cardinal_preprocessing: preprocessing.xml comparison

comparison preprocessing.xml @ 13:6b36be80febb draft

"planemo upload for repository https://github.com/galaxyproteomics/tools-galaxyp/tree/master/tools/cardinal commit 39bd480e8813fa7a96b640150365577a69885d17-dirty"

author	galaxyp
date	Sun, 29 Nov 2020 23:45:08 +0000
parents	e0669b1854b1
children	accf9fb6ea01

comparison

equal deleted inserted replaced

-:e0669b1854b1
+:6b36be80febb
-<tool id="cardinal_preprocessing" name="MSI preprocessing" version="@VERSION@.1">
+<tool id="cardinal_preprocessing" name="MSI preprocessing" version="@VERSION@.2">
 <description>
 mass spectrometry imaging preprocessing
 </description>
 <macros>
 <import>macros.xml</import>
 @READING_MSIDATA_FULLY_COMPATIBLE@
 ## remove duplicated coordinates, otherwise peak picking and log2 transformation will fail
-msidata <- msidata[,!duplicated(coord(msidata)[,1:2])]
 ## set variable to False
 #set $used_peak_picking = False
 #set $used_peak_alignment = False
 #set $continuous_format = False
 #if str( $method.methods_conditional.methods_for_smoothing.smoothing_method) == 'gaussian':
 print('gaussian smoothing')
 msidata = smoothSignal(msidata, method="$method.methods_conditional.methods_for_smoothing.smoothing_method", window=$method.methods_conditional.window_smoothing, sd = $method.methods_conditional.methods_for_smoothing.sd_gaussian)
+msidata <- process(msidata, BPPARAM=MulticoreParam(workers=number_cpu))
 #elif str( $method.methods_conditional.methods_for_smoothing.smoothing_method) == 'sgolay':
 print('sgolay smoothing')
 msidata = smoothSignal(msidata, method="$method.methods_conditional.methods_for_smoothing.smoothing_method", window=$method.methods_conditional.window_smoothing, order = $method.methods_conditional.methods_for_smoothing.order_of_filters)
+msidata <- process(msidata, BPPARAM=MulticoreParam(workers=number_cpu))
 ## if selected replace negative intensities with zero
 #if $method.methods_conditional.methods_for_smoothing.replace_negatives:
-spectra(msidata)[spectra(msidata)<0] = 0
+## bring spectra matrix to disk
+spectra_df = as.matrix(spectra(msidata))
+spectra_df[spectra_df<0] = 0
+spectra(msidata) = spectra_df
 #end if
 #elif str($method.methods_conditional.methods_for_smoothing.smoothing_method) == 'ma':
 print('moving average smoothing')
 msidata = smoothSignal(msidata, method="$method.methods_conditional.methods_for_smoothing.smoothing_method", window=$method.methods_conditional.window_smoothing, coef = $method.methods_conditional.methods_for_smoothing.coefficients_ma_filter)
+msidata <- process(msidata, BPPARAM=MulticoreParam(workers=number_cpu))
 #end if
-msidata <- process(msidata, BPPARAM=MulticoreParam(workers=number_cpu))
 ############################### QC ###########################
 maxfeatures =nrow(msidata)
 pixelcount = ncol(msidata)
 ############################### Mz alignment ###########################
 #elif str( $method.methods_conditional.preprocessing_method ) == 'mz_alignment':
-print('M/z alignment')
+print('m/z alignment')
 ## M/z alignment
 #if str( $method.methods_conditional.mzalign_ref_type.align_reference_datatype) == 'align_table':
 reference_mz = read.delim("$method.methods_conditional.mzalign_ref_type.mz_tabular", header = $method.methods_conditional.mzalign_ref_type.feature_header, stringsAsFactors = FALSE)
 msidata = mzAlign(msidata, ref=reference_mz, tolerance = $method.methods_conditional.alignment_tol, units = "$method.methods_conditional.alignment_units", quantile = $method.methods_conditional.quantile, span = $method.methods_conditional.span)
 #elif str( $method.methods_conditional.mzalign_ref_type.align_reference_datatype) == 'align_noref':
-msidata = mzAlign(msidata,tolerance = $method.methods_conditional.alignment_tol, units = "$method.methods_conditional.alignment_units", , quantile = $method.methods_conditional.quantile, span = $method.methods_conditional.span)
+msidata = mzAlign(msidata,tolerance = $method.methods_conditional.alignment_tol, units = "$method.methods_conditional.alignment_units", quantile = $method.methods_conditional.quantile, span = $method.methods_conditional.span)
 #end if
 msidata <- process(msidata, BPPARAM=MulticoreParam(workers=number_cpu))
