Development manual for vsearch 2.32.0, built from the dev branch: it describes changes that are not released yet. The manual for the current release is published separately.

NAME

vsearch --orient — use a reference database to orient fasta or fastq sequences

SYNOPSIS

vsearch --orient fastxfile --db (fastxfile | udbfile) (--fastaout | --fastqout | --notmatched | --tabbedout) outputfile [options]

DESCRIPTION

The vsearch command --orient detects the orientation of input sequences by comparing them to a reference database specified with --db. The two strands of each input sequence are decomposed into words (k-mers) of length --wordlength (default 12) and compared against words from the database. A word counts for a strand only when it is at least 8 times more frequent in the database than its reverse complement. If one strand has at least one such decisive word and at least 4 times as many as the other strand, that strand is selected as the correct orientation. Otherwise, the orientation is reported as undetermined (?).

Correctly oriented sequences are written with --fastaout or --fastqout. Sequences with undetermined orientation are written with --notmatched. A summary table can be written with --tabbedout. At least one output option is required.

To illustrate the three possible outcomes for an input sequence:

Outcome  Symbol  Meaning
-------  ------  --------------------------------------------------
forward  +       forward strand matches the database
reverse  -       reverse-complementary strand matches the database
unknown  ?       neither strand clearly matches (written to --notmatched)

OPTIONS

mandatory options

--db filename: Read reference sequences from filename, in fasta, fastq, or UDB format. A UDB file may have been created with any --wordlength: the word length stored in the file is used, overriding the configured one with a warning (see vsearch-makeudb_usearch(1)). --db accepts - to read the database from standard input, as well as an explicit stream path such as /dev/stdin, a named pipe, or a process substitution. The query and the database cannot both be -, however, as they would compete for the same standard input; give at least one of them an explicit path.

At least one of the following output options is required:

--fastaout filename

Write correctly oriented sequences to filename, in fasta format.

--fastqout filename

Write correctly oriented sequences to filename, in fastq format (see vsearch-fastq(5)). Requires fastq input.

--notmatched filename

Write sequences with undetermined orientation to filename, in their original format.

--tabbedout filename

Write orientation results to filename as a tab-separated file with four columns:

query label;
orientation (+, -, or ?);
number of decisive words supporting the forward strand;
number of decisive words supporting the reverse-complementary strand (see the strand-selection rule in the DESCRIPTION).

core options

--wordlength positive integer

Set the length of words (i.e. k-mers) used for sequence comparisons. The range of possible values goes from 3 to 15 (default value is 12). Longer words may reduce the sensitivity/recall for weak similarities, but can increase precision. On the other hand, shorter words may increase sensitivity or recall, but may reduce precision.

Computation time generally increases with shorter words and decreases with longer words, but it increases again for very long words. Memory requirements for a part of the index increase by a factor of 4 each time word length increases by one nucleotide, and this generally becomes significant for words longer than 12.

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.

--dbmask none|dust|soft

Mask regions in the database sequences using the dust method or the soft method, or none to suppress masking. See vsearch-fastx_mask(1) for more details. Warning, when using soft masking, search commands become case sensitive: masking excludes masked regions from the k-mer index used to select candidate targets (the pairwise alignment itself always ignores case). A database sequence with no unmasked stretch of at least the word length contributes no k-mer to the index and is never selected as a candidate target; it can still be reported when the query itself samples no k-mer, as such a query is compared against every database sequence (see --qmask). vsearch warns when there are such sequences, giving how many. The default is to mask using dust.

--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.

--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.

--hardmask

Replace masked nucleotides with Ns, rather than lowercasing them.

This is also the only masking option that reaches the alignment. Soft and dust masking (see --qmask and --dbmask) only keep masked words out of the k-mer pre-filter that selects candidates; the masked region is still aligned and scored like any other. An N, on the other hand, scores zero and counts as a matching column, so hard masking makes a masked region match whatever it is aligned against, and the reported identity can only go up. Add --n_mismatch to count those columns as mismatches instead.

--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.

--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: masking excludes masked regions from the k-mer pre-filter that selects candidate targets (the pairwise alignment itself always ignores case). A query with no unmasked stretch of at least the word length samples no k-mers and is therefore compared against every database sequence. The default is to mask using dust.

Note: --orient does not run the DUST algorithm on the queries: masking only controls whether lowercase regions of the query sequences are ignored when counting words, so dust (the default) behaves like soft. The database side does run DUST as documented under --dbmask.

--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.

--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. Existing ;size= annotations are reported unchanged; entries without one receive ;size=1. For this command --sizein is not needed: abundance annotations are always parsed from the input headers.

--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

Use a reference database to orient sequences from a fastq file:

vsearch \
    --orient query.fastq \
    --db db.fastq \
    --fastaout query_oriented.fasta

Search for mis-oriented sequences by comparing a database against itself, writing a summary table:

vsearch \
    --orient db.fasta \
    --db db.fasta \
    --tabbedout db_outliers.tsv

Orient sequences and collect those with undetermined orientation separately:

vsearch \
    --orient query.fasta \
    --db db.fasta \
    --fastaout oriented.fasta \
    --notmatched undetermined.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.