Skip site navigation (1)Skip section navigation (2)

FreeBSD Manual Pages

  
 
  

home | help
cmscan(1)			Infernal Manual			     cmscan(1)

NAME
       cmscan -	search sequence(s) against a covariance	model database

SYNOPSIS
       cmscan [options]	_cmdb_ _seqfile_

DESCRIPTION
       cmscan  is  used	 to search sequences against collections of covariance
       models.	For each sequence in _seqfile_,	use  that  query  sequence  to
       search the target database of CMs in _cmdb_, and	output ranked lists of
       the CMs with the	most significant matches to the	sequence.

       The _seqfile_ may contain more than one query sequence. It  can	be  in
       FASTA  format,  or several other	common sequence	file formats (genbank,
       embl, and among others),	 or  in	 alignment  file  formats  (stockholm,
       aligned	fasta,	and  others).  See the --qformat option	for a complete
       list.

       The _cmdb_ needs	to be press'ed using cmpress before it can be searched
       with  cmscan.  This creates four	binary files, suffixed .i1{fimp}.  Ad-
       ditionally, _cmdb_ must have been calibrated for	E-values with  cmcali-
       brate before being press'ed with	cmpress.

       The  query  _seqfile_  may be '-' (a dash character), in	which case the
       query sequences are read	from a <stdin> pipe instead of	from  a	 file.
       The  _cmdb_  cannot  be read from a <stdin> stream, because it needs to
       have those four auxiliary binary	files generated	by cmpress.

       The output format is designed to	be human-readable, but is often	so vo-
       luminous	 that reading it is impractical, and parsing it	is a pain. The
       --tblout	option saves output in a simple	tabular	format that is concise
       and easier to parse. The	--fmt 2	option modifies	the format of the tab-
       ular output by adding several fields, including markup  of  overlapping
       hits, as	described in section 6 of the Infernal user guide.  The	-o op-
       tion allows redirecting the main	output,	including throwing it away  in
       /dev/null.

       cmscan  reexamines the 5' and 3'	termini	of target sequences using spe-
       cialized	algorithms for detection of truncated hits, in which  part  of
       the  5'	and/or 3' end of the actual full length	homologous sequence is
       missing in the target sequence file. These types	of hits	will  be  most
       common in sequence files	consisting of unassembled sequencing reads. By
       default,	any 5' truncated hit is	required to include the	first  residue
       of  the	target sequence	it derives from	in _seqfile_, and any 3' trun-
       cated hit is required to	include	the final residue of  the  target  se-
       quence  it  derives  from. Any 5' and 3'	truncated hit must include the
       first and final residue of the target sequence  it  derives  from.  The
       --anytrunc  option will relax the requirements for hit inclusion	of se-
       quence endpoints, and truncated hits are	allowed	to start and  stop  at
       any   positions	 of   target   sequences.   Importantly	 though,  with
       --anytrunc, hit E-values	will be	less accurate because  model  calibra-
       tion  does  not	consider  the possibility of truncated hits, so	use it
       with caution.  The --notrunc option can be used to turn	off  truncated
       hit  detection.	--notrunc will reduce the running time of cmscan, most
       significantly for target	_seqfile_ files	that include  many  short  se-
       quences.	  Truncated hit	detection is automatically turned off when the
       --max, --nohmm, --qdb, or --nonbanded options are used because  it  re-
       lies on the use of an accelerated HMM banded alignment strategy that is
       turned off by any of those options.

OPTIONS
       -h     Help; print a brief reminder  of	command	 line  usage  and  all
	      available	options.

       -g     Turn  on	the glocal alignment algorithm,	global with respect to
	      the query	model and local	with respect to	the  target  database.
	      By default, the local alignment algorithm	is used	which is local
	      with respect to both the target sequence and the model. In local
	      mode,  the  alignment to span two	or more	subsequences if	neces-
	      sary (e.g. if the	structures of the query	model and  target  se-
	      quence are only partially	shared), allowing certain large	inser-
	      tions and	deletions in the structure to be penalized differently
	      than  normal  indels.  Local  mode  performs better on empirical
	      benchmarks and is	significantly more sensitive for remote	homol-
	      ogy  detection.  Empirically,  glocal searches return many fewer
	      hits than	local searches,	so glocal may be desired for some  ap-
	      plications.

