vsearch

Versatile open-source tool for microbiome analysis

View on GitHub

NAME

vsearch \-\-uchime_denovo — detect chimeras de novo using the UCHIME algorithm

SYNOPSIS

vsearch \-\-uchime_denovo fastafile (\-\-chimeras | \-\-nonchimeras | \-\-uchimealns | \-\-uchimeout) filename [options]

DESCRIPTION

The vsearch command --uchime_denovo detects chimeric sequences present in the fasta-formatted fastafile, without the use of external references (de novo). Sequences are compared on their plus strand only.

Chimera detection is based on a scoring function controlled by five options: --dn, --mindiffs, --mindiv, --minh, and --xn. The algorithm identifies candidate chimeras by finding three-way alignments where a query sequence can be modelled as a mosaic of two parent sequences.

Input sequences must carry abundance annotations in their headers (e.g. ;size=integer;). --sizein is always implied; it does not need to be specified. Sequences are automatically sorted by decreasing abundance before chimera detection. The assumption is that chimeras appear later in the PCR amplification process and are therefore less abundant than their parents (see --abskew).

At least one output option must be specified.

See also --uchime2_denovo and --uchime3_denovo for improved algorithms designed for denoised amplicons, and --uchime_ref for reference-based chimera detection.

OPTIONS

mandatory options

--uchime_denovo fastafile
Detect chimeras de novo in the fasta-formatted fastafile. This option is mandatory.

At least one of the following output options must be specified:

--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.

--uchimealns filename
Write three-way global alignments (parentA, parentB, chimera) to filename in a human-readable format. All sequences are converted to upper case before alignment. Lower case letters indicate disagreement in the alignment. Output order may vary when using multiple threads. Use --alignwidth to modify the alignment width.

--uchimeout filename
Write chimera detection results to filename using an 18-field, tab-separated uchime-like format. Output row order may vary when using multiple threads. Use --uchimeout5 for a format compatible with usearch version 5 and earlier. The 18 fields are:

  1. score: higher score means a more likely chimeric alignment.
  2. Q: query sequence label.
  3. A: parent A sequence label.
  4. B: parent B sequence label.
  5. T: top parent sequence label (parent most similar to the query).
  6. idQM: percentage of similarity between query (Q) and the model 13) constructed as a part of parent A and a part of parent B.
  7. idQA: percentage of similarity between query (Q) and parent A.
  8. idQB: percentage of similarity between query (Q) and parent B.
  9. idAB: percentage of similarity between parent A and parent B.
  10. idQT: percentage of similarity between query (Q) and top parent (T).
  11. LY: yes votes in the left part of the model.
  12. LN: no votes in the left part of the model.
  13. LA: abstain votes in the left part of the model.
  14. RY: yes votes in the right part of the model.
  15. RN: no votes in the right part of the model.
  16. RA: abstain votes in the right part of the model.
  17. div: divergence, defined as (idQM - idQT).
  18. YN: query is chimeric (Y), not chimeric (N), or borderline (?).

The --borderline option can also produce output, but only when combined with at least one of the output options listed above:

--borderline filename
Write borderline chimeric sequences to filename, in fasta format. Borderline chimeric sequences are sequences with a chimera score high enough to be suspicious, but not sufficiently different from their closest parent to be classified as chimeric.

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 2.0, which means that the parents should be at least 2 times more abundant than the chimera.

--dn strictly positive real number
Set the pseudo-count prior on the number of ‘no’ votes, corresponding to the parameter n in the chimera scoring function. Default value is 1.4. Increasing --dn reduces the likelihood of tagging a sequence as a chimera (fewer false positives, but also more false negatives).

--mindiffs positive integer
Set the minimum number of differences per segment. Default value is 3.

--mindiv real
Set the minimum divergence from the closest parent. Default value is 0.8.

--minh real
Set the minimum chimera score (h). Increasing this value tends to reduce the number of false positives and to decrease sensitivity. Default value is 0.28. Accepted values are strictly greater than 0.0; values above 1.0 are accepted but uncommon.

--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.

--uchimeout5
When using --uchimeout, write chimera detection results using a 17-field, tab-separated uchime-like format. This drops the 5th field (top parent T) from the standard --uchimeout format, for compatibility with usearch version 5 and earlier versions.

--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 the width of the three-way alignments written with --uchimealns. Default width is 80 nucleotides. Set to zero (0) to suppress folding.

--fasta_score
Add the chimera score to the sequence headers in the fasta output files for chimeras, non-chimeras, and borderline sequences, using the format ;uchime_denovo=float;.

--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

Detect chimeras de novo and write non-chimeric sequences to a file:

vsearch \
    --uchime_denovo amplicons.fasta \
    --sizein \
    --nonchimeras clean.fasta

Also write chimeras and borderline cases to separate files, and record detection scores:

vsearch \
    --uchime_denovo amplicons.fasta \
    --sizein \
    --nonchimeras clean.fasta \
    --chimeras chimeras.fasta \
    --borderline borderline.fasta \
    --fasta_score

A typical amplicon sequencing workflow: dereplicate, denoise, then remove chimeras:

vsearch \
    --derep_fulllength reads.fasta \
    --sizeout \
    --output derep.fasta

vsearch \
    --cluster_unoise derep.fasta \
    --sizein \
    --sizeout \
    --centroids denoised.fasta

vsearch \
    --uchime_denovo denoised.fasta \
    --sizein \
    --nonchimeras amplicons.fasta

SEE ALSO

vsearch-uchime2_denovo(1), vsearch-uchime3_denovo(1), vsearch-uchime_ref(1), vsearch-chimeras_denovo(1), vsearch-cluster_unoise(1), vsearch-derep_fulllength(1), vsearch-sortbysize(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:

  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: