comparison dia_umpire_se.xml @ 1:2b785516abfc draft default tip

planemo upload for repository https://github.com/galaxyproject/tools-iuc/tree/master/tools/dia_umpire commit 2379480213ba2e084a93bf82052fac858ffd074f

0:22a1fa7d9d6a

    </macros>
    <expand macro="requirements" />
    <expand macro="stdio" />
    <command>
<![CDATA[
#import re
## want to save all outputs in a directory output.extra_files_path to be used by dia_umpire_quant
## Is file naming going to be a problem? May need to have a name param
#if $se_extraction_data:
#set se_params = $se_ser
#set $ser_dir = $se_ser.extra_files_path
mkdir $ser_dir
&& ln -s '$ser_dir' '$output_dir'
&& cat $se_config > $se_ser 
#else:
#set se_params = $params
mkdir '$output_dir'
&& cat $se_config > $se_params 
#end if
##
&& echo " " >> $se_params 
&& echo "Thread = \$GALAXY_SLOTS" >> $se_params
#if $input_prefix and len($input_prefix.strip()) > 0:
#set $input_path = str($output_dir) + '/' + $input_prefix.__str__ + '_rep' + str($i + 1) + '.mzXML' 
#else:
#set $input_path = str($output_dir) + '/' + $re.sub('\.[mM]\w+$','',$re.sub('[^-a-zA-Z0-9_.]','_',$input.name)) + '.mzXML'
#end if
&& ln -s '${input}' '$input_path'
&&  dia_umpire_se '$input_path' '$se_params'
&& cat $output_dir/*.log >> "$logfile"
#if not $mgfs_as_collection:

    <inputs>
        <param name="input" type="data" format="mzxml" label="Proteomics Spectrum  files in mzXML format"/>
        <param name="input_prefix" type="text" value="" optional="true" label="File name prefix" help="Names inputs: prefix_rep#.mzXML Leave blank to use History names of input">
          <validator type="regex" message="">[a-zA-Z][a-zA-Z0-9_-]*</validator>
        </param>
        <param name="output_dir" type="hidden" value="gx_path"/>

        <conditional name="instrument">
          <param name="model" type="select" label="instrument used" help="Sets default parameters">
            <option value="Thermo_Orbitrap">Thermo Orbitrap</option>
            <option value="AB_SCIEX_Triple_TOF_5600">AB SCIEX Triple TOF 5600</option>
            <option value="other">other</option>
          </param>
          <when value="Thermo_Orbitrap">
       
            <param name="SE_MS1PPM" type="float" value="5" min="1" max="20" optional="true" label="Maximum mass error for two MS1 peaks">
                <help>
SE.MS1PPM: (Unit: ppm) Maximum mass error for two MS1 peaks in consecutive spectra to be considered signal of the same ion. Used in MS1 signal detection and precursor alignment between samples/runs.
Recommended value: Depends on the instrument. Typical values are 5-10 ppm for Thermo Orbitrap.
                </help>

SE.MS2PPM: (Unit: ppm) Maximum mass error for two MS2 peaks in consecutive spectra to be considered signal of the same ion.
Recommended value: Depends on the instrument. If fragmentation spectra are measured with the same detector as MS1 spectra, set the same as Para.MS1PPM or a little higher, e.g. if you've set Para.MS1PPM=30 ppm for AB SCIEX Triple TOF 5600, consider setting to 40ppm.
                </help>
            </param>
            <expand macro="common_se_params" />

          </when>
          <when value="AB_SCIEX_Triple_TOF_5600">
            <param name="SE_MS1PPM" type="float" value="30" min="1" max="50" optional="true" label="Maximum mass error for two MS1 peaks">
                <help>
            SE.MS1PPM: (Unit: ppm) Maximum mass error for two MS1 peaks in consecutive spectra to be considered signal of the same ion. Used in MS1 signal detection and precursor alignment between samples/runs.

