NAME
vsearch \-\-search_exact — search for exact full-length matches against a database
SYNOPSIS
vsearch \-\-search_exact fastxfile \-\-db filename (\-\-alnout | \-\-biomout | \-\-blast6out | \-\-fastapairs | \-\-matched | \-\-mothur_shared_out | \-\-notmatched | \-\-otutabout | \-\-samout | \-\-uc | \-\-userout) filename [options]
DESCRIPTION
The vsearch command --search_exact searches the query sequences in a
fasta or fastq file against a database of target sequences (--db),
reporting only 100% exact full-length matches. It is much faster than
--usearch_global for this use case.
Because only exact matches are reported, --id, --maxaccepts, and
--maxrejects do not apply and are not accepted. Masking is controlled
with --qmask (queries) and --dbmask (database). By default only the
plus strand is searched; use --strand both to also check the reverse
complement.
At least one output option must be specified. This command is multi-threaded.
The --qsegout and --tsegout options do not count as sole output
options: when either is used on its own, vsearch rejects the run with
“No output files specified”. They can still be used, but only in
combination with one of the output options listed in the SYNOPSIS.
OPTIONS
mandatory options
--db must be specified, along with at least one output option.
--db filename
Search query sequences against the target sequences in filename, in
fasta or fastq format. 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.
--qmask none|dust|soft
Mask regions in query sequences using the dust method or the soft
method, or none to suppress masking. Values are case-insensitive, so
DUST, Dust, and dust are all accepted. 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.
--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
--alnout filename
Write pairwise global alignments to filename in a human-readable
format. Use --rowlen to set the alignment line width.
--biomout filename
Write an OTU table to filename in the biom version 1.0 JSON file
format. The OTUs are represented by the cluster centroids. Sample
identifiers are extracted from sequence headers (;sample=abc123; or
;barcodelabel=abc123; patterns, or the initial part of the header).
OTU identifiers are extracted from centroid headers (;otu=def789;
pattern, or the initial part of the header, or via relabelling options).
Taxonomy information is extracted from centroid headers (;tax=...;
pattern) if available.
--blast6out filename
Write results to filename using a BLAST-like tab-separated format with
twelve fields per query-target match: query label, target label,
percentage identity, alignment length, mismatches, gap openings, query
start, query end, target start, target end, expectation value (always
-1), and bit score (always 0). If --output_no_hits is used,
non-matching queries are also written. Note that vsearch uses global
pairwise alignments, not BLAST’s seed-and-extend algorithm.
--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.
--dbmatched filename
Write database target sequences that match at least one query sequence
to filename, in fasta format. If --sizeout is specified, the number
of matching queries is annotated in the header (;size=integer).
--dbnotmatched filename
Write database target sequences that do not match any query sequence to
filename, in fasta format.
--fasta_width positive integer
Set the maximal width of sequences when writing fasta files. Longer
sequences are folded and written on several lines. Default width is 80
nucleotides. Set to zero (0) to suppress folding.
--fastapairs filename
Write pairwise alignments of query and target sequences to filename,
in fasta format.
--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.
--hardmask
Replace masked nucleotides with Ns.
--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’.
--lca_cutoff real
Set the fraction of hits that must agree at a given taxonomic rank for
that rank to be included in the last common ancestor (LCA) output (see
--lcaout). The default value is 1.0, requiring all hits to agree.
Lower values (e.g., 0.95) tolerate a small fraction of disagreeing hits.
The value must be greater than 0.5 and at most 1.0.
--lcaout filename
Write last common ancestor (LCA) information to filename in a
tab-separated format. The first column contains the query identifier and
the second column contains the deepest taxonomic lineage shared by a
sufficient fraction of hits (see --lca_cutoff). Database sequence
headers must carry taxonomic annotations in the format used by
--sintax (e.g., tax=k:Archaea,p:Euryarchaeota,c:Halobacteria). Set
--maxaccepts to a value greater than 1 for the LCA to be meaningful.
--lengthout
Add a sequence length annotation (;length=integer) to each sequence
header when writing fasta or fastq files.
--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.
--matched filename
Write query sequences matching a target sequence to filename, in fasta
format.
--maxhits non-negative integer
Set the maximum number of hits to report once the search is terminated
for a given query; hits are sorted by decreasing identity. Unlimited by
default, or when the argument is zero. When searching both strands,
--maxhits controls the total number of hits reported per query across
both strands.
--maxqsize positive integer
Reject query sequences with an abundance greater than integer.
--maxqt real
Reject the sequence match if the query/target sequence length ratio is
greater than real.
--maxseqlength positive integer
Discard sequences longer than positive integer (50,000 nucleotides by
default).
--maxsizeratio real
Reject the sequence match if the query/target abundance ratio is greater
than real.
--maxsl real
Reject the sequence match if the shorter/longer sequence length ratio is
greater than real.
--mincols positive integer
Reject the sequence match if the alignment length is shorter than
integer columns.
--minqt real
Reject the sequence match if the query/target sequence length ratio is
lower than real.
--minseqlength positive integer
Discard sequences shorter than positive integer (1 nucleotide by
default).
--minsizeratio real
Reject the sequence match if the query/target abundance ratio is lower
than real.
--minsl real
Reject the sequence match if the shorter/longer sequence length ratio is
lower than real.
--mintsize positive integer
Reject target sequences with an abundance lower than integer.
--mothur_shared_out filename
Write an OTU table to filename in the mothur ‘shared’ tab-separated
plain text format. The first line starts with label, group and
numOtus, followed by all OTU identifiers. Each subsequent line starts
with vsearch, the sample identifier, the total number of OTUs, and the
abundance of each OTU in that sample. Sample and OTU identifiers are
extracted from FASTA headers. OTUs are represented by the cluster
centroids.
--no_progress
Suppress the gradually increasing progress indicator normally written to
the standard error stderr(3).
--notmatched filename
Write the sequences that were not extracted to filename, in fasta
format.
--notrunclabels
Retain whole sequence headers in output files. By default, vsearch
truncates sequence headers at first space or tabulation. This option
suppresses truncation.
--otutabout filename
Write an OTU table to filename in a classic tab-separated plain text
format. The first line starts with #OTU ID followed by sample
identifiers. Each subsequent line starts with the OTU identifier
followed by the abundances in each sample. Sample and OTU identifiers
are extracted from FASTA headers (see --sample). OTUs are represented
by the cluster centroids. A taxonomy column is appended if taxonomy
information is available for any OTU.
--output_no_hits
Write both matching and non-matching queries to --alnout,
--blast6out, --samout, or --userout output files. Non-matching
queries are labelled ‘No hits’ in --alnout files.
--qsegout filename
Write the aligned part of each query sequence to filename, in fasta
format.
--quiet
Suppress messages to the standard output stdout(3) and standard
error stderr(3), except for warnings and error messages.
--relabel string
Replace sequence headers with the prefix string and a ticker (1, 2, 3,
etc.). For example, with --relabel "cluster:", the first sequence
header becomes ‘>cluster:1’, the second sequence header becomes
‘>cluster:2’, and so on. To retain annotations, use their corresponding
options (--lengthout, --eeout, and --sizeout). Use
--relabel_keep to also retain old sequence identifiers.
--relabel_keep
Retain old sequence identifiers by including them at the end of the new
headers, after a space.
--relabel_md5
Replace each sequence header with the MD5 digest derived from the
sequence itself. The sequence is converted to upper case, and each ‘U’
is replaced with a ‘T’ before computation of the digest. The MD5 digest
is a 128-bit value (16 bytes), represented using a string of 32 ASCII
characters. Each pair of characters encodes an hexadecimal value,
ranging from x00 to xff. See md5(3) for more details, and
--relabel_sha1 for an alternative hashing algorithm. To retain
annotations, use their corresponding options (--lengthout, --eeout,
and --sizeout). Use --relabel_keep to also retain old sequence
identifiers.
--relabel_self
Replace each sequence header with the sequence itself. To retain
annotations, use their corresponding options (--lengthout, --eeout,
and --sizeout). Use --relabel_keep to also retain old sequence
identifiers.
--relabel_sha1
Replace each sequence header with the SHA1 digest derived from the
sequence itself. The sequence is converted to upper case, and each ‘U’
is replaced with a ‘T’ before computation of the digest. The SHA1 digest
is a 160-bit value (20 bytes), represented using a string of 40 ASCII
characters. Each pair of characters encodes an hexadecimal value,
ranging from x00 to xff. See sha1(3) for more details, and
--relabel_md5 for an alternative hashing algorithm. To retain
annotations, use their corresponding options (--lengthout, --eeout,
and --sizeout). Use --relabel_keep to also retain old sequence
identifiers.