 ############################### Mass binning ###########################
 #elif str( $method.methods_conditional.preprocessing_method) == 'Mass_binning':
 print('mass binning')
-#if str( $method.methods_conditional.mz_range.features_filtering) == 'change_mz_range':
+#if str($method.methods_conditional.mz_range.features_filtering) == 'change_mz_range':
-#if str($processed_cond.processed_file) == "processed":
 msidata = mse_bin = mzBin(msidata,resolution=$method.methods_conditional.bin_width, from=$method.methods_conditional.mz_range.min_mz, to=$method.methods_conditional.mz_range.max_mz, units="$method.methods_conditional.bin_units", fun="$method.methods_conditional.bin_fun")
-#else
-## continuous file cannot be binned from m/z to m/z, therefore first cut m/z range and then do mzbin:
-msidata = msidata[mz(msidata) >= $method.methods_conditional.mz_range.min_mz & mz(msidata) <= $method.methods_conditional.mz_range.max_mz,]
-msidata = mse_bin = mzBin(msidata,resolution=$method.methods_conditional.bin_width, units="$method.methods_conditional.bin_units", fun="$method.methods_conditional.bin_fun")
-#end if
-	    #elif str( $method.methods_conditional.mz_range.features_filtering) == 'none':
+	    #elif str($method.methods_conditional.mz_range.features_filtering) == 'none':
 msidata = mse_bin = mzBin(msidata,resolution=$method.methods_conditional.bin_width, units="$method.methods_conditional.bin_units", fun=$method.methods_conditional.bin_fun)
+	    #elif str($method.methods_conditional.mz_range.features_filtering) == 'bin_to_reference':
+	        bin_reference_mz = read.delim("$method.methods_conditional.mz_range.mz_tabular", header = $method.methods_conditional.mz_range.feature_header, stringsAsFactors = FALSE)
+bin_reference_mz = bin_reference_mz[,$method.methods_conditional.mz_range.feature_column]
+	       msidata = mse_bin = mzBin(msidata,resolution=$method.methods_conditional.bin_width, units="$method.methods_conditional.bin_units", fun=$method.methods_conditional.bin_fun,
+	       ref=bin_reference_mz)
 #end if
 msidata <- process(msidata, BPPARAM=MulticoreParam(workers=number_cpu))
 #if str( $method.methods_conditional.transf_conditional.trans_type) == 'log2':
 print('log2 transformation')
 ## replace 0 with NA to prevent Inf
-spectra_df = spectra(msidata) ## convert into R matrix
+spectra_df = as.matrix(spectra(msidata)) ## convert into R matrix
 spectra_df[spectra_df ==0] = NA
 print(paste0("Number of 0 which were converted into NA:",sum(is.na(spectra_df))))
 spectra(msidata) = spectra_df
 ## log transformation
 spectra(msidata) = log2(spectra(msidata))
 ############# Outputs: RData, imzml and QC report #############
 ################################################################################
 ## save msidata as imzML file, will only work if there is at least 1 m/z left
-#if str($imzml_output) == "cont_format":
-#set $continuous_format = True
-#end if
 if (nrow(msidata) > 0){
 ## make sure that coordinates are integers
 coord(msidata)\$y = as.integer(coord(msidata)\$y)
 coord(msidata)\$x = as.integer(coord(msidata)\$x)
-#if $used_peak_picking:
+## only continuous files can currently be exported
-#if $continuous_format:
+msidata = as(msidata, "MSContinuousImagingExperiment")
-msidata = as(msidata, "MSContinuousImagingExperiment")
+writeImzML(msidata, "out")
-#end if
-#elif $used_peak_alignment
-#if $continuous_format:
-msidata = as(msidata, "MSContinuousImagingExperiment")
-#end if
-#end if
-writeImzML(msidata, "out")
 }
 plot(0,type='n',axes=FALSE,ann=FALSE)
 rownames(QC_numbers) = c("min m/z", "max mz", "# features", "# spectra")
 grid.table(t(QC_numbers))
 <option value="mean" selected="True">mean</option>
 <option value="sum">sum</option>
 </param>
 <param name="replace_NA_bin" type="boolean" label="Replace NA with 0" truevalue="TRUE" falsevalue="FALSE" checked="True" help="Binning can introduce NAs, should they be replaced with 0"/>
 <conditional name="mz_range">
-		 <param name="features_filtering" type="select" label="Select m/z feature filtering option">
+	<param name="features_filtering" type="select" label="Select m/z options">
 <option value="none" selected="True">none</option>
 <option value="change_mz_range">change m/z range</option>