            <param name="RTOverlap" type="float" value=".3" min="0" optional="true" label="RTOverlap" >
              <help>
RTOverlap: Retention time overlap. (Default: 0.3)
              </help>
            </param>

            <param name="DeltaApex" type="float" value=".6" min="0" optional="true" label="DeltaApex" >
              <help>
DeltaApex: (Unit: minute) Maximum retention time difference of LC profile apexes between precursor and fragment (the lower, the more stringent). (Default: 0.6)
              </help>
            </param>

            <option value="no">no</option>
            <option value="yes">yes</option>
          </param>
          <when value="no"/>
          <when value="yes">

            <param name="SE_MinMSIntensity" type="float" value="" optional="true" label="MinMSIntensity" >
              <help>
SE.MinMSIntensity: Minimum signal intensity for a peak in an MS1 spectrum to be considered as a valid signal. Any MS1 peak having intensity lower than this threshold will be ignored. It is the main parameter controlling how many peaks and isotopic envelopes will be detected.
Recommended value: Depends on the data. Check raw data for average noise- levels. E.g. TOF data often have thousands of random small intensity peaks. 
Warning: Setting this parameter too low (or zero) in such a case will significantly increase processing time and memory requirements.

              </help>
            </param>

            <param name="SE_MinRTRange" type="float" value="" optional="true" label="MinRTRange" >
              <help>

              </help>
            </param>
            <param name="SE_MaxNoPeakCluster" type="integer" value="" optional="true" label="MaxNoPeakCluster" >
              <help>
SE.MaxNoPeakCluster (new parameter in v1.4): Maximum number of isotope peaks for a precursor feature.

              </help>
            </param>
            <param name="SE_MinMS2NoPeakCluster" type="integer" value="" optional="true" label="MinMS2NoPeakCluster" >
              <help>
SE.MinMS2NoPeakCluster (new parameter in v1.4): Minimum number of isotope peaks for a MS2 feature. When it is set as 1, the algorithm will group fragments even for peaks without any isotope signal being found. For these cases, the assumed charged states will be from the parameter SE.StartCharge to SE.EndCharge.

              </help>
            </param>
            <param name="SE_RTtol" type="float" value="" optional="true" label="RTtol" >
              <help>
              </help>

               label="ExportPrecursorPeak"
               help="Output detailed information about detected MS1 precursor and MS2 unfragmented precursor signals"/>
        <param name="ExportFragmentPeak" type="boolean" truevalue="true" falsevalue="false" checked="false" 
               label="ExportFragmentPeak"
               help="Output detailed information about detected MS2 signals"/>
        <param name="se_extraction_data" type="boolean" truevalue="Signal Extraction data" falsevalue="diaumpire_se.params" checked="false" 
               label="Output Signal Extraction data for DIA_Umpire_Quant" />
        <param name="mgfs_as_collection" type="boolean" truevalue="true" falsevalue="false" checked="false" 
               label="Output MGFs as a collection" />

    </inputs>

    <outputs>
        <data format="txt" name="logfile" label="${tool.name} ${on_string} log"/>
        <data format="dia_umpire.ser" name="se_ser" label="${tool.name} ${input.name} ${se_extraction_data}">
            <filter>se_extraction_data</filter>
        </data>
        <data format="txt" name="params" label="${tool.name} ${input.name} ${se_extraction_data}">
            <filter>not se_extraction_data</filter>
        </data>
        <data format="csv" name="PrecursorPeak" label="${tool.name} ${input.name} PeakCluster.csv" from_work_dir="gx_path/swath_PeakCurve.csv">
            <filter>ExportPrecursorPeak</filter>
        </data>
        <!--
        <data format="csv" name="FragmentPeak" label="" from_work_dir="gx_path/swath_PeakCurve.csv">

                </conditional>
            </conditional>
            <output name="q2_mgf">
                <assert_contents>
                    <has_text text="BEGIN IONS" />
                    <has_text_matching expression="^PEPMASS=740.\d+$" />
                </assert_contents>
            </output>
        </test>
    </tests>
    <help>

    2. <filename>_Q2.mgf
    3. <filename>_Q3.mgf

    Note: Each file corresponds to a different "quality level" of precursor ions (Q1= More than two isotopic peaks detected in MS1, Q2 = only two isotopic peak detected, Q3 = detected unfragmented precursor in MS2). These spectra are written to separate files, because they must be searched separately against a protein database as a consequence of differences in FDR estimates for these varying quality data.

