cmfetch1

Usage

Overview

Retrieves one or more CMs from an <cmfile> (a large Rfam database, for example).

To enable very fast retrieval, index the <cmfile> first, using cmfetch --index. The index is a binary file named <cmfile>.ssi.

The default mode is to retrieve a single CM by name or accession, called the <key>. For example:

  % cmfetch Rfam.cm tRNA
  % cmfetch Rfam.cm RF00005

With the -f option, a <keyfile> containing a list of one or more keys is read instead. The first whitespace-delimited field on each non-blank non-comment line of the <keyfile> is used as a <key>, and any remaining data on the line is ignored. This allows a variety of whitespace delimited datafiles to be used as <keyfile>s.

When using -f and a <keyfile>, if <cmfile> has been indexed, the keys are retrieved in the order they occur in the <keyfile>, but if <cmfile> isn't indexed, keys are retrieved in the order they occur in the <cmfile>. This is a side effect of an implementation that allows multiple keys to be retrieved even if the <cmfile> is a nonrewindable stream, like a standard input pipe.

In normal use (without --index or -f options), <cmfile> may be '-' (dash), which means reading input from stdin rather than a file. With the --index option, <cmfile> may not be '-'; it does not make sense to index a standard input stream. With the -f option, either <cmfile> or <keyfile> (but not both) may be '-'. It is often particularly useful to read <keyfile> from standard input, because this allows use to use arbitrary command line invocations to create a list of CM names or accessions, then fetch them all to a new file, just with one command.

By default, the CM is printed to standard output in Infernal-1.1 format.

Options

-h

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

-f

The second commandline argument is a <keyfile> instead of a single <key>. The first field on each line of the <keyfile> is used as a retrieval <key> (a CM name or accession). Blank lines and comment lines (that start with a # character) are ignored.

-o <f>

Output CM(s) to file <f> instead of to standard output.

-O

Output CM(s) to individual file(s) named <key> instead of standard output.

--index

Instead of retrieving one or more profiles from <cmfile>, index the <cmfile> for future retrievals. This creates a <cmfile>.ssi binary index file.