#blast #bioinformatics #ncbi #dna #similarity-analysis

blutils-proc-adapter

A utility to make it easier to run and analyze Blast results

30 stable releases (5 major)

8.2.0 Mar 21, 2024
7.2.0 Mar 3, 2024
6.3.3 Feb 2, 2024
6.2.0 Sep 12, 2023
3.0.1 Jun 25, 2023

#402 in Science

Download history 27/week @ 2024-01-30 25/week @ 2024-02-06 123/week @ 2024-02-13 567/week @ 2024-02-20 285/week @ 2024-02-27 64/week @ 2024-03-05 493/week @ 2024-03-12 560/week @ 2024-03-19 5/week @ 2024-03-26 92/week @ 2024-04-02

1,156 downloads per month
Used in blutils-cli

Apache-2.0

155KB
3.5K SLoC

BLUTILS

The Blutils tool allow user to run and generate consensus identities of Blast results. Currently the BlastN is available.

Installation

Blutils package could be installed directly from crates.io using cargo:

cargo install blutils-cli

After installed, Blutils could be evoked using the blu command.

blu --help

The output should be close to:

A utility to make it easier to run and analyze Blast results

Usage: blu <COMMAND>

Commands:
  build-db  Build the blast database as a pre-requisite for the blastn command
  blastn    Execute the parallel blast and run consensus algorithm
  check     Check `Blutils` dependencies
  help      Print this message or the help of the given subcommand(s)

Options:
  -h, --help     Print help information
  -V, --version  Print version information

Check dependencies

Optionally you could check OS dependencies before run Blutils. Naturally BLutils depends on Ncbi-Blast+ tool to be installed on the host system to perform parallel blast search. To check if the host OS has these package installed run the Blutils checker for linux systems:

blu check linux

Note: Currently the system check is available only for linux systems and assumes that dependencies could be evoked directly from terminal.

Run Blast with Blutils

The Blutils execution is simple. To check all available options evoke the blast subcommand help:

blu blastn run-with-consensus --help

After inspect available options, simple run Blutils with test data. First download test data from the project github directory:

export INPUT_DIR=https://raw.githubusercontent.com/sgelias/blutils/main/test/mock/input
curl ${INPUT_DIR}/query/query.fna > query.fna
curl ${INPUT_DIR}/query/ref_databases/mock-16S.fna > mock-16S.fna
curl ${INPUT_DIR}/query/ref_databases/mock-16S_taxonomies.tsv > mock-16S_taxonomies.tsv

Then run Blutils:

blu blastn run-with-consensus \
    query.fna \
    mock-16S.fna \
    mock-16S_taxonomies.tsv \
    output \
    -t 6 \
    --taxon bacteria \
    --strategy relaxed \
    -f

Seems the above commands, the output files could be found in output directory which will contains two additional files named blast.out and blutils.consensus.json. The first contains default Blast tabular response and the former, the Blutils response, which will be close to:

[
  {
    "query": "NR025123.135626.Bac",
    "taxon": {
      "rank": "species",
      "identifier": "shewanella-olleyana",
      "percIdentity": 100.0,
      "bitScore": 2695.0,
      "alignLength": 1459,
      "mismatches": 0,
      "gapOpenings": 0,
      "qStart": 1,
      "qEnd": 1459,
      "sStart": 1,
      "sEnd": 1459,
      "eValue": 0.0,
      "taxonomy": "d__bacteria;p__pseudomonadota;c__gammaproteobacteria;o__alteromonadales;f__shewanellaceae;g__shewanella;s__shewanella-olleyana",
      "mutated": true,
      "consensusBeans": null
    }
  },
  {
    "query": "draft-5123",
    "taxon": {
      "rank": "species",
      "identifier": "bacillus-mojavensis-subgroup",
      "percIdentity": 100.0,
      "bitScore": 815.0,
      "alignLength": 441,
      "mismatches": 0,
      "gapOpenings": 0,
      "qStart": 1,
      "qEnd": 441,
      "sStart": 217,
      "sEnd": 657,
      "eValue": 0.0,
      "taxonomy": "d__bacteria;clade__terrabacteria-group;p__bacillota;c__bacilli;o__bacillales;f__taxid-186817;g__bacillus;species-group__bacillus-subtilis-group;species-subgroup__bacillus-mojavensis-subgroup",
      "mutated": true,
      "consensusBeans": [
        {
          "rank": "species",
          "identifier": "bacillus-halotolerans",
          "occurrences": 5
        },
        {
          "rank": "species",
          "identifier": "bacillus-mojavensis",
          "occurrences": 4
        }
      ]
    }
  },
  {
    "query": "INVALID_SEQUENCE",
    "taxon": null
  }
]

Blast execution

Blast execution try to reaches the full available CPU saturation. At the default multithread blast execution mode, the full saturation is not reached. To run Blast through Blutils it is possible. All the steps taken during this process can be seen in the image below.

Parallel Blast

Consensus generation

Different from consensus generations from QIIME 2, the Blutils consensus algorithm performs a data pre-filtering based on Blast results for bit-score and perc-identity, seems the algorithm described in the image below.

Consensus Generation

Next steps

This project contains only basic features to run BlastN and generate consensus identities. Thus, many features should be created, like create the database extractor to get data from official NCBI taxdump results and build FASTA database simultaneously, and others. We welcome new feature suggestions when needed!

Dependencies

~23–37MB
~603K SLoC