annotate heatmap.xml @ 2:17f627ee8f5a draft

"planemo upload for repository https://github.com/artbio/tools-artbio/tree/master/tools/high_dim_heatmap commit ba6089071a53fbf08df00c0f94dfc8515b7c858b"

<tool id="high_dim_heatmap" name="Plot heatmap" version="1.2.0">
    <description>with high number of rows</description>
    <requirements>
        <requirement type="package" version="3.0.1">r-gplots</requirement>
        <requirement type="package" version="1.1_2">r-rcolorbrewer</requirement>
    </requirements>
    <command detect_errors="exit_code"><![CDATA[
cat '$script' &&
Rscript '$script'
    ]]></command>
    <configfiles>
        <configfile name="script"><![CDATA[
## Setup R error handling to go to stderr
options(show.error.messages=F, error=function(){cat(geterrmessage(), file=stderr()); q("no",1,F)})

## Unify locale settings
loc <- Sys.setlocale("LC_MESSAGES", "en_US.UTF-8")

## Import library
library("RColorBrewer")
library("gplots")

input <- read.delim('$input1', sep='\t', header=TRUE)

mat_input <- data.matrix(input[, 2:ncol(input)])
rownames(mat_input) <- input[, 1]

#if $transform == "none"
    linput <- mat_input
#elif $transform == "log2"
    linput <- log2(mat_input)
#elif $transform == "log2plus1"
    linput <- log2(mat_input + 1)
#elif $transform == "log10"
    linput <- log10(mat_input)
#elif $transform == "log10plus1"
    linput <- log10(mat_input + 1)
#end if

#if $colorscheme == "whrd"
    colfunc <- colorRampPalette(c("white", "red"))
#elif $colorscheme == "whblu"
    colfunc <- colorRampPalette(c("white", "blue"))
#elif $colorscheme == "blwhre"
    colfunc <- colorRampPalette(c("blue","white", "red"))
#elif $colorscheme == "yeblu"
    colfunc <- colorRampPalette(c("lightyellow", "royalblue4"))
#end if

#if $labels== "both"
    rlabs = NULL
    clabs = NULL
#elif $labels== "rows"
    rlabs = NULL
    clabs = FALSE
#elif $labels== "columns"
    rlabs = FALSE
    clabs = NULL
#elif $labels== "none"
    rlabs = FALSE
    clabs = FALSE
#end if

pdf(file = '$output1', paper = 'special', height = $height)
par(cex.main = $title_size)

#if $cluster_cond.cluster:
    hclust_fun = function(x) hclust(x, method = '$cluster_cond.clustering')
    dist_fun = function(x) dist(x, method = '$cluster_cond.distance')
    #if $cluster_cond.cluster_cols_rows == "rows":
        heatmap.2(linput, dendrogram="row", Colv = FALSE, Rowv = TRUE,
            distfun = dist_fun, hclustfun = hclust_fun, scale = '$scale', labRow = rlabs,
            labCol = clabs, col = colfunc(50), trace = "none", density.info = "none",
            margins = c(8,8), main = '$title', key.xlab = '$key', keysize = 1,
            cexCol = 0.8, cexRow = $rowfontsize, srtCol = 45, lhei = c(1, $lineheight))
    #elif $cluster_cond.cluster_cols_rows == "columns":
        heatmap.2(linput, dendrogram = "column", Colv = TRUE, Rowv = FALSE,
            distfun = dist_fun, hclustfun = hclust_fun, scale = '$scale', labRow = rlabs,
            labCol = clabs, col = colfunc(50), trace = "none", density.info = "none",
            margins = c(8,8), main = '$title', key.xlab = '$key', keysize = 1,
            cexCol = 0.8, cexRow = $rowfontsize, srtCol = 45, lhei = c(1, $lineheight))
    #else
        heatmap.2(linput,
            distfun = dist_fun, hclustfun = hclust_fun, scale = '$scale', labRow = rlabs,
            labCol = clabs, col = colfunc(50), trace = "none", density.info = "none",
            margins = c(8,8), main = '$title', key.xlab = '$key', keysize = 1,
            cexCol = 0.8, cexRow = $rowfontsize, srtCol = 45, lhei = c(1, $lineheight))
    #end if
#else
    heatmap.2(linput,
        dendrogram = "none", Colv = FALSE, Rowv = FALSE, scale = '$scale', labRow = rlabs,
        labCol = clabs, col = colfunc(50), trace="none", density.info = "none",
        margins = c(8,8), main = '$title', key.xlab = '$key', keysize = 1, cexCol = 0.8,
        cexRow = $rowfontsize, srtCol = 45, lhei = c(1, $lineheight))
#end if

