commit/galaxy-central: dan: Remove bam index metadata validator from GATK tools. The GATK wrapper will now build missing bam indexes on the fly as needed.

27 Feb 2012

1 new commit in galaxy-central:


https://bitbucket.org/galaxy/galaxy-central/changeset/167fddae9646/
changeset:   167fddae9646
user:        dan
date:        2012-02-27 18:49:30
summary:     Remove bam index metadata validator from GATK tools. The GATK wrapper will now build missing bam indexes on the fly as needed.
affected #:  10 files

diff -r 17caba3f5b808bd62a96ab3076934779dd01fb77 -r 167fddae9646f761a87c99a15984f52a605cc498 tools/gatk/analyze_covariates.xml

--- a/tools/gatk/analyze_covariates.xml
+++ b/tools/gatk/analyze_covariates.xml
@@ -74,7 +74,7 @@
 
 **Outputs**
 
-The output is in and HTML file with links to PDF graphs and a data files.
+The output is in CSV and HTML files with links to PDF graphs and a data files.
 
 
 Go `here <http://www.broadinstitute.org/gsa/wiki/index.php/Input_files_for_the_GATK>`_ for details on GATK file formats.


diff -r 17caba3f5b808bd62a96ab3076934779dd01fb77 -r 167fddae9646f761a87c99a15984f52a605cc498 tools/gatk/count_covariates.xml
--- a/tools/gatk/count_covariates.xml
+++ b/tools/gatk/count_covariates.xml
@@ -2,12 +2,15 @@
   <description>on BAM files</description><requirements><requirement type="package" version="1.4">gatk</requirement>
+      <requirement type="package">samtools</requirement></requirements><command interpreter="python">gatk_wrapper.py
    --max_jvm_heap_fraction "1"
    --stdout "${output_log}"
    -d "-I" "${reference_source.input_bam}" "${reference_source.input_bam.ext}" "gatk_input"
-   -d "" "${reference_source.input_bam.metadata.bam_index}" "bam_index" "gatk_input" ##hardcode galaxy ext type as bam_index
+   #if str( $reference_source.input_bam.metadata.bam_index ) != "None":
+       -d "" "${reference_source.input_bam.metadata.bam_index}" "bam_index" "gatk_input" ##hardcode galaxy ext type as bam_index
+   #end if
    -p 'java 
     -jar "${GALAXY_DATA_INDEX_DIR}/shared/jars/gatk/GenomeAnalysisTK.jar"
     -T "CountCovariates"
@@ -143,7 +146,6 @@
       <when value="cached"><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><param name="ref_file" type="select" label="Using reference genome">
@@ -155,7 +157,6 @@
       </when><when value="history"><!-- FIX ME!!!! --><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><param name="ref_file" type="data" format="fasta" label="Using reference file" /></when>


diff -r 17caba3f5b808bd62a96ab3076934779dd01fb77 -r 167fddae9646f761a87c99a15984f52a605cc498 tools/gatk/depth_of_coverage.xml
--- a/tools/gatk/depth_of_coverage.xml
+++ b/tools/gatk/depth_of_coverage.xml
@@ -2,13 +2,16 @@
   <description>on BAM files</description><requirements><requirement type="package" version="1.4">gatk</requirement>
+      <requirement type="package">samtools</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
+       #if str( $input_bam.input_bam.metadata.bam_index ) != "None":
+           -d "" "${input_bam.input_bam.metadata.bam_index}" "bam_index" "gatk_input_${i}" ##hardcode galaxy ext type as bam_index
+       #end if
    #end for
    -p 'java 
     -jar "${GALAXY_DATA_INDEX_DIR}/shared/jars/gatk/GenomeAnalysisTK.jar"
@@ -190,7 +193,6 @@
         <repeat name="input_bams" title="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>
@@ -204,7 +206,6 @@
       <when value="history"><!-- FIX ME!!!! --><repeat name="input_bams" title="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" />


diff -r 17caba3f5b808bd62a96ab3076934779dd01fb77 -r 167fddae9646f761a87c99a15984f52a605cc498 tools/gatk/gatk_wrapper.py
--- a/tools/gatk/gatk_wrapper.py
+++ b/tools/gatk/gatk_wrapper.py
@@ -43,6 +43,23 @@
         html_out.write(  '<li><a href="%s">%s</a></li>\n' % ( fname, fname ) )
     html_out.write( '</ul>\n</body>\n</html>\n' )
 