  2. *DIA_Umpire_SE Signal Extraction data* - includes the binary files (.ser) containing contain all necessary information for quantitation procedures (parameter settings, all detected precursor and fragment peaks, precursor-fragment grouping information).  

  3. If ExportPrecursorPeak and/or ExportFragmentPeak options were set to true, text files with detailed information about detected MS1 and/or MS2 features will be generated.


]]>
    </help>
    <expand macro="citations" />

1:2b785516abfc

    </macros>
    <expand macro="requirements" />
    <expand macro="stdio" />
    <command>
<![CDATA[
#set $output_dir = 'gx_path'
#import re
mkdir '$output_dir'
&& cat $se_config > $se_params 
&& echo " " >> $se_params 
&& echo "Thread = \$GALAXY_SLOTS" >> $se_params
#if $input_prefix and len($input_prefix.strip()) > 0:
#set $input_path = str($output_dir) + '/' + $input_prefix.__str__ + '_rep' + str($i + 1) + '.mzXML' 
#else:
#set $input_path = str($output_dir) + '/' + $re.sub('\.[mM]\w+$','',$re.sub('[^-a-zA-Z0-9_.]','_',$input.element_identifier)) + '.mzXML'
#end if
&& ln -s '${input}' '$input_path'
&&  dia_umpire_se '$input_path' '$se_params'
&& cat $output_dir/*.log >> "$logfile"
#if not $mgfs_as_collection:

    <inputs>
        <param name="input" type="data" format="mzxml" label="Proteomics Spectrum  files in mzXML format"/>
        <param name="input_prefix" type="text" value="" optional="true" label="File name prefix" help="Names inputs: prefix_rep#.mzXML Leave blank to use History names of input">
          <validator type="regex" message="">[a-zA-Z][a-zA-Z0-9_-]*</validator>
        </param>
        <conditional name="instrument">
          <param name="model" type="select" label="instrument used" help="Sets default parameters">
            <option value="Thermo_Orbitrap">Thermo Orbitrap</option>
            <option value="AB_SCIEX_Triple_TOF_5600">AB SCIEX Triple TOF 5600</option>
            <option value="other">other</option>
          </param>
          <when value="Thermo_Orbitrap">
            <param name="SE_MS1PPM" type="float" value="5" min="1" max="20" optional="true" label="Maximum mass error for two MS1 peaks">
                <help>
SE.MS1PPM: (Unit: ppm) Maximum mass error for two MS1 peaks in consecutive spectra to be considered signal of the same ion. Used in MS1 signal detection and precursor alignment between samples/runs.
Recommended value: Depends on the instrument. Typical values are 5-10 ppm for Thermo Orbitrap.
                </help>

SE.MS2PPM: (Unit: ppm) Maximum mass error for two MS2 peaks in consecutive spectra to be considered signal of the same ion.
Recommended value: Depends on the instrument. If fragmentation spectra are measured with the same detector as MS1 spectra, set the same as Para.MS1PPM or a little higher, e.g. if you've set Para.MS1PPM=30 ppm for AB SCIEX Triple TOF 5600, consider setting to 40ppm.
                </help>
            </param>
            <expand macro="common_se_params" />
          </when>
          <when value="AB_SCIEX_Triple_TOF_5600">
            <param name="SE_MS1PPM" type="float" value="30" min="1" max="50" optional="true" label="Maximum mass error for two MS1 peaks">
                <help>
            SE.MS1PPM: (Unit: ppm) Maximum mass error for two MS1 peaks in consecutive spectra to be considered signal of the same ion. Used in MS1 signal detection and precursor alignment between samples/runs.