--rowlen positive integer
Set the width of alignment lines in --alnout output. The default value
is 64. Set to 0 to disable line wrapping.
--samheader
Include header lines (@HD, @SQ, @PG) in the SAM file produced by
--samout. By default, no header lines are written.
--samout filename
Write alignment results to filename in the SAM format, see
vsearch-sam(5). Use --samheader to
include header lines. Each non-header line is a SAM record representing
either a query-target alignment or the absence of a match. The alignment
column of each record uses the CIGAR format, see
vsearch-cigar(5).
--sample string
Add the given sample identifier string to sequence headers when
writing fasta or fastq files. For instance, if string is ‘ABC’, the
text ;sample=ABC will be added to the headers. string is silently
truncated at the first ‘;’ or whitespace character (space, tab, newline,
carriage return, vertical tab or form feed), so such characters should
not be used in string. Other characters (alphabetical, numerical and
punctuations) are accepted.
--self
Reject the sequence match if the query and target sequence labels are
identical.
--sizein
Use the abundance annotations present in sequence headers when reading
fasta or fastq file. Search for the pattern [>@;]size=integer[;].
Entries without abundance annotations are silently assumed to be of
size=1.
--sizeout
Add abundance annotations to sequence headers when writing fasta or
fastq files. Add the pattern ;size=integer. If option --sizein is
not used, abundance values are set to 1 for all entries. If --sizein
is used, existing abundance annotations are simply reported to output
files.
--top_hits_only
Report only the hits with the highest pairwise identity for each query.
--tsegout filename
Write the aligned part of each target sequence to filename, in fasta
format.
--uc filename
Write search results to filename in a tab-separated uclust-like format
with 10 columns. Each query produces a hit (H) or no-hit (N) record.
Use --uc_allhits to report all hits per query instead of just the top
hit. Columns are:
- record type:
H(hit) orN(no hit); - ordinal number of the target sequence (zero-based;
*for N); - length of the query sequence (
*for N); - percentage of identity with the target (
*for N); - match orientation
+or-(.for N); - not used;
0(H) or*(N); - not used;
0(H) or*(N); - CIGAR alignment string (
*for N); seevsearch-cigar(5); - query label;
- target label (
*for N).
--uc_allhits
When using --uc, report all hits for each query, not just the top hit.
--userfields string
Select and order the fields written to --userout output. Fields are
separated by + (e.g. query+target+id). See
vsearch-userfields(7) for a
complete description of all available fields.
--userout filename
Write user-defined tab-separated output to filename. Select and order
the fields with --userfields.
--xee
Strip expected error (ee) annotations from sequence headers when writing
fasta or fastq files. Search for the pattern [>@;]ee=float[;].
Expected error annotations are added by the synonymous options
--fastq_eeout and --eeout described in
vsearch-fastx_filter(1).
--xlength
Strip sequence length annotations from sequence headers when writing
fasta or fastq files. Search for the pattern [>@;]length=integer[;].
Sequence length annotations are added by the --lengthout option.
--xsize
Strip abundance annotations from sequence headers when writing fasta or
fastq files. Search for the pattern [>@;]size=integer[;]. Abundance
annotations are added by the --sizeout option.
pairwise alignment options
These options modify the pairwise alignment scoring model used when
writing alignment output (e.g., --alnout). Modify with caution.
--match integer
Set the score assigned to a match (i.e. equivalent nucleotides) in
pairwise alignments. The default value is 2.
--mismatch integer
Set the score assigned to a mismatch (i.e. different nucleotides) in
pairwise alignments. The default value is -4.
EXAMPLES
Find all exact matches in a database and write results in BLAST-like tabular format:
vsearch \
--search_exact queries.fasta \
--db reference.fasta \
--blast6out results.tsv
Search both strands and write matching and non-matching queries to separate fasta files:
vsearch \
--search_exact queries.fasta \
--db reference.fasta \
--strand both \
--matched exact_hits.fasta \
--notmatched no_hits.fasta
Build an OTU table by mapping dereplicated reads to OTU centroids using exact matching:
vsearch \
--search_exact reads.fasta \
--db otus.fasta \
--otutabout otu_table.tsv \
--threads 8
SEE ALSO
vsearch-usearch_global(1),
vsearch-cigar(5),
vsearch-fasta(5),
vsearch-fastq(5),
vsearch-userfields(7)
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.