Development manual for vsearch 2.33.0, built from the dev branch: it describes changes that are not released yet. The manual for the current release is published separately, next to every published version.

NAME

usearch compatibility — the commands, options and behaviours vsearch shares with usearch, and those it does not

DESCRIPTION

If you are a usearch user, our objective is to make you feel at home: vsearch was designed to behave like usearch, to some extent. Like any complex software, usearch is not free from quirks and inconsistencies. We decided not to reproduce some of them, and, for complete transparency, this page documents the differences that matter when a usearch command line is ported to vsearch: commands vsearch does not implement, commands and options it spells differently, options it accepts and ignores, options it rejects, and behaviours it changed on purpose.

Two differences are broad enough to state first. vsearch works with nucleotide sequences only; amino acid sequences are not supported (see vsearch-nucleotides(7)). And vsearch always aligns globally, with full dynamic programming, so the options by which usearch tunes or bounds its alignment stage have nothing to act on (see vsearch-pairwise_alignment_parameters(7)).

COMMANDS NOT IMPLEMENTED

cluster_otus: Not implemented, and there is no exact equivalent. It combines greedy clustering with chimera filtering (the UPARSE-OTU algorithm), so the closest vsearch route is a clustering command followed by a chimera-detection command, run separately. usearch’s own manual considers 97% OTU clustering obsolete for most purposes, and recommends denoising instead — a route vsearch does provide (see unoise3 in the next section).
search_pcr, search_pcr2: Not implemented. Extracting the region between two primers is not something vsearch does (--search_oligodb finds the primers, see vsearch-search_oligodb(1)); --cut cuts at a restriction pattern, which is a different operation (see vsearch-cut(1)).
otutab_rare, otutab_norm: Not implemented. Rarefying or normalising an OTU table to a common number of reads is left to downstream tools. --fastx_subsample subsamples sequences, not table columns (see vsearch-fastx_subsample(1)).
otutab_octave: Not implemented.
sintax_summary: Not implemented. --sintax writes per-query classifications; summing them per rank is left to downstream tools (see vsearch-sintax(1)).

COMMANDS SPELLED DIFFERENTLY

The operation is available, under another name, or as an option of another command.

unoise3: --cluster_unoise, then --uchime3_denovo. usearch denoises and removes chimeras in a single command; vsearch splits the two, so the chimera step is a separate run (see vsearch-cluster_unoise(1)).
otutab: --usearch_global with --otutabout. The OTU sequences that usearch passes with -otus or -zotus are the --db here, and the identity threshold that usearch fixes at 0.97 by default is the usual --id (see vsearch-usearch_global(1)). --search_exact also writes --otutabout, when only exact matches should count.
fastx_truncate: --fastx_filter, whose --fastq_stripleft, --fastq_stripright and --fastq_trunclen perform the same trimming; --fastq_trunclen_keep keeps the sequences that are shorter than the requested length, where usearch discards them. Padding (-padlen, -padq) has no equivalent (see vsearch-fastx_filter(1)).
fastx_orient: --orient. usearch 11 accepts both spellings, orient and fastx_orient; usearch 12 keeps only the second (see vsearch-orient(1)).
cluster_fast with -sort size: --cluster_size. vsearch names the sort order rather than passing it: --cluster_fast sorts by decreasing length, --cluster_size by decreasing abundance, and --cluster_smallmem does not sort at all, expecting its input already sorted by decreasing length (see vsearch-cluster_size(1)).

OPTIONS SPELLED DIFFERENTLY

Same quantity, different name, and in one case a different direction.

usearch	vsearch
`--fastq_pctid` x	`--fastq_maxdiffpct` 100 - x
`--nowordcountreject`	`--minwordmatches 0`
`--queryalnfract`	`--query_cov`
`--targetalnfract`	`--target_cov`

--fastq_pctid bounds the identity of the overlap from below, where --fastq_maxdiffpct bounds the percentage of mismatches from above, so the two arguments are complements. Older usearch versions had --fastq_maxdiffpct itself, under that name.

OPTIONS ACCEPTED AND IGNORED

These options are recognised so that an existing command line still runs, but they have no effect. --band, --fulldp, --hspw, --minhsp, --pattern, --slots and --xdrop_nw all describe usearch’s alignment or seeding stage, which vsearch does not share: vsearch always performs a full dynamic programming alignment, so there is no band to widen and no heuristic to relax.

--cons_truncate is also accepted and ignored, with a warning.

--strand plus is accepted by --uchime_ref for compatibility, and is the only value it takes; --strand both is a fatal error there.

OPTIONS NOT RECOGNISED

Some options of usearch 5, 6 and 7, used by pipelines of that era such as QIIME 1, are rejected outright rather than ignored:

--global: vsearch always aligns globally, so the option would assert the only behaviour there is. Remove it.
--evalue: vsearch does not compute E-values for nucleotide alignments. The evalue field of --userfields exists, and always reports -1 (see vsearch-userfields(7)).
--query: Not an option in vsearch. The query file is the argument to the search command itself, as in --usearch_global queries.fasta. The name survives as the query field of --userfields.