dev.off()
        ]]></configfile>
    </configfiles>

    <inputs>
        <param name="input1" type="data" format="tabular" label="Input should have column headers - these will be the columns that are plotted"/>
        <param name="title" type="text" format="txt" label="Plot title"/>
        <param name="key" type="text" format="txt" label="key title"/>
        <param name="transform" type="select" label="Data transformation">
            <option value="none">Plot the data as it is</option>
            <option value="log2">Log2(value) transform my data</option>
            <option value="log2plus1">Log2(value+1) transform my data</option>
            <option value="log10">Log10(value) transform my data</option>
            <option value="log10plus1">Log10(value+1) transform my data</option>
        </param>
        <conditional name="cluster_cond">
            <param name="cluster" type="select" label="Enable data clustering">
                <option value="true">Yes</option>
                <option value="">No</option>
            </param>
            <when value="true">
                <param name="cluster_cols_rows" type="select" label="Clustering columns and rows">
                    <option value="both" selected="true">Cluster rows and columns</option>
                    <option value="rows">Cluster rows and not columns</option>
                    <option value="columns">Cluster columns and not rows</option>
                </param>
                <param name="distance" type="select" label="Distance method" help="The method used to compute the distance (dissimilarity) between both rows and columns. Default: Euclidean" >
                    <option value="euclidean" selected="true">Euclidean</option>
                    <option value="maximum">Maximum</option>
                </param>
                <param name="clustering" type="select" label="Clustering method" help="The method used to compute the hierarchical clustering. Default: Complete" >
                    <option value="complete" selected="true">Complete</option>
                    <option value="average">Average</option>
                </param>
            </when>
            <when value=""/>
        </conditional>
        <param name="labels" type="select" label="Labeling columns and rows" >
            <option value="both" selected="true">Label my columns and rows</option>
            <option value="rows">Label rows and not columns</option>
            <option value="columns">Label columns and not rows</option>
            <option value="none">Do not label rows or columns</option>
        </param>
        <param name="colorscheme" type="select" label="Coloring groups" >
            <option value="whrd" selected="true">White to red</option>
            <option value="whblu">White to blue</option>
            <option value="yeblu">Yellow to blue</option>
            <option value="blwhre">Blue to white to red</option>
        </param>
        <param name="scale" type="select" label="Data scaling" >
            <option value="none" selected="true">Do not scale my data</option>
            <option value="row">Scale my data by row</option>
            <option value="column">Scale my data by column</option>
        </param>

        <param name="title_size" type="float" label="tweak main title size" value="1" />
        <param name="height" type="float" label="tweak plot height" value="7" />
        <param name="rowfontsize" type="float" label="tweak row label size" value="0.8" />
        <param name="lineheight" type="integer" label="tweak line height" value="4" />
    </inputs>
    <outputs>
        <data name="output1" format="pdf"/>
    </outputs>
    <tests>
        <test>
            <param name="input1" value="mtcars.txt"/>
            <param name="title" value="This is a pretty long title"/>
            <param name="cluster" value="true"/>
            <output name="output1" file="result1.pdf"/>
        </test>
        <test>
            <param name="input1" value="mtcars.txt"/>
            <param name="title" value="This is a pretty long title"/>
            <param name="cluster" value=""/>
            <output name="output1" file="result2.pdf"/>
        </test>
        <test>
            <param name="input1" value="complex_table.tsv"/>
            <param name="title" value="This is a pretty long title for a complex plot"/>
            <param name="title_size" value="0.7"/>
            <param name="transform" value="log2plus1"/>
            <param name="cluster" value=""/>
            <param name="colorscheme" value="blwhre"/>
            <param name="height" value="21"/>
            <param name="rowfontsize" value="0.4"/>
            <param name="lineheight" value="15"/>
            <output name="output1" file="complex_result1.pdf"/>
        </test>
    </tests>
    <help><![CDATA[
This tool employs the heatmap.2 function from the R gplots package and will generate a
heatmap of your data. If clustering is enabled, the heatmap uses the Euclidean distance
method and the Complete hierarchical clustering method by default.

Input data should have row labels in the first column and column labels. For example,
the row labels (the first column) should represent gene IDs and the column labels should
represent sample IDs.

To adapt heatmap.2 to high number of lines, tweak the parameters:

- `tweak plot height`. By default 7 (inches), this can be increased to 14, 21, 28, etc..
- `tweak row label size`. By default 0.8. This is usually diminished to 0.3, 0.1, etc to
  adapt the label font to high number of rows
- `tweak line height`. By default 4. This is a metric to define the relative size of the
  heatmap panel in the plot layout. This is a bit tricky to adjust, but should be adapted
  when plot height is increased (typically to 8, 10, 15, etc...)

For instance, upload the dataset
https://raw.githubusercontent.com/ARTbio/tools-artbio/master/tools/high_dim_heatmap/test-data/complex_table.tsv
to your account and test this tools with:

- `tweak plot height` : 21.0
- `tweak row label size`: 0.4
- `tweak line height`: 14


    ]]></help>
    <citations>
    </citations>
</tool>