       -Z _x_ Calculate	E-values as if the search space	size was _x_ megabases
	      (Mb). Without the	use of this  option,  the  search  space  size
	      changes  for each	query sequence,	it is defined as the length of
	      the current query	sequence times 2 (because both strands of  the
	      sequence will be searched) times the number of CMs in _cmdb_.

       --devhelp
	      Print  help,  as	with -h	, but also include expert options that
	      are not displayed	with -h	.  These expert	options	 are  not  ex-
	      pected  to be relevant for the vast majority of users and	so are
	      not described in the manual page.	 The only resources for	under-
	      standing	what  they actually do are the brief one-line descrip-
	      tions output when	--devhelp is enabled, and the source code.

OPTIONS	FOR CONTROLLING	OUTPUT
       -o _f_ Direct the main human-readable output to a file _f_  instead  of
	      the default stdout.

       --tblout	_f_
	      Save  a  simple  tabular	(space-delimited) file summarizing the
	      hits found, with one data	line per hit.  The format of this file
	      is described in section 6	of the Infernal	user guide.

       --fmt _n_
	      specify  the  format  of	the tabular output file	specified with
	      --tblout _f_ be in format	_n_.  Possible values for _n_ are 1 or
	      2.  By  default  _n_  is	1 when --tblout	is used	without	--fmt.
	      With --fmt 2 nine	additional fields are  added  to  the  tabular
	      output file, most	of which pertain to the	annotation of overlap-
	      ping hits.  See section 6	the Infernal user guide	for a descrip-
	      tion of both formats.

       --acc  Use accessions instead of	names in the main output, where	avail-
	      able for profiles	and/or sequences.

       --noali
	      Omit the alignment  section  from	 the  main  output.  This  can
	      greatly reduce the output	volume.

       --notextw
	      Unlimit  the length of each line in the main output. The default
	      is a limit of 120	characters per line, which helps in displaying
	      the output cleanly on terminals and in editors, but can truncate
	      target profile description lines.

       --textw _n_
	      Set the main output's line length	limit to  _n_  characters  per
	      line. The	default	is 120.

       --verbose
	      Include extra search pipeline statistics in the main output, in-
	      cluding filter survival statistics for truncated	hit  detection
	      and number of envelopes discarded	due to matrix size overflows.

OPTIONS	CONTROLLING REPORTING THRESHOLDS
       Reporting  thresholds  control  which hits are reported in output files
       (the main output	and --tblout) Hits are ranked by statistical  signifi-
       cance  (E-value).   By  default,	all hits with an E-value <= 10 are re-
       ported.	The following options allow you	to change the default  E-value
       reporting thresholds, or	to use bit score thresholds instead.

       -E _x_ In  the  per-target  output,  report target sequences with an E-
	      value of <= _x_.	The default is 10.0, meaning that on  average,
	      about  10	false positives	will be	reported per query, so you can
	      see the top of the noise and decide for yourself if it's	really
	      noise.

       -T _x_ Instead  of thresholding per-CM output on	E-value, report	target
	      sequences	with a bit score of >= _x_.

OPTIONS	FOR INCLUSION THRESHOLDS
       Inclusion thresholds are	stricter than reporting	thresholds.  Inclusion
       thresholds  control  which hits are considered to be reliable enough to
       be included in a	possible subsequent search round, or marked as signif-
       icant ("!") as opposed to questionable ("?") in hit output.

       --incE _x_
	      Use  an  E-value	of <= _x_ as the hit inclusion threshold.  The
	      default is 0.01, meaning that on average,	about 1	false positive
	      would be expected	in every 100 searches with different query se-
	      quences.

       --incT _x_
	      Instead of using E-values	for setting the	 inclusion  threshold,
	      instead  use  a bit score	of >= _x_ as the hit inclusion thresh-
	      old.  By default this option is unset.

OPTIONS	FOR MODEL-SPECIFIC SCORE THRESHOLDING
       Curated CM databases may	define specific	bit score thresholds for  each
       CM,  superseding	 any  thresholding  based  on statistical significance
       alone.

       To use these options, the profile must contain the appropriate (GA, TC,
       and/or  NC)  optional  score threshold annotation; this is picked up by
       cmbuild from Stockholm format alignment files. Each thresholding	option
       has  a score of _x_ bits, and acts as if	-T _x_ --incT _x_ has been ap-
       plied specifically using	each model's curated thresholds.

