Mercurial > repos > artbio > small_rna_maps

changeset 7:a96e6a7df2b7 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/small_rna_maps commit 06472d1bd1365e4f7b385d578a69f4646481e22f
author artbio
date Tue, 10 Oct 2017 18:48:37 -0400
parents a3be3601bcb3
children 1827b74f872b
files small_rna_maps.r small_rna_maps.xml test-data/global_merge.pdf test-data/global_nomerge.pdf
diffstat 4 files changed, 103 insertions(+), 19 deletions(-) [+]
line wrap: on
line diff
--- a/small_rna_maps.r	Mon Oct 09 11:07:09 2017 -0400
+++ b/small_rna_maps.r	Tue Oct 10 18:48:37 2017 -0400
@@ -1,7 +1,7 @@
 ## Setup R error handling to go to stderr
 options( show.error.messages=F,
        error = function () { cat( geterrmessage(), file=stderr() ); q( "no", 1, F ) } )
-# options(warn = -1)
+options(warn = -1)
 library(RColorBrewer)
 library(lattice)
 library(latticeExtra)
@@ -9,12 +9,14 @@
 library(gridExtra)
 library(optparse)
 
+
 option_list <- list(
     make_option(c("-f", "--first_dataframe"), type="character", help="path to first dataframe"),
     make_option(c("-e", "--extra_dataframe"), type="character", help="path to additional dataframe"),
     make_option(c("-n", "--normalization"), type="character", help="space-separated normalization/size factors"),
     make_option("--first_plot_method", type = "character", help="How additional data should be plotted"),
     make_option("--extra_plot_method", type = "character", help="How additional data should be plotted"),
+    make_option("--global", type = "character", help="data should be plotted as global size distribution"),
     make_option("--output_pdf", type = "character", help="path to the pdf file with plots")
     )
  
@@ -37,11 +39,9 @@
 if (args$first_plot_method == "Counts" | args$first_plot_method == "Size" | args$first_plot_method == "Coverage") {
     i = 1
     for (sample in samples) {
-        print(norm_factors[i])
         Table[, length(Table)][Table$Dataset==sample] <- Table[, length(Table)][Table$Dataset==sample]*norm_factors[i]
         i = i + 1
     }
-    print(tail(Table))
 }
 genes=unique(levels(Table$Chromosome))
 per_gene_readmap=lapply(genes, function(x) subset(Table, Chromosome==x))
@@ -64,7 +64,42 @@
 }
 
 ## functions
-
+globalbc = function(df, global="", ...) {
+    if (global == "yes") {
+        bc <- barchart(Counts~as.factor(Size)|factor(Dataset, levels=unique(Dataset)),
+              data = df, origin = 0,
+              horizontal=FALSE,
+              col=c("darkblue"),
+              scales=list(y=list(tick.number=4, rot=90, relation="free", cex=0.5, alternating=T), x=list(rot=0, cex=0.6, tck=0.5, alternating=c(3,3))),
+              xlab=list(label=bottom_first_method[[args$first_plot_method]], cex=.85),
+              ylab=list(label=legend_first_method[[args$first_plot_method]], cex=.85),
+              main=title_first_method[[args$first_plot_method]],
+              layout = c(2, 6), newpage=T,
+              as.table=TRUE,
+              aspect=0.5,
+              strip = strip.custom(par.strip.text = list(cex = 1), which.given=1, bg="lightblue"),
+              ...
+              )
+    } else {
+        bc <- barchart(Counts~as.factor(Size)|factor(Dataset, levels=unique(Dataset)),
+              data = df, origin = 0,
+              horizontal=FALSE,
+              group=Polarity,
+              stack=TRUE,
+              col=c('red', 'blue'),
+              scales=list(y=list(tick.number=4, rot=90, relation="free", cex=0.5, alternating=T), x=list(rot=0, cex=0.6, tck=0.5, alternating=c(3,3))),
+              xlab=list(label=bottom_first_method[[args$first_plot_method]], cex=.85),
+              ylab=list(label=legend_first_method[[args$first_plot_method]], cex=.85),
+              main=title_first_method[[args$first_plot_method]],
+              layout = c(2, 6), newpage=T,
+              as.table=TRUE,
+              aspect=0.5,
+              strip = strip.custom(par.strip.text = list(cex = 1), which.given=1, bg="lightblue"),
+              ...
+              )
+    }
+   return(bc)
+}
 plot_unit = function(df, method=args$first_plot_method, ...) {
     if (method == 'Counts') {
         p = xyplot(Counts~Coordinate|factor(Dataset, levels=unique(Dataset))+factor(Chromosome, levels=unique(Chromosome)),
@@ -161,10 +196,8 @@
                     par.strip.text = list(cex=0.7),
                     nrow = 8,
                     as.table=TRUE,
-                    
                     ...)
           p = update(useOuterStrips(p, strip.left=strip.custom(par.strip.text = list(cex=0.5))), layout=c(n_samples, rows_per_page))
-          
           p = combineLimits(p, extend=TRUE)
           return (p)
         }
@@ -228,9 +261,26 @@
 # main
 
