<tool id="blockclust" name="BlockClust" version="1.0.0">
    <description>efficient clustering and classification of non-coding RNAs from short read RNA-seq profiles</description>
    <requirements>
        <requirement type="package" version="1.0">blockclust</requirement>
        <requirement type="package" version="1.1">eden</requirement>
        <requirement type="package" version="3.0.1">R_3_0_1</requirement>
        <requirement type="package" version="0.1.19">samtools</requirement>
        <requirement type="package" version="12.135">mcl</requirement>
        <requirement type="package" version="1.0">blockclust_rlibs</requirement>
    </requirements>
    <version_command>echo '1.0'</version_command>
    <command>
        #if str($tool_mode.operation) == "pre":
            BlockClustPipeLine.pl -m PRE -bam $tool_mode.reads_bam -tbed $tags_bed
        #elif str($tool_mode.operation) == "clust":
            #set $outputdir = $clusters.files_path
            #set $accept_bed=list()
            #set $reject_bed=list()
            ## prepare annotations
            #if str($tool_mode.reference) == "hg19":
                $accept_bed.append("\$BLOCKCLUST_DATA_PATH/annotations/hg19/hg19.accept.bed")
                $reject_bed.append("\$BLOCKCLUST_DATA_PATH/annotations/hg19/hg19.reject.bed")
            #elif str($tool_mode.reference) == "mm10":
                $accept_bed.append("\$BLOCKCLUST_DATA_PATH/annotations/mm10/mm10.accept.bed")
                $reject_bed.append("\$BLOCKCLUST_DATA_PATH/annotations/mm10/mm10.reject.bed")
            #elif str($tool_mode.reference) == "dm3":
                $accept_bed.append("\$BLOCKCLUST_DATA_PATH/annotations/dm3/dm3.accept.bed")
                $reject_bed.append("\$BLOCKCLUST_DATA_PATH/annotations/dm3/dm3.reject.bed")
                #elif str($tool_mode.reference) == "rheMac3":
                $accept_bed.append("\$BLOCKCLUST_DATA_PATH/annotations/rheMac3/rheMac3.accept.bed")
                $reject_bed.append("\$BLOCKCLUST_DATA_PATH/annotations/rheMac3/rheMac3.reject.bed")
            #elif str($tool_mode.reference) == "panTro4":
                $accept_bed.append("\$BLOCKCLUST_DATA_PATH/annotations/panTro3/panTro4.accept.bed")
                $reject_bed.append("\$BLOCKCLUST_DATA_PATH/annotations/panTro3/panTro4.reject.bed")
            #elif str($tool_mode.reference) == "xenTro3":
                $accept_bed.append("\$BLOCKCLUST_DATA_PATH/annotations/xenTro3/xenTro3.accept.bed")
                $reject_bed.append("\$BLOCKCLUST_DATA_PATH/annotations/xenTro3/xenTro3.reject.bed")
            #elif str($tool_mode.reference) == "celWS235":
                $accept_bed.append("\$BLOCKCLUST_DATA_PATH/annotations/celWS235/celWS235.accept.bed")
                $reject_bed.append("\$BLOCKCLUST_DATA_PATH/annotations/celWS235/celWS235.reject.bed")
            #elif str($tool_mode.reference) == "tair10":
                $accept_bed.append("\$BLOCKCLUST_DATA_PATH/annotations/tair10/tair10.accept.bed")
                $reject_bed.append("\$BLOCKCLUST_DATA_PATH/annotations/tair10/tair10.reject.bed")
            #end if
            BlockClustPipeLine.pl -m TEST -c \$BLOCKCLUST_DATA_PATH/blockclust.config
            -t $tool_mode.input_bbo
            -a #echo ''.join( $accept_bed )
            -r #echo ''.join( $reject_bed )
            -o $outputdir
            #if $tool_mode.nochr:
                -nochr
            #end if          
            #if str($tool_mode.pred.enable_pred) == "yes":
                -p
                -pm $tool_mode.pred.pred_mode
                -md \$BLOCKCLUST_DATA_PATH/models;
                #if str($tool_mode.pred.pred_mode) == "nearest_neighbour":
                    cp #echo os.path.join($outputdir,'nearest_neighbour_predictions.txt')# $nearest_neighbour_pred_bed;
                #elif str($tool_mode.pred.pred_mode) == "model_based":
                    cp #echo os.path.join($outputdir,'model_based_predictions.txt')# $model_based_pred_bed;
                #end if
            #else:
                ;
            #end if

