NAME¶

vmatchselect - sort and select matches

SYNOPSIS¶

vmatchselect [options] matchfile

DESCRIPTION¶

vmatchselect allows one to select interesting matches from the output of vmatch as specified by user-defined criteria. It delivers matches of chosen length, degeneracy or significance into further analysis routines.

vmatchselect removes from the input all those matches that are contained in another match. To do this efficiently, the matches are sorted by their position in the database sequence, and hence in the order in which the matches are output, unless the user specifies otherwise. Moreover, the sequences of the virtual suffix tree for which the match filewas produced can be clustered according to the matches. The input for vmatchselect is a file produced by vmatch, called a match file.

The output of vmatchselect goes to standard output and is sorted in ascending order of the positions of the left instance of a match. Two matches where the left instance occurs at the same position, are sorted in descending order of their length. Two matches of the same length where the left instance occurs in the same position, are sorted in ascending order of the position of the right instance of the match.

vmatchselect provides a subset of the options of vmatch. The main difference to vmatch is that vmatchselect gets the matches from a match file, while vmatch computes the matches from scratch. Therefore options specifying the index and/or the query sequences to be matched, as well as options specifying how to match are not available in vmatchselect. The options of vmatchselect have the same meaning as in the program vmatch. Thus, for a description, see the corresponding documentation. Note that vmatchselect also supports the option "-dbcluster". If vmatchselect is called with this option, then it parses the given match file and performs single linkage clustering based on the matches in this file. Thus vmatch and vmatchselect can perform hierarchical clustering. In a first step an initial set of matches with loose matching criteria is computed, using vmatch. Then one clusters these matches by calling vmatchselect. In a second round one applies more strict choices for the matches by the using the options "-l", "-leastscore", "-evalue", or "-identity", etc. This facilitates stepwise refinement of clusters without much computational effort and no new index construction for the sequence of a cluster. The output of vmatchselect is the same as the output of vmatch.

OPTIONS¶

-dbcluster <args>

-nonredundant

-selfun <file>

-l <length>

-leastscore <score>

-evalue <value>

-identity <value>

-sort <mode>

-best <n>

-s

-showdesc

-f

-absolute

-nodist

-noevalue

-noscore

-noidentity

-v

-version

-help

SEE ALSO¶

vmatch(1)