From 718782e103e5e16abfd6936e7690856b2481f7bc Mon Sep 17 00:00:00 2001
From: Daniel Blankenberg <dan@bx.psu.edu>
Date: Tue, 24 Jan 2012 16:29:28 -0500
Subject: [PATCH] Add GATK Print Reads walker tool.

---
 tool_conf.xml.sample       |   1 +
 tools/gatk/print_reads.xml | 424 +++++++++++++++++++++++++++++++++++++
 2 files changed, 425 insertions(+)
 create mode 100644 tools/gatk/print_reads.xml

diff --git a/tool_conf.xml.sample b/tool_conf.xml.sample
index cdb4736d734..037fe8a5b36 100644
--- a/tool_conf.xml.sample
+++ b/tool_conf.xml.sample
@@ -385,6 +385,7 @@
   <section name="NGS: GATK Tools (beta)" id="gatk">
     <label text="Alignment Utilities" id="gatk_bam_utilities"/>
     <tool file="gatk/depth_of_coverage.xml" />
+    <tool file="gatk/print_reads.xml" />
     
     <label text="Realignment" id="gatk_realignment" />
     <tool file="gatk/realigner_target_creator.xml" />
diff --git a/tools/gatk/print_reads.xml b/tools/gatk/print_reads.xml
new file mode 100644
index 00000000000..64bef033bb1
--- /dev/null
+++ b/tools/gatk/print_reads.xml
@@ -0,0 +1,424 @@
+<tool id="gatk_print_reads" name="Print Reads" version="0.0.1">
+  <description>from BAM files</description>
+  <requirements>
+      <requirement type="package" version="1.4">gatk</requirement>
+  </requirements>
+  <command interpreter="python">gatk_wrapper.py
+   --max_jvm_heap_fraction "1"
+   --stdout "${output_log}"
+   #for $i, $input_bam in enumerate( $reference_source.input_bams ):
+       -d "-I" "${input_bam.input_bam}" "${input_bam.input_bam.ext}" "gatk_input_${i}"
+       -d "" "${input_bam.input_bam.metadata.bam_index}" "bam_index" "gatk_input_${i}" ##hardcode galaxy ext type as bam_index
+   #end for
+   -p 'java 
+    -jar "${GALAXY_DATA_INDEX_DIR}/shared/jars/gatk/GenomeAnalysisTK.jar"
+    -T "PrintReads"
+    ##--num_threads 4 ##hard coded, for now
+    --out "${output_bam}"
+    -et "NO_ET" ##ET no phone home
+    #if $reference_source.reference_source_selector != "history":
+        -R "${reference_source.ref_file.fields.path}"
+    #end if
+    --number "${number}"
+    #if $platform:
+        --platform "${platform}"
+    #end if
+    #if $read_group:
+        --readGroup "${read_group}"
+    #end if
+    #for $sample_file in $sample_file_repeat:
+        --sample_file "${sample_file.input_sample_file}"
+    #end for
+    #for $sample_name in $sample_name_repeat:
+        --sample_name "${sample_name.sample_name}"
+    #end for
+   '
+   
+    ##start standard gatk options
+    #if $gatk_param_type.gatk_param_type_selector == "advanced":
+        #for $pedigree in $gatk_param_type.pedigree:
+            -p '--pedigree "${pedigree.pedigree_file}"'
+        #end for
+        #for $pedigree_string in $gatk_param_type.pedigree_string_repeat:
+            -p '--pedigreeString "${pedigree_string.pedigree_string}"'
+        #end for
+        -p '--pedigreeValidationType "${gatk_param_type.pedigree_validation_type}"'
+        #for $read_filter in $gatk_param_type.read_filter:
+            -p '--read_filter "${read_filter.read_filter_type.read_filter_type_selector}"
+            ###raise Exception( str( dir( $read_filter ) ) )
+            #for $name, $param in $read_filter.read_filter_type.iteritems():
+                #if $name not in [ "__current_case__", "read_filter_type_selector" ]:
+                    #if hasattr( $param.input, 'truevalue' ):
+                        ${param}
+                    #else:
+                        --${name} "${param}"
+                    #end if
+                #end if
+            #end for
+            '
+        #end for
+        #for $interval_count, $input_intervals in enumerate( $gatk_param_type.input_interval_repeat ):
+            -d "--intervals" "${input_intervals.input_intervals}" "${input_intervals.input_intervals.ext}" "input_intervals_${interval_count}"
+        #end for
+        
+        #for $interval_count, $input_intervals in enumerate( $gatk_param_type.input_exclude_interval_repeat ):
+            -d "--excludeIntervals" "${input_intervals.input_exclude_intervals}" "${input_intervals.input_exclude_intervals.ext}" "input_exlude_intervals_${interval_count}"