            <param name="RTOverlap" type="float" value=".3" min="0" optional="true" label="RTOverlap" >
              <help>
RTOverlap: Retention time overlap. (Default: 0.3)
              </help>
            </param>
            <param name="DeltaApex" type="float" value=".6" min="0" optional="true" label="DeltaApex" >
              <help>
DeltaApex: (Unit: minute) Maximum retention time difference of LC profile apexes between precursor and fragment (the lower, the more stringent). (Default: 0.6)
              </help>
            </param>

            <option value="no">no</option>
            <option value="yes">yes</option>
          </param>
          <when value="no"/>
          <when value="yes">
            <param name="SE_MinMSIntensity" type="float" value="" optional="true" label="MinMSIntensity" >
              <help>
SE.MinMSIntensity: Minimum signal intensity for a peak in an MS1 spectrum to be considered as a valid signal. Any MS1 peak having intensity lower than this threshold will be ignored. It is the main parameter controlling how many peaks and isotopic envelopes will be detected.
Recommended value: Depends on the data. Check raw data for average noise- levels. E.g. TOF data often have thousands of random small intensity peaks. 
Warning: Setting this parameter too low (or zero) in such a case will significantly increase processing time and memory requirements.

              </help>
            </param>

            <param name="SE_MinRTRange" type="float" value="" optional="true" label="MinRTRange" >
              <help>
              </help>
            </param>
            <param name="SE_MaxNoPeakCluster" type="integer" value="" optional="true" label="MaxNoPeakCluster" >
              <help>
SE.MaxNoPeakCluster (new parameter in v1.4): Maximum number of isotope peaks for a precursor feature.

              </help>
            </param>
            <param name="SE_MinMS2NoPeakCluster" type="integer" value="" optional="true" label="MinMS2NoPeakCluster" >
              <help>
SE.MinMS2NoPeakCluster (new parameter in v1.4): Minimum number of isotope peaks for a MS2 feature. When it is set as 1, the algorithm will group fragments even for peaks without any isotope signal being found. For these cases, the assumed charged states will be from the parameter SE.StartCharge to SE.EndCharge.
              </help>
            </param>
            <param name="SE_RTtol" type="float" value="" optional="true" label="RTtol" >
              <help>
              </help>

               label="ExportPrecursorPeak"
               help="Output detailed information about detected MS1 precursor and MS2 unfragmented precursor signals"/>
        <param name="ExportFragmentPeak" type="boolean" truevalue="true" falsevalue="false" checked="false" 
               label="ExportFragmentPeak"
               help="Output detailed information about detected MS2 signals"/>
        <param name="mgfs_as_collection" type="boolean" truevalue="true" falsevalue="false" checked="false" 
               label="Output MGFs as a collection" />

    </inputs>

    <outputs>
        <data format="txt" name="logfile" label="${tool.name} ${on_string} log"/>
        <data format="txt" name="se_params" label="${tool.name} ${input.name} diaumpire_se.params"/>
        <data format="csv" name="PrecursorPeak" label="${tool.name} ${input.name} PeakCluster.csv" from_work_dir="gx_path/swath_PeakCurve.csv">
            <filter>ExportPrecursorPeak</filter>
        </data>
        <!--
        <data format="csv" name="FragmentPeak" label="" from_work_dir="gx_path/swath_PeakCurve.csv">

                </conditional>
            </conditional>
            <output name="q2_mgf">
                <assert_contents>
                    <has_text text="BEGIN IONS" />
                    <has_text text="PEPMASS=740" />
                </assert_contents>
            </output>
        </test>
    </tests>
    <help>

    2. <filename>_Q2.mgf
    3. <filename>_Q3.mgf

    Note: Each file corresponds to a different "quality level" of precursor ions (Q1= More than two isotopic peaks detected in MS1, Q2 = only two isotopic peak detected, Q3 = detected unfragmented precursor in MS2). These spectra are written to separate files, because they must be searched separately against a protein database as a consequence of differences in FDR estimates for these varying quality data.

  2. If ExportPrecursorPeak and/or ExportFragmentPeak options were set to true, text files with detailed information about detected MS1 and/or MS2 features will be generated.


]]>
    </help>
    <expand macro="citations" />