+def index_bam_files( bam_filenames, tmp_dir ):
+    for bam_filename in bam_filenames:
+        bam_index_filename = "%s.bai" % bam_filename
+        if not os.path.exists( bam_index_filename ):
+            #need to index this bam file
+            stderr_name = tempfile.NamedTemporaryFile( prefix = "bam_index_stderr" ).name
+            command = 'samtools index %s %s' % ( bam_filename, bam_index_filename )
+            proc = subprocess.Popen( args=command, shell=True, stderr=open( stderr_name, 'wb' ) )
+            return_code = proc.wait()
+            if return_code:
+                for line in open( stderr_name ):
+                    print >> sys.stderr, line
+                os.unlink( stderr_name ) #clean up
+                cleanup_before_exit( tmp_dir )
+                raise Exception( "Error indexing BAM file" )
+            os.unlink( stderr_name ) #clean up
+
 def __main__():
     #Parse Command Line
     parser = optparse.OptionParser()
@@ -67,11 +84,15 @@
         cmd = cmd.replace( 'java ', 'java -Xmx%s ' % ( options.max_jvm_heap ), 1 )
     elif options.max_jvm_heap_fraction is not None:
         cmd = cmd.replace( 'java ', 'java -XX:DefaultMaxRAMFraction=%s  -XX:+UseParallelGC ' % ( options.max_jvm_heap_fraction ), 1 )
+    bam_filenames = []
     if options.datasets:
         for ( dataset_arg, filename, galaxy_ext, prefix ) in options.datasets:
             gatk_filename = gatk_filename_from_galaxy( filename, galaxy_ext, target_dir = tmp_dir, prefix = prefix )
             if dataset_arg:
                 cmd = '%s %s "%s"' % ( cmd, gatk_filetype_argument_substitution( dataset_arg, galaxy_ext ), gatk_filename )
+            if galaxy_ext == "bam":
+                bam_filenames.append( gatk_filename )
+    index_bam_files( bam_filenames, tmp_dir )
     #set up stdout and stderr output options
     stdout = open_file_from_option( options.stdout, mode = 'wb' )
     stderr = open_file_from_option( options.stderr, mode = 'wb' )


diff -r 17caba3f5b808bd62a96ab3076934779dd01fb77 -r 167fddae9646f761a87c99a15984f52a605cc498 tools/gatk/indel_realigner.xml
--- a/tools/gatk/indel_realigner.xml
+++ b/tools/gatk/indel_realigner.xml
@@ -2,12 +2,15 @@
   <description>- perform local realignment</description><requirements><requirement type="package" version="1.4">gatk</requirement>
+      <requirement type="package">samtools</requirement></requirements><command interpreter="python">gatk_wrapper.py
    --max_jvm_heap_fraction "1"
    --stdout "${output_log}"
    -d "-I" "${reference_source.input_bam}" "${reference_source.input_bam.ext}" "gatk_input"
-   -d "" "${reference_source.input_bam.metadata.bam_index}" "bam_index" "gatk_input" ##hardcode galaxy ext type as bam_index
+   #if str( $reference_source.input_bam.metadata.bam_index ) != "None":
+       -d "" "${reference_source.input_bam.metadata.bam_index}" "bam_index" "gatk_input" ##hardcode galaxy ext type as bam_index
+   #end if
    -p 'java 
     -jar "${GALAXY_DATA_INDEX_DIR}/shared/jars/gatk/GenomeAnalysisTK.jar"
     -T "IndelRealigner"
@@ -121,7 +124,6 @@
       <when value="cached"><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><param name="ref_file" type="select" label="Using reference genome">
@@ -133,7 +135,6 @@
       </when><when value="history"><!-- FIX ME!!!! --><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><param name="ref_file" type="data" format="fasta" label="Using reference file"><options>