+<option value="bin_to_reference">bin m/z to reference</option>
 </param>
 			    <when value="none"/>
 			    <when value="change_mz_range">
 				<param name="min_mz" type="float" value="1" label="Minimum value for m/z"/>
 				<param name="max_mz" type="float" value="10000" label="Maximum value for m/z"/>
 			    </when>
+			    <when value="bin_to_reference">
+		                <expand macro="reading_1_column_mz_tabular" label="Tabular file with m/z features as reference for binning. Only the m/z values from the tabular file will be kept."/>
+	                   </when>
 </conditional>
 </when>
 <when value="Transformation">
 <conditional name="transf_conditional">
 <param name="trans_type" type="select" label="Intensity transformations" help="logarithm base 2 (log2) or squareroot (sqrt)">
 <when value="sqrt"/>
 </conditional>
 </when>
 </conditional>
 </repeat>
-<param name="imzml_output" type="boolean" label="imzML output in processed format" checked="True" truevalue="proc_format" falsevalue="cont_format"/>
 </inputs>
 <outputs>
 <data format="imzml" name="outfile_imzml" label="${tool.name} on ${on_string}: imzML"/>
 <data format="pdf" name="QC_overview" from_work_dir="Preprocessing.pdf" label = "${tool.name} on ${on_string}: QC"/>
 </outputs>
 </repeat>
 <repeat name="methods">
 <conditional name="methods_conditional">
 <param name="preprocessing_method" value="Smoothing"/>
 <conditional name="methods_for_smoothing">
-<param name="smoothing_method" value="gaussian"/>
+<param name="smoothing_method" value="sgolay"/>
-<param name="sd_gaussian" value="4"/>
+</conditional>
-</conditional>
-<param name="window_smoothing" value="9"/>
 </conditional>
 </repeat>
 <repeat name="methods">
 <conditional name="methods_conditional">
 <param name="preprocessing_method" value="Peak_picking"/>
 <conditional name="transf_conditional">
 <param name="trans_type" value="sqrt"/>
 </conditional>
 </conditional>
 </repeat>
-<param name="imzml_output" value="cont_format"/>
 <output name="QC_overview" file="preprocessing_results1.pdf" compare="sim_size"/>
 <output name="outfile_imzml" ftype="imzml" file="preprocessing_results1.imzml.txt" compare="sim_size">
 <extra_files type="file" file="preprocessing_results1.imzml" name="imzml" lines_diff="6"/>
 <extra_files type="file" file="preprocessing_results1.ibd" name="ibd" compare="sim_size"/>
 </output>
 <repeat name="methods">
 <conditional name="methods_conditional">
 <param name="preprocessing_method" value="Peak_alignment"/>
 </conditional>
 </repeat>
-<param name="imzml_output" value="cont_format"/>
 <output name="QC_overview" file="preprocessing_results2.pdf" compare="sim_size"/>
 <output name="outfile_imzml" ftype="imzml" file="preprocessing_results2.imzml.txt" compare="sim_size">
 <extra_files type="file" file="preprocessing_results2.imzml" name="imzml" lines_diff="6"/>
 <extra_files type="file" file="preprocessing_results2.ibd" name="ibd" compare="sim_size"/>
 </output>
 <param name="window_picking" value="5"/>
 <param name="SNR_picking_method" value="2"/>
 <conditional name="methods_for_picking">
 <param name="picking_method" value="mad"/>
 </conditional>
-<param name="imzml_output" value="proc_format"/>
 </conditional>
 </repeat>
 <repeat name="methods">
 <conditional name="methods_conditional">
 <param name="preprocessing_method" value="Peak_alignment"/>
-<param name="imzml_output" value="proc_format"/>
 </conditional>
 </repeat>
 <output name="QC_overview" file="preprocessing_results3.pdf" compare="sim_size"/>
 <output name="outfile_imzml" ftype="imzml" file="preprocessing_results3.imzml.txt" compare="sim_size">
 <extra_files type="file" file="preprocessing_results3.imzml" name="imzml" lines_diff="6"/>
 - Transformation: log2 or squareroot transformation of all intensities; when using log2 transformation zero intensities will become NA, this can lead to compatibility problems.
 **Output**
-- MSI data as continuous or processed imzML file
+- MSI data as continuous imzML file
 - pdf with key values and four random mass spectra after each processing step
 ]]>
 </help>
 <expand macro="citations"/>

Mercurial > repos > galaxyp > cardinal_preprocessing

comparison preprocessing.xml @ 13:6b36be80febb draft