/abricate

:mag_right: :pill: Mass screening of contigs for antimicrobial and virulence genes

Primary LanguagePerlGNU General Public License v2.0GPL-2.0

Build Status License: GPL v2 Don't judge me

ABRicate

Mass screening of contigs for antimicrobial resistance or virulence genes. It comes bundled with multiple databases: NCBI, CARD, ARG-ANNOT, Resfinder, MEGARES, EcOH, PlasmidFinder, Ecoli_VF and VFDB.

Is this the right tool for me?

  1. It only supports contigs, not FASTQ reads
  2. It only detects acquired resistance genes, NOT point mutations
  3. It uses a DNA sequence database, not protein
  4. It needs BLAST+ >= 2.7 and any2fasta to be installed
  5. It's written in Perl 🐫

If you are happy with the above, then please continue! Otherwise consider using Ariba, Resfinder, RGI, SRST2, AMRFinderPlus, etc.

Quick Start

% abricate 6159.fasta
Using database resfinder:  2130 sequences -  Mar 17, 2017
Processing: 6159.fna
Found 3 genes in 6159.fna
#FILE     SEQUENCE     START   END     STRAND GENE     COVERAGE     COVERAGE_MAP     GAPS  %COVERAGE  %IDENTITY  DATABASE  ACCESSION  PRODUCT        RESISTANCE
6159.fna  NC_017338.1  39177   41186   +      mecA_15  1-2010/2010  ===============  0/0   100.00     100.000    ncbi      AB505628   n/a	     FUSIDIC_ACID
6159.fna  NC_017338.1  727191  728356  -      norA_1   1-1166/1167  ===============  0/0   99.91      92.367     ncbi      M97169     n/a            FOSFOMYCIN
6159.fna  NC_017339.1  10150   10995   +      blaZ_32  1-846/846    ===============  0/0   100.00     100.000    ncbi      AP004832   betalactamase  BETA-LACTAM;PENICILLIN

Installation

Brew

If you are using the MacOS Homebrew or LinuxBrew packaging system:

brew install brewsci/bio/abricate
abricate --check
abricate --list

Bioconda

If you use Conda follow the instructions to add the Bioconda channel:

conda install -c conda-forge -c bioconda -c defaults abricate
abricate --check
abricate --list

Source

If you install from source, Abricate has the following package dependencies:

  • any2fasta for sequence file format conversion
  • BLAST+ >2.7.0 for blastn, makeblastdb, blastdbcmd
  • Perl modules: LWP::Simple, Bio::Perl, JSON, Path::Tiny
  • git, unzip, gzip for updating databases

Most of these are easy to install on an Ubuntu-based system:

sudo apt-get install bioperl ncbi-blast+ gzip unzip git \
  libjson-perl libtext-csv-perl libpath-tiny-perl liblwp-protocol-https-perl libwww-perl
git clone https://github.com/tseemann/abricate.git
./abricate/bin/abricate --check
./abricate/bin/abricate --setupdb
./abricate/bin/abricate ./abricate/test/assembly.fa

Input

Abricate takes any sequence file that any2fasta can convert to FASTA files (eg. Genbank, EMBL), and they can be optionally gzip or bzip2 compressed.

abricate assembly.fa 
abricate assembly.fa.gz
abricate assembly.gbk 
abricate assembly.gbk.bz2

It can take multiple files at once too:

abricate assembly.*
abricate /mnt/ncbi/bacteria/*.gbk.gz 

Or you can provide it a "file of file names" (a "FOFN"):

% cat test/fofn.txt

assembly.fa
assembly.fa.gz
assembly.gbk
assembly.gbk.bz2

% abricate --fofn test/fofn.txt

It does not accept raw FASTQ reads; please use Ariba or SRTS2 for that.

Output

Abricate produces a tap-separated output file with the following columns:

Column Example Description
FILE Ecoli.fna The filename this hit came from
SEQUENCE contig000324 The sequence in the filename
START 23423 Start coordinate in the sequence
END 24117 End coordinate
STRAND + Strand + or -
GENE tet(M) AMR gene name
COVERAGE 1-1920/1920 What proportion of the gene is in our sequence
COVERAGE_MAP =============== A visual represenation of the hit. ==aligned, .=unaligned, /=has_gaps
GAPS 1/4 Openings / gaps in subject and query - possible psuedogene?
%COVERAGE 100.00% Proportion of gene covered
%IDENTITY 99.95% Proportion of exact nucleotide matches
DATABASE ncbi The database this sequence comes from
ACCESSION NC_009632:49744-50476 The genomic source of the sequence
PRODUCT aminoglycoside O-phosphotransferase APH(3')-IIIa Gene product (if available)
RESISTANCE TETRACYCLINE;FUSIDIC_ACID putative antibiotic resistance phenotype, ;-separated