DELIBERATE DIFFERENCES

Corrections and extensions vsearch applies on purpose. They are differences in output or in accepted input, not accidents.

--search_global ignores --maxaccepts and --maxrejects, which usearch honours. An exhaustive search has no ranked candidate order for an early stop to cut short, so honouring them would make the reported hits depend on how the database happens to be sorted, which is the one property the command exists to remove. They are accepted and have no effect, so a command line written for --usearch_global can be reused unchanged; --maxhits limits how many hits are reported. --search_global also defaults --minseqlength to 1 rather than 32, nothing in it requiring a sequence to hold a whole word (see vsearch-search_global(1)).
--search_oligodb allows gaps (usearch’s command aligns without gaps; --maxgaps 0 does the same), lets any number of Ns in a sequence match (usearch drops a site with two Ns or more), reports occurrences truncated by a sequence end, orders occurrences by position rather than by score, and swaps qlo and qhi on the minus strand as every vsearch command does. Its defaults are --strand both and --qmask none (see vsearch-search_oligodb(1)).
With --blast6out and --output_no_hits, usearch reports 13 fields for a query with no match, where the format has 12. vsearch reports the 12 the format calls for.
With --output_no_hits, usearch lists queries without a match in the --blast6out file but not in the alignment output. vsearch lists them in both, as No hits.
The raw field of --userfields is not informative in usearch. vsearch reports the alignment score.
The fields qlo, qhi, tlo and thi have counterparts qilo, qihi, tilo and tihi reporting alignment coordinates that ignore terminal gaps.
--iddef, and with it the alternative pairwise identity definitions that usearch removed, is available in vsearch.
--topn extends to the sorting commands.
--sizein extends to dereplication and clustering.
T and U count as the same nucleotide during dereplication, so an RNA and a DNA spelling of one sequence collapse into a single entry.
Sorting is stabilised, using sequence abundances or labels as secondary and tertiary keys, so equal keys do not come out in an arbitrary order.
Low-complexity regions are masked with the DUST algorithm by default, and masking behaviour is more consistent (see vsearch-fastx_mask(1)).
This includes --sintax, which DUST-masks a fasta reference database by default (--dbmask dust). usearch’s -sintax masks nothing unless -dbmask is given, so low-complexity reference regions can take part in its classifications and not in vsearch’s; --dbmask none restores usearch’s default. The two also treat lowercase letters differently: vsearch’s dust ignores the case of the input and masks only low-complexity regions, whereas usearch’s -dbmask dust keeps any lowercase region masked as well, so an all-lowercase reference database leaves every query unclassified in usearch (see vsearch-sintax(1)).
vsearch adds the command --cluster_size, which sorts sequences by decreasing abundance before clustering.
--relabel @, which derives a sample identifier from the input file name, is accepted by every command that accepts --relabel. usearch honours -relabel @ for fastq_filter and fastq_mergepairs only, and takes it as a literal @ prefix everywhere else, so a command such as --fastx_uniques writes ‘>sampleA.1’ where usearch writes ‘>@1’. vsearch also refuses --relabel @ on standard input, which has no file name to derive from, and warns when the name yields an empty identifier; usearch cannot read standard input at all. The derivation itself follows usearch’s implementation rather than its documentation (see --relabel in any command’s manual page).
The output-label column of --fastx_uniques --tabbedout follows --relabel_md5, --relabel_self and --relabel_sha1 as well as --relabel. usearch has none of those three options, so the column is byte-identical to its own whenever -relabel is what was given.

The --tabbedout report of --fastq_mergepairs follows the file usearch writes for that option, with the same one-line-per-pair shape and the same result=merged or result=notmerged ending each line. It differs in the following ways, all deliberate (see vsearch-fastq_mergepairs(1)):

vsearch reports the percentage of differences in the overlap (diffpct=), where usearch reports the percentage of identity (pctid=). The two are complements, for the same reason --fastq_pctid and --fastq_maxdiffpct are (see OPTIONS SPELLED DIFFERENTLY above); the field is named after the option that bounds it.
The three components of aln= are never negative. A pair whose reads run past each other is reported with a separate stagger= field giving the bases trimmed, where usearch encodes the same information as a negative first or third component of aln= – which cannot be read by splitting the value on -.
trunc= reports the read lengths that remain after --fastq_truncqual, where usearch’s tailf= and tailr= report the number of bases removed.
relabel= gives the label the merged sequence was written with, including any ;size=, ;ee= or ;length= annotation requested. usearch has no such annotations, so its relabel= is always a bare label.
Lines appear in input order whatever --threads is set to. usearch emits them in the order its threads finish, unless -threads 1 is given.
The first field is the whole header of the forward read, description included, which is the label the merged-read outputs of the command carry. usearch truncates its first field at the first blank.

usearch also reports a minq= field, for its -fastq_minqual. vsearch has --fastq_minqual, but --fastq_mergepairs does not accept it (only --fastq_filter and --fastx_filter do), so no such field is written.

vsearch also provides commands and options that usearch does not; they are listed in vsearch(1) and described one page per command.

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/. That address always serves the manual of the current release; the manuals of earlier releases are listed at https://torognes.github.io/vsearch/versions/.

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.