+        #end for
+
+        -p '--interval_set_rule "${gatk_param_type.interval_set_rule}"'
+        
+        -p '--downsampling_type "${gatk_param_type.downsampling_type.downsampling_type_selector}"'
+        #if str( $gatk_param_type.downsampling_type.downsampling_type_selector ) != "NONE":
+            -p '--${gatk_param_type.downsampling_type.downsample_to_type.downsample_to_type_selector} "${gatk_param_type.downsampling_type.downsample_to_type.downsample_to_value}"'
+        #end if
+        -p '
+        --baq "${gatk_param_type.baq}"
+        --baqGapOpenPenalty "${gatk_param_type.baq_gap_open_penalty}"
+        ${gatk_param_type.use_original_qualities}
+        --defaultBaseQualities "${gatk_param_type.default_base_qualities}"
+        --validation_strictness "${gatk_param_type.validation_strictness}"
+        --interval_merging "${gatk_param_type.interval_merging}"
+        ${gatk_param_type.disable_experimental_low_memory_sharding}
+        ${gatk_param_type.non_deterministic_random_seed}
+        '
+        #for $rg_black_list_count, $rg_black_list in enumerate( $gatk_param_type.read_group_black_list_repeat ):
+            #if $rg_black_list.read_group_black_list_type.read_group_black_list_type_selector == "file":
+                -d "--read_group_black_list" "${rg_black_list.read_group_black_list_type.read_group_black_list}" "txt" "input_read_group_black_list_${rg_black_list_count}"
+            #else
+                -p '--read_group_black_list "${rg_black_list.read_group_black_list_type.read_group_black_list}"'
+            #end if
+        #end for
+    #end if
+    
+    #if $reference_source.reference_source_selector == "history":
+        -d "-R" "${reference_source.ref_file}" "${reference_source.ref_file.ext}" "gatk_input"
+    #end if
+    ##end standard gatk options
+    
+  </command>
+  <inputs>
+    <conditional name="reference_source">
+      <param name="reference_source_selector" type="select" label="Choose the source for the reference list">
+        <option value="cached">Locally cached</option>
+        <option value="history">History</option>
+      </param>
+      <when value="cached">
+        <repeat name="input_bams" title="Sample BAM file" min="1">
+            <param name="input_bam" type="data" format="bam" label="BAM file">
+              <validator type="unspecified_build" />
+              <validator type="metadata" check="bam_index" message="Metadata missing, click the pencil icon in the history item and use the auto-detect feature to correct this issue."/>
+              <validator type="dataset_metadata_in_data_table" table_name="gatk_picard_indexes" metadata_name="dbkey" metadata_column="dbkey" message="Sequences are not currently available for the specified build." /> <!-- fixme!!! this needs to be a select -->
+            </param>
+        </repeat>
+        <param name="ref_file" type="select" label="Using reference genome">
+          <options from_data_table="gatk_picard_indexes">
+            <!-- <filter type="data_meta" key="dbkey" ref="input_bam" column="dbkey"/> does not yet work in a repeat...--> 
+          </options>
+          <validator type="no_options" message="A built-in reference genome is not available for the build associated with the selected input file"/>
+        </param>
+      </when>
+      <when value="history"> <!-- FIX ME!!!! -->
+        <repeat name="input_bams" title="Sample BAM file" min="1">
+            <param name="input_bam" type="data" format="bam" label="BAM file" >
+              <validator type="metadata" check="bam_index" message="Metadata missing, click the pencil icon in the history item and use the auto-detect feature to correct this issue."/>
+            </param>
+        </repeat>
+        <param name="ref_file" type="data" format="fasta" label="Using reference file" />
+      </when>
+    </conditional>
+    
+    <param name="number" type="integer" value="-1" label="Print the first n reads from the file, discarding the rest" />
+    <param name="platform" type="text" value="" label="Exclude all reads with this platform from the output" />