 if (args$extra_plot_method != '') { double_plot() }
-if (args$extra_plot_method == '') {
+if (args$extra_plot_method == '' & !exists('global', where=args)) {
     single_plot()
 }
+if (exists('global', where=args)) {
+    pdf(file=args$output, paper="special", height=11.69)
+    Table <- within(Table, Counts[Polarity=="R"] <- abs(Counts[Polarity=="R"])) # retropedalage
+    library(reshape2)
+    ml = melt(Table, id.vars = c("Dataset", "Chromosome", "Polarity", "Size"))
+    if (args$global == "nomerge") {
+        castml = dcast(ml, Dataset+Polarity+Size ~ variable, function(x) sum(x))
+        castml <- within(castml, Counts[Polarity=="R"] <- (Counts[Polarity=="R"]*-1))
+        bc = globalbc(castml, global="no")
+    } else {
+        castml = dcast(ml, Dataset+Size ~ variable, function(x) sum(x))
+        bc = globalbc(castml, global="yes")
+    }
+    plot(bc)
+    devname=dev.off()
+}
+    
 
 
 
--- a/small_rna_maps.xml	Mon Oct 09 11:07:09 2017 -0400
+++ b/small_rna_maps.xml	Tue Oct 10 18:48:37 2017 -0400
@@ -1,4 +1,4 @@
-<tool id="small_rna_maps" name="small_rna_maps" version="2.1.0">
+<tool id="small_rna_maps" name="small_rna_maps" version="2.2.0">
   <description></description>
   <requirements>
         <requirement type="package" version="1.11.2=py27_0">numpy</requirement>
@@ -6,6 +6,8 @@
         <requirement type="package" version="1.3.2=r3.3.2_0">r-optparse</requirement>
         <requirement type="package" version="0.6_28=r3.3.2_0">r-latticeextra</requirement>
         <requirement type="package" version="2.2.1=r3.3.2_0">r-gridextra</requirement>
+        <requirement type="package" version="1.4.2=r3.3.2_0">r-reshape2</requirement>
+        
   </requirements>
   <stdio>
       <exit_code range="1:" level="fatal" description="Tool exception" />
@@ -26,6 +28,9 @@
       #if str($plots_options.plots_options_selector ) == "two_plot":
           --plot_methods '${plots_options.first_plot}' '${plots_options.extra_plot}'
           --outputs '$output_tab' '$extra_output_tab' &&
+      #elif str($plots_options.plots_options_selector ) == "global":
+          --plot_methods 'Size'
+          --outputs '$output_tab' &&
       #else:
           --plot_methods '${plots_options.first_plot}'
           --outputs '$output_tab' &&
@@ -37,30 +42,35 @@
           #if str($plots_options.plots_options_selector ) == "two_plot":
               --first_plot_method '${plots_options.first_plot}'
               --extra_plot_method '${plots_options.extra_plot}'
+          #elif str($plots_options.plots_options_selector ) == "global":
+              --first_plot_method 'Size'
+              --extra_plot_method ''
+              --global '${plots_options.mergestrands}'
           #else:
               --first_plot_method '${plots_options.first_plot}'
               --extra_plot_method ''
           #end if
-          --output_pdf '$output_pdf'
+              --output_pdf '$output_pdf'
   ]]></command>
  <inputs>
     <param name="inputs" type="data" format="bam" label="Select multiple alignments to parse" multiple="True"/>
     <param name="normalization" type="text" label="Enter size/normalization factors as a space-separated list. Leave blank for no normalization (default)"
            help="e.g. '1 0.75 1.23'. Values of the list map to the above selected files from bottom to top" />
     <conditional name="plots_options">
-        <param name="plots_options_selector" type="select" label="Number of plots per chromosome">
+        <param name="plots_options_selector" type="select" display="radio" label="Plot Options">
             <option value="one_plot">Just one plot per chromosome</option>
             <option value="two_plot" selected="True">Two plots per chromosome</option>
+            <option value="global">Global read size distributions of aligned reads</option> 
         </param>
         <when value="two_plot">
-            <param name="first_plot" type="select" label="Select the type of the first plot">
+            <param name="first_plot" type="select" display="radio" label="Select the type of the first plot">
                 <option value="Counts">Counts</option> 
                 <option value="Coverage">Coverage</option> 
                 <option value="Mean">Mean Sizes</option> 
                 <option value="Median">Median Sizes</option>
                 <option value="Size">Size Distributions</option>
             </param>
-            <param name="extra_plot" type="select" label="Select the type of the second plot">
+            <param name="extra_plot" type="select" display="radio" label="Select the type of the second plot">
                 <option value="Counts">Counts</option> 
                 <option value="Coverage">Coverage</option> 
                 <option value="Mean">Mean Sizes</option> 
@@ -69,7 +79,7 @@
             </param>
         </when>
         <when value="one_plot">
-            <param name="first_plot" type="select" label="select the type of the first plot">
+            <param name="first_plot" type="select" display="radio" label="select the type of the first plot">
                 <option value="Counts">Counts</option> 
                 <option value="Coverage">Coverage</option> 
                 <option value="Mean">Mean Sizes</option> 
@@ -77,21 +87,45 @@
                 <option value="Size">Size Distributions</option>
             </param>
         </when>
+        <when value="global">
+            <param name="first_plot" type="hidden" value="Size"/>
+            <param name="mergestrands" type="select" display="radio" label="Whether forward and reverse aligned reads should be merged or not in the histogram">
+                <option value="nomerge">Do not merge</option>
+                <option value="merge">Merge forward and reverse reads</option>
+            </param>
+        </when>
     </conditional>
  </inputs>
 
