NAME
vsearch \-\-fastq_mergepairs — merge paired-end reads into one sequence
SYNOPSIS
vsearch \-\-fastq_mergepairs fwdfile \-\-reverse revfile (\-\-fastaout | \-\-fastqout) filename [options]
DESCRIPTION
The vsearch command --fastq_mergepairs merges paired-end sequence
reads into a single sequence by aligning the forward and reverse reads
and combining their overlapping regions. The forward reads are specified
as the argument to this option; the reverse reads are specified with
--reverse. Reads are matched by position: the first forward read is
paired with the first reverse read, the second with the second, and so
on. Labels are not used for matching, but a warning is emitted if the
two files contain different numbers of reads.
The reverse read is reverse-complemented before alignment. Merging
requires an overlap between the two reads of at least --fastq_minovlen
bases (default 10, minimum 5). Read pairs with too many mismatches in
the overlap — more than --fastq_maxdiffs (default 10) or more than
--fastq_maxdiffpct percent (default 100.0%) — are discarded.
Additional heuristics prevent merging of read pairs that cannot be
aligned reliably.
In the merged region, quality scores from the two reads are combined
using the Phred score formula. Outside the overlap, the quality scores
from the contributing read are used directly. Output quality scores can
be clamped with --fastq_qmaxout and --fastq_qminout (these apply
only to the merged region).
Staggered pairs — where the 3’ end of the reverse read extends past the
5’ end of the forward read — are discarded by default. Use
--fastq_allowmergestagger to allow them; the overhanging portion is
excluded from the merged sequence.
Reads can be pre-filtered with --fastq_truncqual, --fastq_maxee,
--fastq_minlen, --fastq_maxlen, and --fastq_maxns. Bounds on the
merged sequence length are set with --fastq_minmergelen and
--fastq_maxmergelen.
To illustrate a merge with a 6-base overlap:
Forward (5'→3'): AAAAAAAAATTTTTTTGGG
Reverse (5'→3'): CCCCCCCCCCGGGGGG (reverse complement of rev read)
Aligned:
Forward: AAAAAAAAATTTTTTTGGG
Rev-comp: TTTTTTTGGGCCCCCCCCCC
Overlap: TTTTTTTGGG (6 bases; scores combined)
Merged: AAAAAAAAATTTTTTTGGGCCCCCCCCCC
OPTIONS
mandatory options
--reverse filename
Specify the FASTQ file containing the reverse reads.
At least one of the following output options is required:
--fastaout filename
Write merged sequences to filename, in fasta format.
--fastqout filename
Write merged sequences to filename, in fastq format (see
vsearch-fastq(5)).
core options
--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_minovlen positive integer
Set the minimum length of the overlap region between the forward and
reverse reads. Must be at least 5. The default is 10.
--fastq_maxdiffs positive integer
Set the maximum number of mismatches allowed in the overlap region. This
option has a strong influence on the merging success rate. The default
is 10.
--fastq_maxdiffpct real
Set the maximum percentage of mismatches allowed in the overlap region
(0.0 to 100.0). Additional heuristics in the merging algorithm may still
discard pairs with a high mismatch rate. The default is 100.0.
--fastq_allowmergestagger
Allow merging of staggered read pairs. Staggered pairs arise when a very
short fragment is sequenced: the 3’ end of the reverse read extends
beyond the 5’ end of the forward read. The overhanging portion of the
reverse read is discarded. By default, staggered pairs are not merged
(see --fastq_nostagger).
--fastq_nostagger
Discard staggered read pairs. This is the default behaviour. See
--fastq_allowmergestagger to change it.
--fastq_minmergelen positive integer
Discard merged sequences shorter than positive integer bases. The
default is 1.
--fastq_maxmergelen positive integer
Discard merged sequences longer than positive integer bases. The
default is 1,000,000.
--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.
--fastq_qmaxout positive integer
Specify the maximum quality score used when writing fastq files. The
default is 41, which is usual for recent Sanger/Illumina 1.8+ files.
Older formats may use a maximum quality score of 40.
--fastq_qminout positive integer
Specify the minimum quality score used when writing fastq files. The
default is 0, which is usual for recent Sanger/Illumina 1.8+ files.
Older formats may use scores between -5 and 2.
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.
--eeout
Add the expected error count to each sequence header in output fasta or
fastq files, as the annotation ;ee=float. Synonym of --fastq_eeout.
Use --xee to remove this annotation from headers. See
vsearch-expected_error(7).
--eetabbedout filename
Write per-pair expected error statistics to filename, in a
tab-separated format with four columns: the expected errors in the
forward read, the expected errors in the reverse read, the observed
differences in the forward read within the overlap region, and the
observed differences in the reverse read within the overlap region. See
vsearch-expected_error(7).
--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.
--fastaout_notmerged_fwd filename
Write forward reads that could not be merged to filename, in fasta
format.
--fastaout_notmerged_rev filename
Write reverse reads that could not be merged to filename, in fasta
format.
--fastq_eeout
Add the expected error count to each sequence header in output fasta or
fastq files, as the annotation ;ee=float. Synonym of --eeout. Use
--xee to remove this annotation from headers. See
vsearch-expected_error(7).
--fastq_maxee positive real
Discard sequences with an expected error greater than real. The
expected error is the sum of error probabilities for all positions in
the sequence, and is strictly positive (zero or negative arguments are
rejected, as they would discard every sequence). Applied after trimming.
See vsearch-expected_error(7).
--fastq_maxlen positive integer
Discard sequences longer than positive integer bases. Applied after
trimming.
--fastq_maxns positive integer
Discard sequences containing more than positive integer ambiguous
bases (N).
--fastq_minlen positive integer
Discard sequences shorter than positive integer bases. Applied after
trimming. Default is 1.
--fastq_truncqual positive integer
Truncate reads starting at the first base whose quality score is at or
below positive integer.
--fastqout_notmerged_fwd filename
Write forward reads that could not be merged to filename, in fastq
format (see vsearch-fastq(5)).
--fastqout_notmerged_rev filename
Write reverse reads that could not be merged to filename, in fastq
format (see vsearch-fastq(5)).
--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’.
--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.
--no_progress
Suppress the gradually increasing progress indicator normally written to
the standard error stderr(3).
--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.
--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.
--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.
--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.
--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.
EXAMPLES
Merge paired-end reads and write merged sequences to a fastq file:
vsearch \
--fastq_mergepairs fwd.fastq \
--reverse rev.fastq \
--fastqout merged.fastq
Merge with a stricter overlap and mismatch threshold, and save unmerged reads for inspection:
vsearch \
--fastq_mergepairs fwd.fastq \
--reverse rev.fastq \
--fastq_minovlen 20 \
--fastq_maxdiffs 5 \
--fastqout merged.fastq \
--fastqout_notmerged_fwd unmerged_fwd.fastq \
--fastqout_notmerged_rev unmerged_rev.fastq
Allow staggered pairs and filter on expected error after merging:
vsearch \
--fastq_mergepairs fwd.fastq \
--reverse rev.fastq \
--fastq_allowmergestagger \
--fastq_maxee 1.0 \
--fastqout merged.fastq
SEE ALSO
vsearch-fastx_filter(1),
vsearch-fastq_eestats(1),
vsearch-fastq(5),
vsearch-expected_error(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.