+    <param name="read_group" type="text" value="" label="Exclude all reads with this read group from the output" />
+    <repeat name="sample_file_repeat" title="File containing a list of samples to include">
+        <param name="input_sample_file" type="data" format="text" label="Sample file" />
+    </repeat>
+    <repeat name="sample_name_repeat" title="Sample name to be included in the analysis">
+        <param name="sample_name" type="text" label="Sample name" />
+    </repeat>
+    
+    <conditional name="gatk_param_type">
+      <param name="gatk_param_type_selector" type="select" label="Basic or Advanced GATK options">
+        <option value="basic" selected="True">Basic</option>
+        <option value="advanced">Advanced</option>
+      </param>
+      <when value="basic">
+        <!-- Do nothing here -->
+      </when>
+      <when value="advanced">
+        <repeat name="pedigree" title="Pedigree file">
+            <param name="pedigree_file" type="data" format="txt" label="Pedigree files for samples" />
+        </repeat>
+        <repeat name="pedigree_string_repeat" title="Pedigree string">
+            <param name="pedigree_string" type="text" value="" label="Pedigree string for samples" />
+        </repeat>
+        <param name="pedigree_validation_type" type="select" label="How strict should we be in validating the pedigree information">
+          <option value="STRICT" selected="True">STRICT</option>
+          <option value="SILENT">SILENT</option>
+        </param>
+        <repeat name="read_filter" title="Read Filter">
+            <conditional name="read_filter_type">
+		      <param name="read_filter_type_selector" type="select" label="Read Filter Type">
+		        <option value="BadCigar">BadCigar</option>
+		        <option value="BadMate">BadMate</option>
+		        <option value="DuplicateRead">DuplicateRead</option>
+		        <option value="FailsVendorQualityCheck">FailsVendorQualityCheck</option>
+		        <option value="MalformedRead">MalformedRead</option>
+		        <option value="MappingQuality">MappingQuality</option>
+		        <option value="MappingQualityUnavailable">MappingQualityUnavailable</option>
+		        <option value="MappingQualityZero">MappingQualityZero</option>
+		        <option value="MateSameStrand">MateSameStrand</option>
+		        <option value="MaxInsertSize">MaxInsertSize</option>
+		        <option value="MaxReadLength" selected="True">MaxReadLength</option>
+		        <option value="MissingReadGroup">MissingReadGroup</option>
+		        <option value="NoOriginalQualityScores">NoOriginalQualityScores</option>
+		        <option value="NotPrimaryAlignment">NotPrimaryAlignment</option>
+		        <option value="Platform454">Platform454</option>
+		        <option value="Platform">Platform</option>
+		        <option value="PlatformUnit">PlatformUnit</option>
+		        <option value="ReadGroupBlackList">ReadGroupBlackList</option>
+		        <option value="ReadName">ReadName</option>
+		        <option value="ReadStrand">ReadStrand</option>
+		        <option value="ReassignMappingQuality">ReassignMappingQuality</option>
+		        <option value="Sample">Sample</option>
+		        <option value="SingleReadGroup">SingleReadGroup</option>
+		        <option value="UnmappedRead">UnmappedRead</option>
+		      </param>
+	          <when value="BadCigar">
+	              <!-- no extra options -->
+	          </when>
+	          <when value="BadMate">
+	              <!-- no extra options -->
+	          </when>
+	          <when value="DuplicateRead">
+	              <!-- no extra options -->
+	          </when>
+	          <when value="FailsVendorQualityCheck">
+	              <!-- no extra options -->
+	          </when>
+	          <when value="MalformedRead">
+	              <!-- no extra options -->
+	          </when>
+	          <when value="MappingQuality">
+	              <param name="min_mapping_quality_score" type="integer" value="10" label="Minimum read mapping quality required to consider a read for calling"/>
+	          </when>
+	          <when value="MappingQualityUnavailable">
+	              <!-- no extra options -->