       --cut_ga
	      Use the GA (gathering) bit scores	in the model to	 set  hit  re-
	      porting  and  inclusion  thresholds. GA thresholds are generally
	      considered to be the reliable curated thresholds defining	family
	      membership;  for	example, in Rfam, these	thresholds define what
	      gets included in Rfam Full alignments  based  on	searches  with
	      Rfam Seed	models.

       --cut_nc
	      Use  the	NC (noise cutoff) bit score thresholds in the model to
	      set hit reporting	and inclusion thresholds.  NC  thresholds  are
	      generally	 considered  to	 be  the  score	of the highest-scoring
	      known false positive.

       --cut_tc
	      Use the TC (trusted cutoff) bit score thresholds in the model to
	      set  hit	reporting  and inclusion thresholds. TC	thresholds are
	      generally	considered to be the score of the lowest-scoring known
	      true positive that is above all known false positives.

OPTIONS	CONTROLLING THE	ACCELERATION PIPELINE
       Infernal	 searches  are accelerated in a	six-stage filter pipeline. The
       first five stages use a profile HMM to define envelopes that are	passed
       to  the stage six CM CYK	filter.	Any envelopes that survive all filters
       are assigned final scores using the the CM Inside algorithm.

       The profile HMM filter is built by the cmbuild program and is stored in
       _cmfile_.

       Each successive filter is slower	than the previous one, but better than
       it at disciminating between subsequences	that may contain  high-scoring
       CM  hits	 and  those that do not. The first three HMM filter stages are
       the same	as those used in HMMER3.  Stage	1 (F1) is the  local  HMM  SSV
       filter  modified	 for  long  sequences.	Stage  2 (F2) is the local HMM
       Viterbi filter. Stage 3 (F3) is the local HMM Forward filter.  Each  of
       the first three stages uses the profile HMM in local mode, which	allows
       a target	subsequence to align to	any region of the HMM. Stage 4 (F4) is
       a  glocal  HMM  filter, which requires a	target subsequence to align to
       the full-length profile HMM. Stage 5 (F5) is the	 glocal	 HMM  envelope
       definition filter, which	uses HMMER3's domain identification heursitics
       to define envelope boundaries. After each stage from 2 to 5 a bias fil-
       ter  step (F2b, F3b, F4b, and F5b) is used to remove sequences that ap-
       pear to have passed the filter due to biased composition	alone. Any en-
       velopes that survive stages F1 through F5b are then passed with the lo-
       cal CM CYK filter. The CYK filter uses constraints (bands) derived from
       an  HMM alignment of the	envelope to reduce the number of required cal-
       culations and save time.	 Any envelopes that pass CYK are  scored  with
       the local CM Inside algorithm, again using HMM bands for	acceleration.

       The  default  filter  thresholds	that define the	minimum	score required
       for a subsequence to survive each stage are defined based on  the  size
       of  the search space (Z), which is defined as the length	of the current
       query sequence times 2 (because both strands will  be  searched)	 times
       the  number  of	profiles  in _cmdb_.  However, if either the -Z	_x_ or
       --FZ _x_	options	are used then the search space will be	considered  to
       be _x_ for purposes of defining the filter thresholds.

       For  larger  databases, the filters are more strict leading to more ac-
       celeration but potentially a greater loss of sensitivity. The rationale
       is  that	 for larger databases, hits must have higher scores to achieve
       statistical significance, so  stricter  filtering  that	removes	 lower
       scoring insignificant hits is acceptable.

       The P-value thresholds for all possible search space sizes and all fil-
       ter stages are listed next. (A P-value threshold	 of  0.01  means  that
       roughly	1%  of	the  highest scoring nonhomologous subsequence are ex-
       pected to pass the filter.) Z is	defined	as the number  of  nucleotides
       in  the complete	target sequence	file times 2 because both strands will
       be searched with	each model.

       If Z is less than 2 Mb: F1 is 0.35; F2 and F2b are off;	F3,  F3b,  F4,
       F4b and F5 are 0.02; F6 is 0.0001.

       If  Z  is  between  2 Mb	and 20 Mb: F1 is 0.35; F2 and F2b are off; F3,
       F3b, F4,	F4b and	F5 are 0.005; F6 is 0.0001.