            cp #echo os.path.join($outputdir, 'mcl_clusters','all_clusters.bed')# $clusters;
            cp #echo os.path.join($outputdir, 'hclust_tree.pdf')# $hclust_plot;
            cp #echo os.path.join($outputdir, 'discretized.gspan.tab')# $sim_tab_out
        #elif str($tool_mode.operation) == "post":
            BlockClustPipeLine.pl -m POST -cbed $tool_mode.clusters_bed -cm $tool_mode.cmsearch_out -tab $tool_mode.sim_tab_in -rfam \$BLOCKCLUST_DATA_PATH/rfam_map.txt -o ./;
        #end if
    </command>
    <inputs>
        <conditional name="tool_mode">
            <param name="operation" type="select" label="Select mode of operation">
                <option value="pre">Pre-processing </option>
                <option value="clust">Clustering and classification</option>
                <option value="post">Post-processing</option>
            </param>
            <when value="pre">
                <param name="reads_bam" type="data" format="bam" label="BAM file containing alignments" />
            </when>
            <when value="clust">
                <param name="input_bbo" type="data" format="tabular" label="Input blockgroups file" />
                <param name="reference" type="select" label="Select reference genome">
                    <option value="hg19">Human (hg19)</option>
                    <option value="mm10">Mouse (mm10)</option>
                    <option value="dm3">Fly (dm3)</option>
                    <option value="rheMac3">Monkey (rheMac3)</option>
                    <option value="panTro4">Chimp (panTro4)</option>
                    <option value="xenTro3">Frog (xenTro3)</option>
                    <option value="celWS235">C. elegans (celWS235)</option>
                    <option value="tair10">Arabidopsis thaliana (tair10)</option>
                </param>
                <param name="nochr" type="boolean" label="My input files have no 'chr' for chromosome names" checked="False"/>
                <conditional name="pred">
                    <param name="enable_pred" type="select" label="Would you like to perform classification?">
                        <option value="no">No</option>
                        <option value="yes">Yes</option>
                    </param>
                    <when value="yes">
                        <param name="pred_mode" type="select" label="Mode of classification">
                            <option value="model_based">Model based</option>
                            <option value="nearest_neighbour">Nearest neighbour</option>
                        </param>
                    </when>
                </conditional>
            </when>
            <when value="post">
                <param name="cmsearch_out" type="data" format="tabular" label="Output of cmsearch tool" />
                <param name="clusters_bed" type="data" format="bed" label="BED file containing clusters (output of BlockClust)" />
                <param name="sim_tab_in" type="data" format="tabular" label="Pairwise similarities file" />
            </when>
        </conditional>
    </inputs>

    <outputs>
        <data format="bed" name="tags_bed" label="BlockClust: BAM to BED on ${on_string}">
            <filter> tool_mode["operation"]=="pre"</filter>
        </data>
        <data format="pdf" name="hclust_plot" label="BlockClust: Hierarchical clustering plot on ${on_string}" >
            <filter> tool_mode["operation"]=="clust"</filter>
        </data>
        <data format="bed" name="clusters" label="BlockClust: BED of predicted clusters on ${on_string}">
            <filter> tool_mode["operation"]=="clust"</filter>
        </data>
        <data format="bed" name="model_based_pred_bed" label="BlockClust: Model based predictions BED on ${on_string}">
            <filter>
            ((
                tool_mode["operation"] == 'clust' and 
                tool_mode["pred"]["enable_pred"] == "yes" and
                tool_mode["pred"]["pred_mode"] == "model_based"
             ))
             </filter>
        </data>
        <data format="bed" name="nearest_neighbour_pred_bed" label="BlockClust: Nearest neighbor predictions BED on ${on_string}">
            <filter>
            ((
                tool_mode["operation"] == 'clust' and 
                tool_mode["pred"]["enable_pred"] == "yes" and
                tool_mode["pred"]["pred_mode"] == "nearest_neighbour"
             ))
             </filter>
        </data>
        <data format="tabular" name="sim_tab_out" label="BlockClust: Pairwise similarities on ${on_string}">
            <filter> tool_mode["operation"]=="clust"</filter>
        </data>
        <data format="pdf" name="cluster_dist" from_work_dir="cluster_distribution.pdf" label="BlockClust: Cluster distribution on ${on_string}" >
            <filter> tool_mode["operation"]=="post"</filter>
        </data>
        <data format="pdf" name="cluster_hclust" from_work_dir="hclust_tree_clusters.pdf" label="BlockClust: Hierarchical clustering plot of cluster centroids on ${on_string}" >
            <filter> tool_mode["operation"]=="post"</filter>
        </data>
    </outputs>
    <help>

.. class:: infomark

**What it does** 

BlockClust is an efficient approach to detect transcripts with similar 
processing patterns. We propose a novel way to encode expression profiles
in compact discrete structures, which can then be processed using 
fast graph-kernel techniques. BlockClust allows both clustering and 
classification of small non-coding RNAs.

BlockClust runs in three operating modes: 

1) Pre-processing - converts given mapped reads (BAM) into BED file of tags

2) Clustering and classification - of given input blockgroups (output of blockbuster tool) as explained in the original paper.

3) Post-processing - plots for overview of predicted clusters.

For a thorough analysis of your data, we suggest you to use complete blockclust workflow, which contains all three modes of operation.

**Inputs**

BlockClust input files are dependent on the mode of operation:

1. Pre-processing mode:
    * Binary Sequence Alignment Map (BAM) file

2. Clustering and classification:
    * A blockgroups file generated by blockbuster tool
    * Select reference genome

3. Post-processing:
    * Output of cmsearch, searched clusters generated by BlockClust against Rfam
    * BED file containing clusters generated by BlockClust
    * Pairwise similarities of blockgroups generated by BlockClust

**Outputs**

1. Pre-processing mode:
    * BED file of tags with expressions

2. Clustering and classification:
    * Hierarchical clustering plot of all input blockgroups by their similarity
    * Pairwise similarities of all input blockgroups
    * BED file containing predicted clusters
    * BED file containing prediction of blockgroups by pre-compiled SVM binary classification model.

3. Post-processing:
    * Plot of distribution of ncRNA families per predicted cluster (overview of cluster precissions). The annotation of ncRNA families are retrieved by searching cluster instances against Rfam database.
    * Hierarchical clustering made out of centroids of each BlockClust predicted cluster

------

**References** 

Pavankumar Videm, Dominic Rose, Fabrizio Costa, and Rolf Backofen. "BlockClust: efficient clustering and classification of non-coding RNAs from short read RNA-seq profiles." Bioinformatics 30, no. 12 (2014): i274-i282.


    </help>
</tool>