NAME
vsearch \-\-sintax — classify sequences using the SINTAX algorithm
SYNOPSIS
vsearch \-\-sintax fastxfile \-\-db filename \-\-tabbedout filename [\-\-sintax_cutoff real] [options]
DESCRIPTION
The vsearch command --sintax classifies query sequences from a fasta
or fastq file using the SINTAX algorithm (Edgar 2016,
doi:10.1101/074161), a non-Bayesian method for taxonomic classification.
The reference database is specified with --db. Results are written to
--tabbedout.
Classification works by finding the database sequence with the most
shared k-mers with each query across 100 bootstrap resamples of the
query’s k-mers. For each bootstrap replicate, the taxonomy of the
best-matching database sequence is recorded. The frequency of agreement
at each taxonomic rank across the 100 replicates is reported as a
bootstrap confidence value. Use --sintax_cutoff to filter ranks below
a confidence threshold.
The reference database must contain taxonomic annotations in the
sequence headers. Each header must include a ;tax= field followed by a
comma-separated list of taxonomic identifiers. Each identifier starts
with a rank letter (d domain, k kingdom, p phylum, c class, o
order, f family, g genus, s species, t strain), a colon, and the
taxon name. Commas and semicolons are not allowed in taxon names.
Example:
>X80725;tax=d:Bacteria,p:Proteobacteria,c:Gammaproteobacteria,o:Enterobacteriales,f:Enterobacteriaceae,g:Escherichia,s:Escherichia_coli
ACGT...
The --notrunclabels option is on by default for --sintax,
allowing spaces in taxonomic identifiers.
When ties occur between database sequences with equally many k-mer
matches, the shortest (and then earliest) sequence is chosen by default.
The option --sintax_random is strongly recommended instead, as it
breaks ties by a random draw and avoids a bias towards shorter reference
sequences.
For reproducible results with a fixed random seed, use --randseed
together with --threads 1. With multiple threads, sequences may be
processed in varying order across runs, making results non-reproducible
even with a fixed seed.
Both strands can be searched with --strand both. Databases in UDB
format are supported (see
vsearch-udb(5)). This command is
multi-threaded.
OPTIONS
mandatory options
--db filename
Read reference sequences from filename, in fasta, fastq, or UDB format
(see vsearch-udb(5)). Reference
sequence headers must carry taxonomic annotations (see DESCRIPTION).
This option is mandatory.
--tabbedout filename
Write classification results to filename as a tab-separated file. Each
query produces one row with the following columns:
- query label;
- predicted taxonomy with bootstrap confidence in parentheses after
each rank (e.g.,
d:Bacteria(1.00),p:Proteobacteria(0.95)); - strand (
+or-); - predicted taxonomy filtered by
--sintax_cutoff, showing only ranks with bootstrap support at or above the threshold and omitting the values (e.g.,d:Bacteria,p:Proteobacteria). Absent if--sintax_cutoffis not specified.
This option is mandatory.
core options
--dbmask none|dust|soft
Mask regions in the database sequences using the dust method or the
soft method, or none to suppress masking. See
vsearch-fastx_mask(1) for more details.
Warning, when using soft masking, search commands become case
sensitive. The default is to mask using dust.
--randseed integer
Set the seed for the pseudo-random generator. A given seed always
produces the same results, which is useful for replicability. By
default, vsearch uses a pseudo-random seed if --randseed is not set,
or set to the special value zero (0).
--sintax_cutoff real
Set the minimum bootstrap support required for a taxonomic rank to be
included in column 4 of the output (e.g., 0.8 for 80%). Without this
option, column 4 is absent.
--sintax_random
Break ties between database sequences with equally many k-mer matches by
a random draw, rather than preferring the shortest or earliest sequence
in the database. This option is strongly recommended as it avoids a bias
towards shorter reference sequences.
--strand plus|both
Check the plus strand only (default), or check both strands when
comparing sequences.
--threads positive integer
Set the number of computation threads to use, from 1 to 1024. The number
of threads should not exceed the number of available CPU cores. The
default is to use all available cores.
secondary options
--bzip2_decompress
Specify that the input pipe is streaming data compressed using Huffman
coding. See bzip2(1) for more details. This option is not needed when
reading from a regular file compressed with bzip2.
--fastq_ascii 33|64
Specify the offset used as the basis for the fastq quality score when
reading fastq files. For example, an offset of 33 means that a quality
value of 41 is represented by the 74th ASCII symbol (33 + 41 = 74),
which is ‘J’. See ascii(7) for a view of the ASCII character set. The
offset value is either 33 or 64, default is 33.
--fastq_qmax positive integer
Specify the maximal quality score accepted when reading fastq sequences.
Stop with an error message if a quality score higher than the specified
value is read. Accepted values range from 0 to 93 if the offset is 33
(see --fastq_ascii), or range from 0 to 62 if the offset is 64. The
default is 41, which is usual for recent Sanger/Illumina 1.8+ files.