       If Z is between 20 Mb and 200 Mb: F1 is 0.35; F2	and F2b	are 0.15;  F3,
       F3b, F4,	F4b and	F5 are 0.003; F6 is 0.0001.

       If  Z  is between 200 Mb	and 2 Gb: F1 is	0.15; F2 and F2b are 0.15; F3,
       F3b, F4,	F4b, F5, and F5b are 0.0008; and F6 is 0.0001.

       If Z is between 2 Gb and	20 Gb: F1 is 0.15; F2 and F2b  are  0.15;  F3,
       F3b, F4,	F4b, F5, and F5b are 0.0002; and F6 is 0.0001.

       If  Z is	more than 20 Gb: F1 is 0.06; F2	and F2b	are 0.02; F3, F3b, F4,
       F4b, F5,	and F5b	are 0.0002; and	F6 is 0.0001.

       These thresholds	were chosen based on performance on an internal	bench-
       mark testing many different possible settings.

       There  are  five	 options  for controlling the general filtering	level.
       These options are, in order from	least strict (slowest but most	sensi-
       tive)  to  most	strict	(fastest but least sensitive): --max, --nohmm,
       --mid, --default, (this is the default setting) --rfam.	and --hmmonly.
       With  --default	the filter thresholds will be database-size dependent.
       See the explanation of each of these individual options below for  more
       information.

       Additionally,  an  expert  user can precisely control each filter stage
       score threshold with the	--F1, --F1b, --F2, --F2b, --F3,	 --F3b,	 --F4,
       --F4b,  --F5, --F5b, and	--F6 options. As well as turn each stage on or
       off with	the --noF1, --doF1b, --noF2, --noF2b, --noF3, --noF3b, --noF4,
       --noF4b,	 --noF5,  and  --noF6.	 options.  These options are only dis-
       played if the --devhelp option is used to keep the number of  displayed
       options	with  -h  reasonable, and because they are only	expected to be
       useful to a small minority of users.

       As a special case, for any models in _cmfile_  which  have  zero	 base-
       pairs,  profile	HMM searches are run instead of	CM searches. HMM algo-
       rithms are more efficient than CM algorithms, and the benefit of	CM al-
       gorithms	 is  lost  for	models with no secondary structure (zero base-
       pairs). These profile HMM searches will run significantly  faster  than
       the CM searches.	You can	force HMM-only searches	with the --hmmonly op-
       tion. For more information on HMM-only searches see the user guide.

       --max  Turn off all filters, and	run non-banded Inside on  every	 full-
	      length  target sequence. This increases sensitivity somewhat, at
	      an extremely large cost in speed.

       --nohmm
	      Turn off all HMM filter stages (F1 through F5b). The CYK filter,
	      using QDBs, will be run on every full-length target sequence and
	      will enforce a P-value threshold	of  0.0001.  Each  subsequence
	      that  survives CYK will be passed	to Inside, which will also use
	      QDBs (but	a looser set). This increases sensitivity somewhat, at
	      a	very large cost	in speed.

       --mid  Turn off the HMM SSV and Viterbi filter stages (F1 through F2b).
	      Set remaining HMM	filter thresholds (F3 through F5b) to 0.02  by
	      default,	but  changeable	 to _x_	with --Fmid _x_	sequence. This
	      may increase sensitivity,	at a significant cost in speed.

       --default
	      Use the default filtering	strategy. This option  is  on  by  de-
	      fault.  The  filter thresholds are determined based on the data-
	      base size.

       --rfam Use a strict filtering  strategy	devised	 for  large  databases
	      (more  than  20 Gb). This	will accelerate	the search at a	poten-
	      tial cost	to sensitivity.

       --hmmonly
	      Only use the filter profile HMM for searches, do not use the CM.
	      Only  filter stages F1 through F3	will be	executed, using	strict
	      P-value thresholds (0.02 for F1, 0.001 for F2  and  0.00001  for
	      F3).   Additionally  a bias composition filter is	used after the
	      F1 stage (with P=0.02 survival threshold).  Any  hit  that  sur-
	      vives  all  stages and has an HMM	E-value	or bit score above the
	      reporting	threshold will be output.  The	user  can  change  the
	      HMM-only	filter	thresholds  and	options	with --hmmF1, --hmmF2,
	      --hmmF3, --hmmnobias, --hmmnonull2, and --hmmmax.	  By  default,
	      searches	for  any model with zero basepairs will	be run in HMM-
	      only mode. This can be turned off, forcing CM searches for these
	      models with the --nohmmonly option.