diff -r 17caba3f5b808bd62a96ab3076934779dd01fb77 -r 167fddae9646f761a87c99a15984f52a605cc498 tools/gatk/print_reads.xml
--- a/tools/gatk/print_reads.xml
+++ b/tools/gatk/print_reads.xml
@@ -2,13 +2,16 @@
   <description>from BAM files</description><requirements><requirement type="package" version="1.4">gatk</requirement>
+      <requirement type="package">samtools</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
+       #if str( $input_bam.input_bam.metadata.bam_index ) != "None":
+           -d "" "${input_bam.input_bam.metadata.bam_index}" "bam_index" "gatk_input_${i}" ##hardcode galaxy ext type as bam_index
+       #end if
    #end for
    -p 'java 
     -jar "${GALAXY_DATA_INDEX_DIR}/shared/jars/gatk/GenomeAnalysisTK.jar"
@@ -106,7 +109,6 @@
         <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>
@@ -120,7 +122,6 @@
       <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" />


diff -r 17caba3f5b808bd62a96ab3076934779dd01fb77 -r 167fddae9646f761a87c99a15984f52a605cc498 tools/gatk/realigner_target_creator.xml
--- a/tools/gatk/realigner_target_creator.xml
+++ b/tools/gatk/realigner_target_creator.xml
@@ -2,12 +2,15 @@
   <description>for use in local realignment</description><requirements><requirement type="package" version="1.3">gatk</requirement>
+      <requirement type="package">samtools</requirement></requirements><command interpreter="python">gatk_wrapper.py
    --max_jvm_heap_fraction "1"
    --stdout "${output_log}"
    -d "-I" "${reference_source.input_bam}" "${reference_source.input_bam.ext}" "gatk_input"
-   -d "" "${reference_source.input_bam.metadata.bam_index}" "bam_index" "gatk_input" ##hardcode galaxy ext type as bam_index
+   #if str( $reference_source.input_bam.metadata.bam_index ) != "None":
+       -d "" "${reference_source.input_bam.metadata.bam_index}" "bam_index" "gatk_input" ##hardcode galaxy ext type as bam_index
+   #end if
    -p 'java 
     -jar "${GALAXY_DATA_INDEX_DIR}/shared/jars/gatk/GenomeAnalysisTK.jar"
     -T "RealignerTargetCreator"
@@ -109,7 +112,6 @@
       <when value="cached"><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><param name="ref_file" type="select" label="Using reference genome">
@@ -121,7 +123,6 @@
       </when><when value="history"><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><param name="ref_file" type="data" format="fasta" label="Using reference file"><options>


diff -r 17caba3f5b808bd62a96ab3076934779dd01fb77 -r 167fddae9646f761a87c99a15984f52a605cc498 tools/gatk/table_recalibration.xml
--- a/tools/gatk/table_recalibration.xml
+++ b/tools/gatk/table_recalibration.xml
@@ -2,12 +2,15 @@
   <description>on BAM files</description><requirements><requirement type="package" version="1.4">gatk</requirement>
+      <requirement type="package">samtools</requirement></requirements><command interpreter="python">gatk_wrapper.py
    --max_jvm_heap_fraction "1"
    --stdout "${output_log}"
    -d "-I" "${reference_source.input_bam}" "${reference_source.input_bam.ext}" "gatk_input"
-   -d "" "${reference_source.input_bam.metadata.bam_index}" "bam_index" "gatk_input" ##hardcode galaxy ext type as bam_index
+   #if str( $reference_source.input_bam.metadata.bam_index ) != "None":
+       -d "" "${reference_source.input_bam.metadata.bam_index}" "bam_index" "gatk_input" ##hardcode galaxy ext type as bam_index
+   #end if
    -p 'java 
     -jar "${GALAXY_DATA_INDEX_DIR}/shared/jars/gatk/GenomeAnalysisTK.jar"
     -T "TableRecalibration"
@@ -126,7 +129,6 @@
       <when value="cached"><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><param name="ref_file" type="select" label="Using reference genome">
@@ -138,7 +140,6 @@
       </when><when value="history"><!-- FIX ME!!!! --><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><param name="ref_file" type="data" format="fasta" label="Using reference file" /></when>


