NAME
vsearch \-\-chimeras_denovo — detect chimeras de novo in long exact sequences
SYNOPSIS
vsearch \-\-chimeras_denovo inputfile (\-\-chimeras | \-\-nonchimeras | \-\-alnout | \-\-tabbedout) outputfile [options]
DESCRIPTION
The vsearch command --chimeras_denovo detects chimeras de novo
(i.e. without external references) in long exact sequences (inputfile,
in fasta or fastq format). It uses a modified UCHIME algorithm that can
automatically adapt to a wide range of sequence lengths.
Abundance annotations (pattern [>@;]size=integer[;]) present in
sequence headers are taken into account by default. This means that
option --sizein is always implied, and does not need to be specified.
Sequences are sorted into chimeras and non-chimeras, and can be written
to fasta files (see output options --chimeras, --nonchimeras).
Additional information on each chimera can be collected with the output
options --tabbedout and --alnout. The latter outputs for each
chimera (i.e. ‘Query’) a multi-way alignment and a model showing the
most likely parent sequence of each section of the chimera. Here is an
example of a short chimeric sequence (Q), with two parents (A and B):
------------------------------------------------------------------------
Query ( 20 nt) Q
ParentA ( 20 nt) A
ParentB ( 20 nt) B
Q 1 GTAGGCCGTGCTGAGCCGTA 20
A 1 GTAGGCCGTGgTagGCCGTg 20
B 1 cTgaGCCGTaCTGAGCCGTA 20
Diffs A AA AB BB B
Model AAAAAAAAAABBBBBBBBBB
Ids. QA 80.00%, QB 80.00%, QC 0.00%, QT 80.00%, QModel 100.00%, Div. +25.00%
Lowercase positions indicate a mismatch between the parent and the query. The line Diffs indicates the positions that favour a particular parent when modelling the chimera. The line Ids gives global similarity percentages with the different parents (QA, QB, and QC), the closest parent (QT), the model (QModel, always 100.00%), and the divergence of the model with the closest parent (Div). If there are only two parents (A and B), QC is set to 0.00%. If there are more than three parents, only QA, QB and QC are reported.
OPTIONS
mandatory options
--chimeras_denovo inputfile
Detect chimeras de novo in inputfile (fasta or fastq format) using a
modified UCHIME algorithm that adapts to a wide range of sequence
lengths. This option is mandatory.
At least one of the following output options must be specified:
--alnout filename
Write multi-way chimera alignments to filename in a human-readable
format (see the example in the DESCRIPTION). Use --alignwidth to set
the alignment width (default 60 nucleotides).
--chimeras filename
Write chimeric sequences to filename, in fasta format. Output order
may vary when using multiple threads.
--nonchimeras filename
Write non-chimeric sequences to filename, in fasta format. Output
order may vary when using multiple threads.
--tabbedout filename
Write chimera detection results to filename as an eighteen-column
tab-separated file, with one row per chimera. Columns are:
- score: dummy value, always set to 99.9999
- query label
- parent A label
- parent B label
- parent C label (“*” if there are only two parents)
- QModel: maximum global similarity percentage (always 100.0%)
- QA: global similarity percentage with parent A
- QB: global similarity percentage with parent B
- QC: global similarity percentage with parent C (0.00 if only two parents)
- QT: highest similarity percentage with any parent
- left yes: ignored, always set to zero
- left no: ignored, always set to zero
- left abstain: ignored, always set to zero
- right yes: ignored, always set to zero
- right no: ignored, always set to zero
- right abstain: ignored, always set to zero
- dummy value, always set to 0.00
- chimeric status, always set to Y (only chimeras are reported)
core options
--abskew real
Set the minimum abundance skew ratio between a chimera and its potential
parent. The assumption is that chimeras appear later in the PCR
amplification process and are therefore less abundant than their
parents. Any positive value equal or greater than 1.0 can be used.
Default is 1.0, which means that the parents should be at least as
abundant than the chimera.
--chimeras_diff_pct real
Set the maximal mismatch percentage allowed in each chimeric region
(excluding insertion and deletions). Accepted values range from 0.0 to
50.0%. Default is 0.0 (no mismatch allowed).
--chimeras_length_min positive non-null integer
Set the minimum length of each chimeric region. Default is 10.
--chimeras_parents_max positive non-null integer
Set the maximum number of parent sequences. Accepted values range from 2
to 20. Default is 3.
--chimeras_parts integer from 2 to 100
Set the number of parts to divide sequences into. Accepted values range
from 2 to 100 (inclusive); values outside this range are rejected.
Default is (sequence length / 100).
--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.
Always implied.
--xn real number strictly greater than 1.0
Set the weight of ‘no’ votes, corresponding to the parameter beta in
the chimera scoring function. Default value is 8.0. Increasing --xn
reduces the likelihood of tagging a sequence as a chimera (less false
positives, but also more false negatives). Decreasing --xn reduces
false negative, but increases false positives.
secondary options
--alignwidth positive integer
Set maximal width of three-way alignments when writing alignments with
--alnout. Default width is 60 nucleotides. Set to zero (0) to suppress
folding.
--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.
--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’.
--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.
--maxseqlength positive integer
Discard sequences longer than positive integer (50,000 nucleotides by
default).
--minseqlength positive integer
Discard sequences shorter than positive integer (1 nucleotide 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.
--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.
--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.
--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.
--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 parameters of the pairwise alignment model. Modify with caution.
--gapext string
Set penalties for a gap extension. See
vsearch-pairwise_alignment_parameters(7)
for a complete description of the gap penalty declaration system. By
default, the penalty is set to 2 for extending internal gaps and to 1
for extending terminal gaps, in both query and target sequences.
--gapopen string
Set penalties for a gap opening. See
vsearch-pairwise_alignment_parameters(7)
for a complete description of the gap penalty declaration system. By
default, the penalty is set to 20 for opening internal gaps and to 2 for
opening terminal gaps, in both query and target sequences.
--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.
ignored options
--threads positive non-null integer
Command is not multithreaded, option has no effect.
EXAMPLES
A simple way to filter out chimeras:
vsearch \
--chimeras_denovo input.fasta \
--quiet \
--nonchimeras clean.fasta
Add option --tabbedout to log the sequences identified as chimeras,
and option --log to record run parameters:
vsearch \
--chimeras_denovo input.fasta \
--quiet \
--nonchimeras clean.fasta \
--tabbedout chimeras.tsv \
--log chimera_filtering.log
SEE ALSO
vsearch-uchime_denovo(1),
vsearch-uchime2_denovo(1),
vsearch-uchime3_denovo(1),
vsearch-uchime_ref(1),
vsearch-fasta(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.