       --FZ _x_
	      Set  filter thresholds as	the defaults used if the database were
	      <x> megabases (Mb). If used with <x> greater than	20000 (20  Gb)
	      this option has the same effect as --rfam.

       --Fmid _x_
	      With  the	--mid option set the HMM filter	thresholds (F3 through
	      F5b) to _x_.  By default,	_x_ is 0.02.

OTHER OPTIONS
       --notrunc
	      Turn off truncated hit detection.

       --anytrunc
	      Allow truncated hits to begin and	end at any position in a  tar-
	      get  sequence.  By  default,  5' truncated hits must include the
	      first residue of their target sequence  and  3'  truncated  hits
	      must  include  the  final	residue	of their target	sequence. With
	      this option you may observe fewer	full length hits  that	extend
	      to the beginning and end of the query CM.

       --nonull3
	      Turn  off	the null3 CM score corrections for biased composition.
	      This correction is not used during the HMM filter	stages.

       --mxsize	_x_
	      Set the maximum allowable	CM DP matrix size to _x_ megabytes. By
	      default  this  size  is 128 Mb.  This should be large enough for
	      the vast majority	of searches, especially	with  smaller  models.
	      If cmscan	encounters an envelope in the CYK or Inside stage that
	      requires a larger	matrix,	the envelope will be  discounted  from
	      consideration.  This  behavior is	like an	additional filter that
	      prevents expensive (slow)	CM DP calculations, but	at a potential
	      cost  to	sensitivity.   Note that if cmscan is being run	in _n_
	      multiple threads on a multicore machine  then  each  thread  may
	      have an allocated	matrix of up to	size _x_ Mb at any given time.

       --smxsize _x_
	      Set  the	maximum	 allowable  CM	search	DP  matrix size	to _x_
	      megabytes. By default this size is 128 Mb.  This option is  only
	      relevant if the CM will not use HMM banded matrices, i.e.	if the
	      --max, --nohmm, --qdb, --fqdb, --nonbanded, or --fnonbanded  op-
	      tions  are  also used. Note that if cmsearch is being run	in _n_
	      multiple threads on a multicore machine  then  each  thread  may
	      have an allocated	matrix of up to	size _x_ Mb at any given time.

       --cyk  Use  the CYK algorithm, not Inside, to determine the final score
	      of all hits.

       --acyk Use  the	CYK  algorithm	to  align  hits.   By	default,   the
	      Durbin/Holmes  optimal  accuracy	algorithm is used, which finds
	      the alignment  that  maximizes  the  expected  accuracy  of  all
	      aligned residues.

       --wcx _x_
	      For each CM, set the W parameter,	the expected maximum length of
	      a	hit, to	_x_ times the consensus	length of the  model.  By  de-
	      fault,  the  W parameter is read from the	CM file	and was	calcu-
	      lated based on the transition probabilities of the model by  cm-
	      build.  You can find out what the	default	W is for a model using
	      cmstat.  This option should be used with caution as  it  impacts
	      the filtering pipeline at	several	different stages in nonobvious
	      ways. It is only recommended for expert users searching for hits
	      that  are	much longer than any of	the homologs used to build the
	      model in cmbuild,	e.g. ones with large introns  or  other	 large
	      insertions.   It cannot be used in combination with the --nohmm,
	      --fqdb or	--qdb options because in those cases W is  limited  by
	      query-dependent bands.

       --toponly
	      Only search the top (Watson) strand of target sequences in _seq-
	      file_.  By default, both strands are searched. This  will	 halve
	      the search space size (Z).

       --bottomonly
	      Only  search  the	 bottom	 (Crick) strand	of target sequences in
	      _seqfile_.  By default, both strands  are	 searched.  This  will
	      halve the	search space size (Z).

       --qformat _s_
	      Assert  that  the	query sequence database	file is	in format _s_.
	      Accepted formats include fasta, embl, genbank, ddbj,  stockholm,
	      pfam, a2m, afa, clustal, and phylip The default is to autodetect
	      the format of the	file.

