Manual for vsearch 2.32.0, the current release. Changes that are not released yet are described in the development manual.

NAME

vsearch --derep_smallmem — merge strictly identical sequences using minimal memory

SYNOPSIS

vsearch --derep_smallmem fastxfile --fastaout filename [options]

DESCRIPTION

The vsearch command --derep_smallmem groups strictly identical sequences, like --derep_fulllength (see vsearch-derep_fulllength(1)), but uses significantly less memory (approximately 24 bytes per unique sequence). The output is written to --fastaout, in fasta format.

Key differences from --derep_fulllength:

Output is written in the order that unique sequences first appear in the input, not sorted by decreasing abundance.
Cannot read from a pipe; the input must be a regular file (it is read twice).
Can read fastq files, but output is always fasta.
Dereplication uses a 128-bit hash function; grouped sequences are not explicitly verified to be identical. The false positive probability is approximately 1.5e-21 for a dataset of one billion unique sequences.
Does not support --topn or --uc.

This command is not multithreaded. The inverse operation is --rereplicate (see vsearch-rereplicate(1)).

See vsearch-fasta(5) and vsearch-fastq(5) for a description of the input formats.

OPTIONS

mandatory options

--fastaout filename: Write the dereplicated sequences to filename, in fasta format, in the order they first appear in the input.

core options

--maxuniquesize positive integer

Discard sequences with a post-dereplication abundance greater than positive integer. When that value is smaller than the --minuniquesize value, no sequence can satisfy the abundance filter: vsearch issues a warning and writes no sequence.

The filter applies to the sequence output only. A cluster summary written with --uc, where the command offers it, still lists every cluster, discarded ones included. This matches usearch.

--minuniquesize positive integer

Discard sequences with a post-dereplication abundance smaller than positive integer. When that value is larger than the --maxuniquesize value, no sequence can satisfy the abundance filter: vsearch issues a warning and writes no sequence.

The filter applies to the sequence output only. A cluster summary written with --uc, where the command offers it, still lists every cluster, discarded ones included. This matches usearch.

--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, where the integer is the total abundance of the cluster: the sum of the members’ abundance annotations when --sizein is used, or the number of merged sequences otherwise.

--strand plus|both

Check the plus strand only (default), or check both strands when comparing sequences. Keywords are case-insensitive.

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 required when compressed data arrives on standard input through a pipe (‘-’), where the format cannot be detected without consuming the stream. It is not needed when reading from a regular file compressed with bzip2, nor when such a file is redirected to standard input: compression is then detected automatically, and a contradicting option is ignored (with a warning when the input is standard input). Pipes other than standard input, such as shell process substitutions and named FIFOs, are always read as uncompressed data; compressed data must arrive on standard input or as a named file.
--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.
--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.

The offset matters even to a command that never decodes a quality score: it is what the reader compares the observed quality symbols against before warning that the file may use the other encoding, it sets the default of --fastq_qmax (the highest score the offset can represent), and the sum rules on --fastq_qmin and --fastq_qmax are stated in terms of it.

--fastq_qmax positive integer: Option is ignored: quality scores are not checked when reading fastq sequences, so no score is rejected for being too high. The argument itself is still validated, as the offset (see --fastq_ascii) plus the maximal score may not exceed 126, the last printable ASCII character.
--fastq_qmin positive integer: Option is ignored: quality scores are not checked when reading fastq sequences, so no score is rejected for being too low. The argument itself is still validated, as the offset (see --fastq_ascii) plus the minimal score must be at least 33, the first printable ASCII character.
--gzip_decompress: Specify that the input pipe is streaming data compressed using Lempel-Ziv coding. See gzip(1) for more details. This option is required when compressed data arrives on standard input through a pipe (‘-’), where the format cannot be detected without consuming the stream. It is not needed when reading from a regular file compressed with gzip, nor when such a file is redirected to standard input: compression is then detected automatically, and a contradicting option is ignored (with a warning when the input is standard input). Pipes other than standard input, such as shell process substitutions and named FIFOs, are always read as uncompressed data; compressed data must arrive on standard input or as a named file.
--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. Messages are written in addition to the standard error, not instead of it; use --quiet to silence the standard error copy (see the --quiet entry for what it spares).
--maxseqlength positive integer: Discard sequences longer than positive integer (50,000 nucleotides by default). The value must not exceed 2,147,481,646 (INT_MAX minus 2,001). When that value is smaller than the effective --minseqlength value (whose default is command-specific), no sequence can pass the length filter: vsearch issues a warning.
--minseqlength non-negative integer: Discard sequences shorter than non-negative integer (1 nucleotide by default). A value of 0 retains empty sequences. When that value is larger than the --maxseqlength value, no sequence can pass the length filter: vsearch issues a warning.
--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.
--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. When nothing is left after truncation — an empty string, or one starting with ‘;’ or a blank character — vsearch issues a warning and writes a bare ;sample= annotation.
--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.

ignored options

--threads positive non-null integer: Command is not multithreaded, option has no effect (a warning is printed when more than one thread is requested).

EXAMPLES

Dereplicate sequences in input.fasta with minimal memory usage:

vsearch \
    --derep_smallmem input.fasta \
    --sizeout \
    --fastaout derep.fasta

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.

These manual pages are also published online at https://torognes.github.io/vsearch/.

COPYRIGHT

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.