annotate picard_CollectInsertSizeMetrics.xml @ 0:b76a4f17bbbb draft

Uploaded
author devteam
date Thu, 23 Oct 2014 11:31:30 -0400
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
1 <tool name="CollectInsertSizeMetrics" id="picard_CollectInsertSizeMetrics" version="1.122.0">
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
2 <description>plots distribution of insert sizes</description>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
3 <requirements>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
4 <requirement type="package" version="1.122.0">picard</requirement>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
5 </requirements>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
6
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
7 <macros>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
8 <import>picard_macros.xml</import>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
9 </macros>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
10
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
11 <command>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
12 @java_options@
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
13 ##set up input files
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
14
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
15 #set $reference_fasta_filename = "localref.fa"
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
16
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
17 #if str( $reference_source.reference_source_selector ) == "history":
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
18 ln -s "${reference_source.ref_file}" "${reference_fasta_filename}" &amp;&amp;
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
19 #else:
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
20 #set $reference_fasta_filename = str( $reference_source.ref_file.fields.path )
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
21 #end if
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
22
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
23 java -jar \$JAVA_JAR_PATH/CollectInsertSizeMetrics.jar
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
24 INPUT="${inputFile}"
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
25 OUTPUT="${outFile}"
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
26 HISTOGRAM_FILE="${histFile}"
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
27 DEVIATIONS="${deviations}"
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
28
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
29 #if str( $hist_width ):
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
30 HISTOGRAM_WIDTH="${hist_width}"
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
31 #end if
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
32
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
33 MINIMUM_PCT="${min_pct}"
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
34 REFERENCE_SEQUENCE="${reference_fasta_filename}"
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
35 ASSUME_SORTED="${assume_sorted}"
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
36 METRIC_ACCUMULATION_LEVEL="${metric_accumulation_level}"
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
37
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
38 VALIDATION_STRINGENCY="${validation_stringency}"
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
39 QUIET=true
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
40 VERBOSITY=ERROR
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
41
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
42 </command>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
43 <inputs>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
44 <param format="sam,bam" name="inputFile" type="data" label="Select SAM/BAM dataset or dataset collection" help="If empty, upload or import a SAM/BAM dataset."/>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
45 <conditional name="reference_source">
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
46 <param name="reference_source_selector" type="select" label="Load reference genome from">
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
47 <option value="cached">Local cache</option>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
48 <option value="history">History</option>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
49 </param>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
50 <when value="cached">
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
51 <param name="ref_file" type="select" label="Using reference genome" help="REFERENCE_SEQUENCE">
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
52 <options from_data_table="all_fasta">
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
53 </options>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
54 <validator type="no_options" message="A built-in reference genome is not available for the build associated with the selected input file"/>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
55 </param>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
56 </when>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
57 <when value="history">
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
58 <param name="ref_file" type="data" format="fasta" label="Use the folloing dataset as the reference sequence" help="REFERENCE_SEQUENCE; You can upload a FASTA sequence to the history and use it as reference" />
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
59 </when>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
60 </conditional>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
61 <param name="deviations" type="float" value="10.0" label="Generate mean, sd and plots by trimming the data down to MEDIAN + DEVIATIONS*MEDIAN_ABSOLUTE_DEVIATION" help="DEVIATIONS; This option is offered because insert size data typically includes enough anomalous values from chimeras and other artifacts to make the mean and SD grossly misleading regarding the real distribution. default=10.0"/>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
62 <param name="hist_width" type="integer" optional="True" label="Explicitly sets the Histogram width, overriding automatic truncation of Histogram tail" help="HISTOGRAM_WIDTH; optional"/>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
63 <param name="min_pct" type="float" value="0.05" label="When generating the Histogram, discard any data categories (out of FR, TANDEM, RF) that have fewer than this percentage of overall reads" help="MINIMUM_PCT; (Range: 0 to 1). default=0.05. "/>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
64 <param name="assume_sorted" type="boolean" label="Assume the input file is already sorted" checked="true" truevalue="true" falsevalue="false" help="ASSUME_SORTED"/>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
65 <param name="metric_accumulation_level" type="select" label="The level(s) at which to accumulate metrics" multiple="true" help="METRIC_ACCUMULATION_LEVEL">
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
66 <option value="ALL_READS" selected="True">All reads</option>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
67 <option value="SAMPLE">Sample</option>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
68 <option value="LIBRARY">Library</option>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
69 <option value="READ_GROUP">Read group</option>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
70 </param>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
71
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
72 <expand macro="VS" />
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
73
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
74 </inputs>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
75
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
76 <outputs>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
77 <data format="tabular" name="outFile"/>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
78 <data format="pdf" name="histFile"/>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
79 </outputs>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
80
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
81 <tests>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
82 <test>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
83 <param name="metric_accumulation_level" value="ALL_READS"/>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
84 <param name="deviations" value="10.0" />
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
85 <param name="hist_width" value="500" />
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
86 <param name="min_pct" value="0.05" />
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
87 <param name="assume_sorted" value="true" />
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
88 <param name="reference_source_selector" value="history" />
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
89 <param name="ref_file" value="picard_CollectInsertSizeMetrics_ref.fa" />
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
90 <param name="inputFile" value="picard_CollectInsertSizeMetrics.bam" ftype="bam" />
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
91 <output name="outFile" file="picard_CollectInsertSizeMetrics_test1.tab" ftype="tabular" lines_diff="4"/>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
92 </test>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
93 </tests>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
94
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
95 <stdio>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
96 <exit_code range="1:" level="fatal"/>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
97 </stdio>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
98
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
99 <help>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
100
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
101 .. class:: infomark
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
102
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
103 **Purpose**
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
104
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
105 Reads a SAM or BAM dataset and writes a file containing metrics about the statistical distribution of insert size (excluding duplicates) and generates a Histogram plot.
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
106
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
107 @dataset_collections@
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
108
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
109 @description@
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
110
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
111
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
112 DEVIATIONS=Double Generate mean, sd and plots by trimming the data down to MEDIAN +
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
113 DEVIATIONS*MEDIAN_ABSOLUTE_DEVIATION. This is done because insert size data typically
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
114 includes enough anomalous values from chimeras and other artifacts to make the mean and
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
115 sd grossly misleading regarding the real distribution. Default value: 10.0.
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
116
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
117 HISTOGRAM_WIDTH=Integer
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
118 W=Integer Explicitly sets the Histogram width, overriding automatic truncation of Histogram tail.
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
119 Also, when calculating mean and standard deviation, only bins &lt;= Histogram_WIDTH will be
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
120 included. Default value: not set.
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
121
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
122 MINIMUM_PCT=Float
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
123 M=Float When generating the Histogram, discard any data categories (out of FR, TANDEM, RF) that
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
124 have fewer than this percentage of overall reads. (Range: 0 to 1). Default value: 0.05.
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
125
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
126 METRIC_ACCUMULATION_LEVEL=MetricAccumulationLevel
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
127 LEVEL=MetricAccumulationLevel The level(s) at which to accumulate metrics. Possible values: {ALL_READS, SAMPLE,
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
128 LIBRARY, READ_GROUP} This option may be specified 0 or more times.
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
129
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
130 ASSUME_SORTED=Boolean
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
131 AS=Boolean If true (default), then the sort order in the header file will be ignored. Default
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
132 value: true. This option can be set to 'null' to clear the default value. Possible
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
133 values: {true, false}
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
134
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
135 @more_info@
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
136
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
137 </help>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
138 </tool>
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
139
b76a4f17bbbb Uploaded
devteam
parents:
diff changeset
140