+	          </when>
+	          <when value="MappingQualityZero">
+	              <!-- no extra options -->
+	          </when>
+	          <when value="MateSameStrand">
+	              <!-- no extra options -->
+	          </when>
+	          <when value="MaxInsertSize">
+	              <param name="maxInsertSize" type="integer" value="1000000" label="Discard reads with insert size greater than the specified value"/>
+	          </when>
+	          <when value="MaxReadLength">
+	              <param name="maxReadLength" type="integer" value="76" label="Max Read Length"/>
+	          </when>
+	          <when value="MissingReadGroup">
+	              <!-- no extra options -->
+	          </when>
+	          <when value="NoOriginalQualityScores">
+	              <!-- no extra options -->
+	          </when>
+	          <when value="NotPrimaryAlignment">
+	              <!-- no extra options -->
+	          </when>
+	          <when value="Platform454">
+	              <!-- no extra options -->
+	          </when>
+	          <when value="Platform">
+	              <param name="PLFilterName" type="text" value="" label="Discard reads with RG:PL attribute containing this string"/>
+	          </when>
+	          <when value="PlatformUnit">
+	              <!-- no extra options -->
+	          </when>
+	          <when value="ReadGroupBlackList">
+	              <!-- no extra options -->
+	          </when>
+	          <when value="ReadName">
+	              <param name="readName" type="text" value="" label="Filter out all reads except those with this read name"/>
+	          </when>
+	          <when value="ReadStrand">
+	              <param name="filterPositive" type="boolean" truevalue="--filterPositive" falsevalue="" label="Discard reads on the forward strand"/>
+	          </when>
+	          <when value="ReassignMappingQuality">
+	              <param name="default_mapping_quality" type="integer" value="60" label="Default read mapping quality to assign to all reads"/>
+	          </when>
+	          <when value="Sample">
+	              <param name="sample_to_keep" type="text" value="" label="The name of the sample(s) to keep, filtering out all others"/>
+	          </when>
+	          <when value="SingleReadGroup">
+	              <param name="read_group_to_keep" type="integer" value="76" label="The name of the read group to keep, filtering out all others"/>
+	          </when>
+	          <when value="UnmappedRead">
+	              <!-- no extra options -->
+	          </when>
+            </conditional>
+        </repeat>
+        <repeat name="input_interval_repeat" title="Operate on Genomic intervals">
+          <param name="input_intervals" type="data" format="bed,gatk_interval,picard_interval_list,vcf" label="Genomic intervals" />
+        </repeat>
+        <repeat name="input_exclude_interval_repeat" title="Exclude Genomic intervals">
+          <param name="input_exclude_intervals" type="data" format="bed,gatk_interval,picard_interval_list,vcf" label="Genomic intervals" />
+        </repeat>
+        
+        <param name="interval_set_rule" type="select" label="Interval set rule">
+          <option value="UNION" selected="True">UNION</option>
+          <option value="INTERSECTION">INTERSECTION</option>
+        </param>
+        
+        <conditional name="downsampling_type">
+          <param name="downsampling_type_selector" type="select" label="Type of reads downsampling to employ at a given locus" help="Downsampling Type">
+            <option value="NONE" selected="True">NONE</option>
+            <option value="ALL_READS">ALL_READS</option>
+            <option value="BY_SAMPLE">BY_SAMPLE</option>
+          </param>
+          <when value="NONE">
+	          <!-- no more options here -->
+	      </when>
+          <when value="ALL_READS">
+	          <conditional name="downsample_to_type">
+	              <param name="downsample_to_type_selector" type="select" label="Type of reads downsampling to employ at a given locus" help="Downsampling Type">
+	                  <option value="downsample_to_fraction" selected="True">Downsample by Fraction</option>
+	                  <option value="downsample_to_coverage">Downsample by Coverage</option>
+	              </param>
+	              <when value="downsample_to_fraction">
+	                  <param name="downsample_to_value" type="float" label="Fraction [0.0-1.0] of reads to downsample to" value="1" min="0" max="1"/>
+	              </when>
+	              <when value="downsample_to_coverage">
+	                  <param name="downsample_to_value" type="integer" label="Coverage to downsample to at any given locus" value="0"/>
+	              </when>
+	          </conditional>
+	      </when>
+          <when value="BY_SAMPLE">
+	          <conditional name="downsample_to_type">
+	              <param name="downsample_to_type_selector" type="select" label="Type of reads downsampling to employ at a given locus" help="Downsampling Type">
+	                  <option value="downsample_to_fraction" selected="True">Downsample by Fraction</option>
+	                  <option value="downsample_to_coverage">Downsample by Coverage</option>
+	              </param>
+	              <when value="downsample_to_fraction">
+	                  <param name="downsample_to_value" type="float" label="Fraction [0.0-1.0] of reads to downsample to" value="1" min="0" max="1"/>
+	              </when>
+	              <when value="downsample_to_coverage">
+	                  <param name="downsample_to_value" type="integer" label="Coverage to downsample to at any given locus" value="0"/>
+	              </when>
+	          </conditional>
+	      </when>
+        </conditional>
+        <param name="baq" type="select" label="Type of BAQ calculation to apply in the engine">
+          <option value="OFF" selected="True">OFF</option>
+          <option value="CALCULATE_AS_NECESSARY">CALCULATE_AS_NECESSARY</option>
+          <option value="RECALCULATE">RECALCULATE</option>
+        </param>
+        <param name="baq_gap_open_penalty" type="float" label="BAQ gap open penalty (Phred Scaled)" value="40" help="Default value is 40. 30 is perhaps better for whole genome call sets."/>
+        <param name="use_original_qualities" type="boolean" truevalue="--useOriginalQualities" falsevalue="" label="Use the original base quality scores from the OQ tag" />
+        <param name="default_base_qualities" type="integer" label="Value to be used for all base quality scores, when some are missing" value="-1"/>
+        <param name="validation_strictness" type="select" label="How strict should we be with validation">
+          <option value="STRICT" selected="True">STRICT</option>
+          <option value="LENIENT">LENIENT</option>
+          <option value="SILENT">SILENT</option>
+          <!-- <option value="DEFAULT_STRINGENCY">DEFAULT_STRINGENCY</option> listed in docs, but not valid value...-->
+        </param>
+        <param name="interval_merging" type="select" label="Interval merging rule">
+          <option value="ALL" selected="True">ALL</option>
+          <option value="OVERLAPPING_ONLY">OVERLAPPING_ONLY</option>
+        </param>
+        
+        <repeat name="read_group_black_list_repeat" title="Read group black list">
+          <conditional name="read_group_black_list_type">
+            <param name="read_group_black_list_type_selector" type="select" label="Type of reads read group black list">
+              <option value="file" selected="True">Filters in file</option>
+              <option value="text">Specify filters as a string</option>
+            </param>
+            <when value="file">
+	          <param name="read_group_black_list" type="data" format="txt" label="Read group black list file" />
+	        </when>
+            <when value="text">
+	          <param name="read_group_black_list" type="text" value="tag:string" label="Read group black list tag:string" />
+	        </when>
+          </conditional>
+        </repeat>
+        
+        <param name="disable_experimental_low_memory_sharding" type="boolean" truevalue="--disable_experimental_low_memory_sharding" falsevalue="" label="Disable experimental low-memory sharding functionality." checked="False"/>