       --glist _f_
	      Configure	a subset of models from	_cmfile_ in  glocal  alignment
	      mode,  instead  of  local	mode, namely the models	listed in file
	      _f_.  Configure all other	models (those not listed  in  _f_)  in
	      local mode.  This	option is incompatible with -g.	 File _f_ must
	      list valid names of models from _cmfile_,	each separated by  any
	      whitespace character (e.g. a newline character).

       --clanin	_f_
	      Read  clan  information on the models in _cmfile_	from file _f_.
	      Not all models in	_cmfile_ need to be a member of	a clan.	  This
	      option must be used in combination with --fmt 2 and --tblout be-
	      cause clan annotation is only output in format 2 of the  tabular
	      output file.  See	section	9 of the Infernal user guide for spec-
	      ifications on the	format of the clan input file _f_.

       --oclan
	      Only mark	overlaps between models	in the same clan.  This	option
	      must be used in combination with --fmt 2 , --tblout and --clanin
	      because clan annotation is only output in	format 2 of the	 tabu-
	      lar  output  file,  and clan information can only	be input using
	      the --clanin option.

       --oskip
	      Omit any hit h from the tabular output file that satisifies  the
	      following:  another hit h2 overlaps with h and the E-value of h2
	      is lower than that of h. Hit h will not appear  in  the  tabular
	      output  file,  although it will still exist in the standard out-
	      put.  This option	must be	 used  in  combination	with  --fmt  2
	      --tblout	because	 overlap annotation is only output in format 2
	      of the tabular output  file.   When  used	 in  combination  with
	      --oclan  only hits h that	satisfy	the following are omitted: an-
	      other hit	h2 overlaps with h, the	E-value	of h2  is  lower  than
	      that  of h, and both h and h2 are	hits to	models that are	in the
	      same clan.

       --cpu _n_
	      Set the number of	parallel worker	threads	to _n_.	  By  default,
	      Infernal sets this to the	number of CPU cores it detects in your
	      machine -	that is, it tries to maximize the use of  your	avail-
	      able  processor  cores.  Setting	_n_  higher than the number of
	      available	cores is of little if any value, but you may  want  to
	      set  it  to  something less. You can also	control	this number by
	      setting an environment variable, INFERNAL_NCPU.  This option  is
	      only  available if Infernal was compiled with POSIX threads sup-
	      port. This is the	default, but it	may have been  turned  off  at
	      compile-time for your site or machine for	some reason.

       --stall
	      For  debugging the MPI master/worker version: pause after	start,
	      to enable	the developer to attach	debuggers to the running  mas-
	      ter  and worker(s) processes. Send SIGCONT signal	to release the
	      pause.  (Under gdb: (gdb)	signal SIGCONT)	(Only available	if op-
	      tional MPI support was enabled at	compile-time.)

       --mpi  Run in MPI master/worker mode, using mpirun.  (Only available if
	      optional MPI support was enabled at compile-time.)

SEE ALSO
       See infernal(1) for a master man	page with a list of all	the individual
       man pages for programs in the Infernal package.

       For  complete documentation, see	the user guide that came with your In-
       fernal distribution (Userguide.pdf); or see the Infernal	web page ().

COPYRIGHT
       Copyright (C) 2019 Howard Hughes	Medical	Institute.
       Freely distributed under	the BSD	open source license.

       For additional information on copyright and  licensing,	see  the  file
       called  COPYRIGHT  in your Infernal source distribution,	or see the In-
       fernal web page ().

AUTHOR
       The Eddy/Rivas Laboratory
       Janelia Farm Research Campus
       19700 Helix Drive
       Ashburn VA 20147	USA
       http://eddylab.org

Infernal 1.1.3			   Nov 2019			     cmscan(1)

NAME | SYNOPSIS | DESCRIPTION | OPTIONS | OPTIONS FOR CONTROLLING OUTPUT | OPTIONS CONTROLLING REPORTING THRESHOLDS | OPTIONS FOR INCLUSION THRESHOLDS | OPTIONS FOR MODEL-SPECIFIC SCORE THRESHOLDING | OPTIONS CONTROLLING THE ACCELERATION PIPELINE | OTHER OPTIONS | SEE ALSO | COPYRIGHT | AUTHOR

Want to link to this manual page? Use this URL:
<https://www.freebsd.org/cgi/man.cgi?query=cmscan&sektion=1&manpath=FreeBSD+12.2-RELEASE+and+Ports>

home | help