Caveats

  • Does not find mutational resistance, only acquired genes.
  • Gap reporting incomplete
  • Sometimes two heavily overlapping genes will be reported for the same locus
  • Possible coverage calculation issues

Databases

ABRicate comes with some pre-downloaded databases:

You can check what you have installed with the --list command. This lists the available databases in TSV (or CSV with --csv) and three columns:

% abricate --list

DATABASE       SEQUENCES  DBTYPE  DATE
argannot       1749       nucl    2019-Jul-28
card           2241       nucl    2019-Jul-28
ecoh           597        nucl    2019-Jul-28
ecoli_vf       2701       nucl    2019-Jul-28
megares        6635       nucl    2020-Feb-20
ncbi           4324       nucl    2019-Jul-28
plasmidfinder  263        nucl    2019-Jul-28
resfinder      2434       nucl    2019-Jul-28
vfdb           2597       nucl    2019-Jul-28

The default database is ncbi. You can choose a different database using the --db option:

% abricate --db vfdb --quiet 6159.fa

6159.fna  NC_017338.1  2724620  2726149  aur      1-1530/1530     ===============  0/0    100.00     99.346     vfdb      NP_647375	zinc metalloproteinase aureolysin
6159.fna  NC_017338.1  2766595  2767155  icaR     1-561/561       ===============  0/0    100.00     98.930     vfdb      NP_647402	N-acetylglucosaminyltransferase
6159.fna  NC_017338.1  2767319  2768557  icaA     1-1239/1239     ===============  0/0    100.00     99.677     vfdb      NP_647403	n/a
6159.fna  NC_017338.1  2768521  2768826  icaD     1-306/306       ===============  0/0    100.00     99.020     vfdb      NP_647404	n/a
6159.fna  NC_017338.1  2768823  2769695  icaB     1-873/873       ===============  0/0    100.00     99.542     vfdb      NP_647405	n/a
6159.fna  NC_017338.1  2769682  2770734  icaC     1-1053/1053     ===============  0/0    100.00     98.955     vfdb      NP_647406	n/a
6159.fna  NC_017338.1  2771040  2773085  lip      1-2046/2046     ===============  0/0    100.00     98.778     vfdb      NP_647407	triacylglycerol lipase precursor

Combining reports across samples

ABRicate can combine results into a simple matrix of gene presence/absence. An absent gene is denoted . and a present gene is represented by its '%COVERAGE`. This can be individual abricate reports, or a combined one.

# Run abricate on each .fa file
% abricate 1.fna > 1.tab
% abricate 2.fna > 2.tab

# Combine
% abricate --summary 1.tab 2.tab

#FILE     NUM_FOUND  aac(6')-aph(2'')_1  aadD_1  blaZ_32  blaZ_36  erm(A)_1  mecA_15  norA_1  spc_1  tet(M)_7
1.tab     8          100.00              100.00  .        100.00   100.00    100.00   99.91   100.00  100.00
2.tab     3          .                   .       100.00   .        .         100.00   99.91   .       .

Or if you ran everything in a single report, it will work too.

% abricate *.fna > results.tab
% abricate --summary results.tab > summary.tab

Updating the databases

# force download of latest version
% abricate-get_db --db ncbi --force

# re-use existing download and just regenerate the database
% abricate-get_db --db ncbi

Making your own database

Let's say you want to make your own database called tinyamr. All you need is a FASTA file of nucleotide sequences, say tinyamr.fa. Ideally the sequence IDs would have the format >DB~~~ID~~~ACC~~~RESISTANCES DESC where DB is tinyamr, ID is the gene name, ACC is an accession number of the sequence source, RESISTANCES is the phenotype(s) to report, and DESC can be any textual description.

% cd /path/to/abricate/db     # this is the --datadir default option
% mkdir tinyamr
% cp /path/to/tinyamr.fa sequences
% head -n 1 sequences
>tinyamr~~~GENE_ID~~~GENE_ACC~~RESISTANCES some description here
% abricate --setupdb
% # or just do this: makeblastdb -in sequences -title tinyamr -dbtype nucl -hash_index

% abricate --list
DATABASE  SEQUENCES  DBTYPE  DATE
tinyamr   173        nucl    2019-Aug-28

% abricate --db tinyamr screen_this.fasta

Etymology

The name "ABRicate" was chosen as the first 3 letters are a common acronym for "Anti-Biotic Resistance". It also has the form of an English verb, which suggests the tool actual taking "action" against the problem of antibiotic resistance. It is also relatively unique in Google, and is unlikely to receive an infamous JABBA Award.

Citation

If you publish the results of Abricate please cite both the software and the appropriate database you used with --db

Issues

Please report problems to the Issues Page.

License

GPLv2

Author

Torsten Seemann | @torstenseemann | blog