Skip to content

Update

The update command downloads MLST schemes and creates a BLAST database. This is the main command for keeping your mlst databases current.

Basic Usage

mlstdb update

By default, this uses the built-in curated list of ~300 MLST schemes from both PubMLST and Pasteur.

Options

Option Description Default
--input, -i Path to custom scheme list file Built-in mlst_schemes_all.tab
--directory, -d Output directory for scheme data pubmlst
--blast-directory, -b Output directory for BLAST database blast
--no-auth Skip OAuth; use unauthenticated access Off
--resume, -r Skip already-downloaded schemes Off
--threads, -t Parallel download threads (max recommended: 4) 1
--verbose, -v Show detailed debug output Off
-h, --help Show help message

Examples

Download all curated schemes

mlstdb update

Uses the built-in curated scheme list. Requires prior setup with mlstdb connect for both PubMLST and Pasteur.

Resume an interrupted download

mlstdb update --resume

Skips schemes whose profile file already exists in the output directory.

Speed up with parallel downloads

mlstdb update --threads 4

Downloads up to 4 schemes simultaneously. Keep this at 4 or below to avoid overwhelming the API servers.

Custom output directories

mlstdb update --directory my_schemes --blast-directory my_blast

Use a custom scheme list

mlstdb update --input my_filtered_schemes.tab

The input file must be tab-delimited with these columns:

database    species scheme_description  scheme  URI
pubmlst Klebsiella pneumoniae   MLST    klebsiella  https://rest.pubmlst.org/db/pubmlst_klebsiella_seqdef/schemes/1
pasteur Bordetella  MLST    bordetella_3    https://bigsdb.pasteur.fr/api/db/pubmlst_bordetella_seqdef/schemes/3

Generating a custom scheme list

Use mlstdb fetch with filters to generate a scheme list, then edit it to keep only the schemes you want. See the Fetch guide for details.

Unauthenticated access

mlstdb update --no-auth

Limited availability

Unauthenticated access only works for data created before 2024-12-31. Newer schemes require authentication. Use mlstdb connect for full access.

What gets downloaded

For each scheme, mlstdb update creates a directory containing:

File Description
<scheme>.txt ST profile definitions (tab-delimited)
<locus>.tfa Allele sequences for each locus (FASTA)
<scheme>_info.json Metadata: source, download date, locus count
database_version.txt Database version number

Profile file format

The profile file maps sequence type (ST) numbers to allele combinations:

ST  gapA    infB    mdh pgi phoE    rpoB    tonB
1   4   4   1   1   7   4   10
2   3   4   1   1   9   4   17

Allele file format

Each .tfa file contains FASTA-formatted allele sequences:

>gapA_1
AACCTGAAGTGGGACGAAGTTGGTGTTGACGTTGTTGCTGAAG...
>gapA_2
AACCTGAAGTGGGACGAAGTTGGTGTTGACGTTGTTGCTGAAG...

BLAST Database

After downloading all schemes, update concatenates the allele sequences and runs makeblastdb to produce the BLAST database:

blast/
├── mlst.fa           # Combined allele sequences
├── mlst.fa.ndb       # BLAST index files
├── mlst.fa.nhr
├── mlst.fa.nin
└── ...

This is the database you pass to mlst via --blastdb blast/mlst.fa.

Handling authentication errors

If some schemes fail to download due to authentication issues, update will:

  1. Report which schemes were skipped
  2. Ask if you want to continue and build the BLAST database with the schemes that succeeded

Common causes:

  • Not registered with a specific database: register via the database's web interface
  • Expired tokens: re-run mlstdb connect to register new credentials

Troubleshooting

"Input file not found"

Run mlstdb connect first, and then mlstdb update. Or provide a valid scheme file with --input.

"No credentials available"

Run mlstdb connect --db pubmlst and/or mlstdb connect --db pasteur to set up authentication.

"No schemes were successfully downloaded"

  • Check your internet connection
  • Verify your credentials are valid: mlstdb connect --db pubmlst
  • Try with --verbose to see detailed error messages

Download is slow

  • Use --threads 4 for parallel downloads
  • Use --resume if you need to restart, it won't re-download completed schemes