  <outputs>
-   <data format="tabular" name="output_tab" label="Read Count dataframe" />
-   <data format="tabular" name="extra_output_tab" label="$plots_options.extra_plot dataframe">
-    <filter>plots_options['plots_options_selector'] == 'two_plot'</filter>
-   </data> 
-   <data format="pdf" name="output_pdf" label="PDF file" />
-
+    <data format="tabular" name="output_tab" label="$plots_options.first_plot dataframe" />
+    <data format="tabular" name="extra_output_tab" label="$plots_options.extra_plot dataframe">
+        <filter>plots_options['plots_options_selector'] == 'two_plot'</filter>
+    </data> 
+    <data format="pdf" name="output_pdf" label="small RNA maps" />
 </outputs>
 
     <tests>
         <test>
             <param name="inputs" value="input1.bam,input2.bam" ftype="bam"/>
+            <param name="plots_options_selector" value="global" />
+            <param name="mergestrands" value="nomerge" />
+            <param name="first_plot" value="Size" />
+            <param name="normalization" value="1 1" />
+            <output file="size.tab" name="output_tab" />
+            <output file="global_nomerge.pdf" name="output_pdf" />
+        </test>
+        <test>
+            <param name="inputs" value="input1.bam,input2.bam" ftype="bam"/>
+            <param name="plots_options_selector" value="global" />
+            <param name="mergestrands" value="merge" />
+            <param name="first_plot" value="Size" />
+            <param name="normalization" value="1 1" />
+            <output file="size.tab" name="output_tab" />
+            <output file="global_merge.pdf" name="output_pdf" />
+        </test>
+        <test>
+            <param name="inputs" value="input1.bam,input2.bam" ftype="bam"/>
             <param name="plots_options_selector" value="two_plot" />
             <param name="first_plot" value="Counts" />
             <param name="extra_plot" value="Mean" />
Binary file test-data/global_merge.pdf has changed
Binary file test-data/global_nomerge.pdf has changed