view rapidcluster.xml @ 1:0d1cf9024f0e draft

Uploaded

<tool id="rapidcluster_2" name="RapidCluster" version="2">

	<description>Cluster closely-related sequences using Levenshtein edit distance filtering.</description>
	
	<version_command>rapidcluster -v</version_command>
	
	<command interpreter="perl">rapidcluster -i $input -o $output -d $distance -f $filter -c $max_clusters > $report
	</command>

	<inputs>
		<param name="input" type="data" format="fasta" label="Input file"  help="Must use FASTA output from FASTAptamer-Count"></param>
		<param name="distance" type="integer" label="Levenshtein Edit Distance" value="1" help="Minimum number of insertions, deletions, or substitutions required to transfer a sequence into another"></param>
		<param name="filter" type="integer" label="Read Filter" optional="true" value="1" help="Only sequences with total reads greater than the value supplied will be clustered."></param> 
		<param name="max_clusters" type="integer" label="Maximum number of clusters to find" optional="true" value="500" help="Script will stop after finding this much clusters"></param> 
 	</inputs>
 	
    <outputs>
    	<data name="output" format="fasta" label="$input RapidCluster output"></data>
	<data name="report" format="txt" label="$input RapidCluster Report"></data>
 	</outputs>
 	
    <help>

.. class:: warningmark

RapidCluster requires a FASTA formatted input file generated by FASTAptamer-Count.

.. class:: warningmark

RapidCluster uses an exhaustive approach to clustering and can take *several* hours to process. For faster processing utilize the "Read Filter" option to exclude low read sequences and define a reasonable number of clusters to find.

------

This version does not calculate exact Levenshtein distance for each pair of sequences, instead it simply checks if this distance is lower or greater than user-defined value. This makes script much faster for clustering highly-similar sequences.

RapidCluster begins with the most abundant sequence in a population, referred to as the "seed sequence," and clusters with it every sequence in the file within an edit distance less than or equal to the specified edit distance (Cluster #1). The next most abundant unclustered sequence then serves as the next seed sequence for assembling the second cluster from the remaining sequences (Cluster #2), followed by the next most abundant unclustered sequence (Cluster #3), and so on. This process is iterated until every sequence is clustered.

Output is FASTA formatted with the following information on the FASTA identifier line:

	>Rank-Reads-RPM-Cluster#-RankWithinCluster-EditDistanceFromSeedSequence

.. class:: infomark

The "Read Filter" excludes from the clustering process sequences with a total number of reads less than or equal to the integer supplied. Because of the computational complexity of clustering large datasets, the default filter setting of 1 is designed to eliminate singleton sequences from clustering.

------


    </help>

	
</tool>