--fastq_qmin positive integer
Specify the minimal quality score accepted when reading fastq sequences.
Stop with an error message if a quality score lower than the specified
value is read. Accepted values range from 0 to 93 if the offset is 33
(see --fastq_ascii), or range from 0 to 62 if the offset is 64. The
default is 0, which is usual for recent Sanger/Illumina 1.8+ files.
Older formats may use scores between -5 and 2.
--gzip_decompress
Specify that the input pipe is streaming data compressed using
Lempel-Ziv coding. See gzip(1) for more details. This option is not
needed when reading from a regular file compressed with gzip.
--label_suffix string
Add the suffix string to sequence headers when writing fasta or fastq
files. For example, with --label_suffix ";status=healthy", sequence
header ‘>seq1’ becomes ‘>seq1;status=healthy’.
--log filename
Write messages to filename. Messages include program version, start
and finish times, elapsed time, amount of memory available, maximum
amount of memory consumed, number of cores and command line options, and
if need be, command-specific informational messages, warnings, and
errors.
--maxseqlength positive integer
Discard database sequences longer than positive integer (50,000
nucleotides by default).
--minseqlength positive integer
Discard database sequences shorter than positive integer (32
nucleotides by default).
--no_progress
Suppress the gradually increasing progress indicator normally written to
the standard error stderr(3).
--notrunclabels
Retain whole sequence headers in output files. By default, vsearch
truncates sequence headers at first space or tabulation. This option
suppresses truncation.
--quiet
Suppress messages to the standard output stdout(3) and standard
error stderr(3), except for warnings and error messages.
--wordlength positive integer
Set the length of words (i.e. k-mers) used for sequence indexing and
comparisons. Valid values range from 3 to 15. The default is 8.
EXAMPLES
Classify query sequences against a taxonomy-annotated reference database, writing all confidence values to the output:
vsearch \
--sintax queries.fasta \
--db reference.fasta \
--tabbedout classification.tsv
Classify and report only ranks with at least 80% bootstrap support
(--sintax_cutoff 0.8), using the recommended tie-breaking option and
searching both strands:
vsearch \
--sintax queries.fasta \
--db reference.fasta \
--sintax_cutoff 0.8 \
--sintax_random \
--strand both \
--tabbedout classification.tsv
Produce reproducible results using a fixed random seed and a single thread:
vsearch \
--sintax queries.fasta \
--db reference.fasta \
--sintax_cutoff 0.8 \
--sintax_random \
--randseed 42 \
--threads 1 \
--tabbedout classification.tsv
SEE ALSO
vsearch-usearch_global(1),
vsearch-makeudb_usearch(1),
vsearch-fasta(5),
vsearch-fastq(5),
vsearch-udb(5)
CITATION
Rognes T, Flouri T, Nichols B, Quince C, Mahé F. (2016) VSEARCH: a versatile open source tool for metagenomics. PeerJ 4:e2584 doi: 10.7717/peerj.2584
REPORTING BUGS
Submit suggestions and bug-reports at https://github.com/torognes/vsearch/issues, send a pull request on https://github.com/torognes/vsearch, or compose a friendly or curmudgeont e-mail to Torbjørn Rognes (torognes@ifi.uio.no).
AVAILABILITY
Source code and binaries are available at https://github.com/torognes/vsearch.
COPYRIGHT
Copyright (C) 2014-2026, Torbjørn Rognes, Frédéric Mahé and Tomás Flouri
All rights reserved.
Contact: Torbjørn Rognes torognes@ifi.uio.no, Department of Informatics, University of Oslo, PO Box 1080 Blindern, NO-0316 Oslo, Norway
This software is dual-licensed and available under a choice of one of two licenses, either under the terms of the GNU General Public License version 3 or the BSD 2-Clause License.
GNU General Public License version 3
This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.
This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.
You should have received a copy of the GNU General Public License along with this program. If not, see http://www.gnu.org/licenses/.
The BSD 2-Clause License
Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:
-
Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.
-
Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.
THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS “AS IS” AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
ACKNOWLEDGMENTS
We would like to thank the authors of the following projects for making their source code available:
- vsearch includes code from Google’s CityHash project by Geoff Pike and Jyrki Alakuijala, providing some excellent hash functions available under a MIT license.
- vsearch includes code derived from Tatusov and Lipman’s DUST program that is in the public domain.
- vsearch includes public domain code written by Alexander Peslyak for the MD5 message digest algorithm.
- vsearch includes public domain code written by Steve Reid and others for the SHA1 message digest algorithm.
- vsearch binaries may include code from the zlib library, copyright Jean-Loup Gailly and Mark Adler.
- vsearch binaries may include code from the bzip2 library, copyright Julian R. Seward.