diff -r 17caba3f5b808bd62a96ab3076934779dd01fb77 -r 167fddae9646f761a87c99a15984f52a605cc498 tools/gatk/unified_genotyper.xml
--- a/tools/gatk/unified_genotyper.xml
+++ b/tools/gatk/unified_genotyper.xml
@@ -2,13 +2,16 @@
   <description>SNP and indel caller</description><requirements><requirement type="package" version="1.4">gatk</requirement>
+      <requirement type="package">samtools</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
+       #if str( $input_bam.input_bam.metadata.bam_index ) != "None":
+           -d "" "${input_bam.input_bam.metadata.bam_index}" "bam_index" "gatk_input_${i}" ##hardcode galaxy ext type as bam_index
+       #end if
    #end for
    -p 'java 
     -jar "${GALAXY_DATA_INDEX_DIR}/shared/jars/gatk/GenomeAnalysisTK.jar"
@@ -155,7 +158,6 @@
         <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>
@@ -169,7 +171,6 @@
       <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" />


diff -r 17caba3f5b808bd62a96ab3076934779dd01fb77 -r 167fddae9646f761a87c99a15984f52a605cc498 tools/gatk/variant_annotator.xml
--- a/tools/gatk/variant_annotator.xml
+++ b/tools/gatk/variant_annotator.xml
@@ -2,13 +2,16 @@
   <description></description><requirements><requirement type="package" version="1.4">gatk</requirement>
+      <requirement type="package">samtools</requirement></requirements><command interpreter="python">gatk_wrapper.py
    --max_jvm_heap_fraction "1"
    --stdout "${output_log}"
    #if str( $reference_source.input_bam ) != "None":
        -d "-I" "${reference_source.input_bam}" "${reference_source.input_bam.ext}" "gatk_input"
-       -d "" "${reference_source.input_bam.metadata.bam_index}" "bam_index" "gatk_input" ##hardcode galaxy ext type as bam_index
+       #if str( $reference_source.input_bam.metadata.bam_index ) != "None":
+           -d "" "${reference_source.input_bam.metadata.bam_index}" "bam_index" "gatk_input" ##hardcode galaxy ext type as bam_index
+       #end if
    #end if
    -d "--variant" "${reference_source.input_variant}" "${reference_source.input_variant.ext}" "input_variant"
    -p 'java 
@@ -151,7 +154,6 @@
         <param name="input_variant_bti" type="boolean" truevalue="-BTI variant" falsevalue="" label="Increase efficiency for small variant files." /><param name="input_bam" type="data" format="bam" label="BAM file" optional="True" help="Not needed for all annotations." ><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><param name="ref_file" type="select" label="Using reference genome">
@@ -165,7 +167,6 @@
         <param name="input_variant" type="data" format="vcf" label="Variant file to annotate" /><param name="input_variant_bti" type="boolean" truevalue="-BTI variant" falsevalue="" label="Increase efficiency for small variant files." /><param name="input_bam" type="data" format="bam" label="BAM file" optional="True" >
-          <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><param name="ref_file" type="data" format="fasta" label="Using reference file" /></when>

Repository URL: https://bitbucket.org/galaxy/galaxy-central/

--

This is a commit notification from bitbucket.org. You are receiving
this because you have the service enabled, addressing the recipient of
this email.

    

Bitbucket

tags

participants (1)