vsearch

Versatile open-source tool for microbiome analysis

View on GitHub

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:

  1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.

  2. 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: