; scale cores to your query size
#SBATCH --time=12:00:00             # illustrative; scale to your query size
#SBATCH --output=%x_%j.out
#SBATCH --error=%x_%j.err

module --force purge
module load biocontainers blast/2.13.0

export BLASTDB=/depot/datasets/bioinformatics/blast/latest
echo "BLAST snapshot: $(readlink ${BLASTDB})"

blastn -task megablast \
    -query contigs.fasta \
    -db core_nt \
    -outfmt "6 qseqid sacc pident length evalue bitscore staxids sscinames" \
    -max_target_seqs 5 \
    -evalue 1e-10 \
    -num_threads ${SLURM_CPUS_PER_TASK} \
    -out contigs_vs_core_nt.tsv

# Keep the top hit for each query
awk '!seen[$1]++' contigs_vs_core_nt.tsv > contigs_top_hit.tsv

Submit it with sbatch blast_screen.sh. To annotate predicted proteins as well, add a blastp search against uniprot_sprot:

1
2
3
4
blastp -query proteins.faa -db uniprot_sprot \
    -outfmt "6 qseqid sacc pident length evalue bitscore stitle" \
    -max_target_seqs 5 -evalue 1e-5 -num_threads ${SLURM_CPUS_PER_TASK} \
    -out proteins_vs_sprot.tsv

Using the ncbi/blast container image

Set APPTAINERENV_BLASTDB inside the ncbi/blast image

The ncbi/blast container image sets its own BLASTDB, which overrides the one you export. When you run that image, set APPTAINERENV_BLASTDB instead:

export APPTAINERENV_BLASTDB=/depot/datasets/bioinformatics/blast/latest

Updates and reproducibility

  • The databases are refreshed monthly, on the first Saturday.
  • Each refresh creates a dated snapshot named week-WW-YYYY. latest is switched to the new snapshot only after every database passes blastdbcheck.
  • The current and previous snapshots are kept; older ones are removed. To keep a snapshot longer, request a hold through a ticket.
  • Each snapshot has a MANIFEST.tsv (database, source, upstream version, and sequence and letter counts) and a SNAPSHOT_INFO file.

For reproducibility, record the snapshot name and the versions of the databases you used in your methods:

readlink /depot/datasets/bioinformatics/blast/latest    # snapshot name, e.g. week-45-2026
column -t /depot/datasets/bioinformatics/blast/latest/MANIFEST.tsv

See Reproducible Genomics for more on recording software and data versions. Other shared datasets are listed under RCAC Datasets.

Getting help

For questions about the databases, or to request a database, submit a ticket through the Purdue Service Portal.

Submit a ticket to RCAC