+        <param name="non_deterministic_random_seed" type="boolean" truevalue="--nonDeterministicRandomSeed" falsevalue="" label="Makes the GATK behave non deterministically, that is, the random numbers generated will be different in every run" checked="False" />
+        
+      </when>
+    </conditional>
+    
+  </inputs>
+  <outputs>
+    <data format="bam" name="output_bam" label="${tool.name} on ${on_string} (BAM)" />
+    <data format="txt" name="output_log" label="${tool.name} on ${on_string} (log)" />
+  </outputs>
+    <param name="number" type="integer" value="-1" label="Print the first n reads from the file, discarding the rest" />
+    <param name="platform" type="text" value="" label="Exclude all reads with this platform from the output" />
+    <param name="read_group" type="text" value="" label="Exclude all reads with this read group from the output" />
+    <repeat name="sample_file_repeat" title="File containing a list of samples to include">
+        <param name="input_sample_file" type="data" format="text" label="Sample file" />
+    </repeat>
+    <repeat name="sample_name_repeat" title="Sample name to be included in the analysis">
+        <param name="sample_name" type="text" label="Sample name" />
+    </repeat>
+  <tests>
+      <test>
+          <param name="reference_source_selector" value="history" />
+          <param name="ref_file" value="phiX.fasta" ftype="fasta" />
+          <param name="input_bam" value="gatk/gatk_table_recalibration/gatk_table_recalibration_out_1.bam" ftype="bam" />
+          <param name="number" value="-1" />
+          <param name="platform" value="" />
+          <param name="read_group" value="" />
+          <param name="sample_file_repeat" value="0" />
+          <param name="sample_name_repeat" value="0" />
+          <param name="gatk_param_type_selector" value="basic" />
+          <output name="output_bam" file="gatk/gatk_table_recalibration/gatk_table_recalibration_out_1.bam" ftype="bam" /> 
+          <output name="output_log" file="gatk/gatk_print_reads/gatk_print_reads_out_1.log.contains" compare="contains" />
+      </test>
+  </tests>
+  <help>
+**What it does**
+
+PrintReads can dynamically merge the contents of multiple input BAM files, resulting in merged output sorted in coordinate order.
+
+For more information on the GATK Print Reads Walker, see this `tool specific page &lt;http://www.broadinstitute.org/gsa/gatkdocs/release/org_broadinstitute_sting_gatk_walkers_PrintReadsWalker.html&gt;`_.
+
+To learn about best practices for variant detection using GATK, see this `overview &lt;http://www.broadinstitute.org/gsa/wiki/index.php/Best_Practice_Variant_Detection_with_the_GATK_v3&gt;`_.
+
+If you encounter errors, please view the `GATK FAQ &lt;http://www.broadinstitute.org/gsa/wiki/index.php/Frequently_Asked_Questions&gt;`_.
+
+------
+
+**Inputs**
+
+GenomeAnalysisTK: PrintReads accepts one or more BAM or SAM input files.
+
+
+**Outputs**
+
+The output is in BAM format.
+
+
+Go `here &lt;http://www.broadinstitute.org/gsa/wiki/index.php/Input_files_for_the_GATK&gt;`_ for details on GATK file formats.
+
+-------
+
+**Settings**::
+
+ number 	    int 	-1 	Print the first n reads from the file, discarding the rest
+ platform 	    String 	NA 	Exclude all reads with this platform from the output
+ readGroup 	    String 	NA 	Exclude all reads with this read group from the output
+ sample_file 	Set[File] 	[] 	File containing a list of samples (one per line). Can be specified multiple times
+ sample_name 	Set[String] 	[] 	Sample name to be included in the analysis. Can be specified multiple times.
+
+------
+
+**Citation**
+
+For the underlying tool, please cite `DePristo MA, Banks E, Poplin R, Garimella KV, Maguire JR, Hartl C, Philippakis AA, del Angel G, Rivas MA, Hanna M, McKenna A, Fennell TJ, Kernytsky AM, Sivachenko AY, Cibulskis K, Gabriel SB, Altshuler D, Daly MJ. A framework for variation discovery and genotyping using next-generation DNA sequencing data. Nat Genet. 2011 May;43(5):491-8. &lt;http://www.ncbi.nlm.nih.gov/pubmed/21478889&gt;`_
+
+If you use this tool in Galaxy, please cite Blankenberg D, et al. *In preparation.*
+
+  </help>
+</tool>