Merge pull request #5360 from jj-umn/dev

Add Proteomics datatypes proBAM and proBED
galaxyproject · Feb 11, 2018 · 0367a35 · 0367a35
2 parents 587f20c + 007e3cc
commit 0367a35
Show file tree

Hide file tree

Showing 3 changed files with 16 additions and 0 deletions.
diff --git a/config/datatypes_conf.xml.sample b/config/datatypes_conf.xml.sample
@@ -20,6 +20,7 @@
       <converter file="bam_to_bigwig_converter.xml" target_datatype="bigwig"/>
       <converter file="bam_native_to_bam_converter.xml" target_datatype="bam"/>
     </datatype>
+    <datatype extension="probam" type="galaxy.datatypes.binary:ProBam" mimetype="application/octet-stream" display_in_upload="true"/>
     <datatype extension="cram" type="galaxy.datatypes.binary:CRAM" mimetype="application/octet-stream" display_in_upload="true" description="CRAM is a file format for highly efficient and tunable reference-based compression of alignment data." description_url="http://www.ebi.ac.uk/ena/software/cram-usage">
         <converter file="cram_to_bam_converter.xml" target_datatype="bam"/>
     </datatype>
@@ -40,6 +41,7 @@
     <datatype extension="bed6" type="galaxy.datatypes.interval:Bed6" display_in_upload="true">
     </datatype>
     <datatype extension="bed12" type="galaxy.datatypes.interval:Bed12" display_in_upload="true"/>
+    <datatype extension="probed" type="galaxy.datatypes.interval:ProBed" display_in_upload="true"/>
     <datatype extension="len" type="galaxy.datatypes.chrominfo:ChromInfo" display_in_upload="true">
       <converter file="len_to_linecount.xml" target_datatype="linecount" />
     </datatype>

diff --git a/lib/galaxy/datatypes/binary.py b/lib/galaxy/datatypes/binary.py
@@ -489,6 +489,13 @@ def samtools_dataprovider(self, dataset, **settings):
         return dataproviders.dataset.SamtoolsDataProvider(dataset_source, **settings)
 
 
+class ProBam(Bam):
+    """Class describing a BAM binary file - extended for proteomics data"""
+    edam_format = "format_3826"
+    edam_data = "data_0863"
+    file_ext = "probam"
+
+
 class CRAM(Binary):
     file_ext = "cram"
     edam_format = "format_3462"

diff --git a/lib/galaxy/datatypes/interval.py b/lib/galaxy/datatypes/interval.py
@@ -572,6 +572,13 @@ def sniff(self, filename):
             return False
 
 
+class ProBed(Bed):
+    """Tab delimited data in proBED format - adaptation of BED for proteomics data."""
+    edam_format = "format_3827"
+    file_ext = "probed"
+    column_names = ['Chrom', 'Start', 'End', 'Name', 'Score', 'Strand', 'ThickStart', 'ThickEnd', 'ItemRGB', 'BlockCount', 'BlockSizes', 'BlockStarts', 'ProteinAccession', 'PeptideSequence', 'Uniqueness', 'GenomeReferenceVersion', 'PsmScore', 'Fdr', 'Modifications', 'Charge', 'ExpMassToCharge', 'CalcMassToCharge', 'PsmRank', 'DatasetID', 'Uri']
+
+
 class BedStrict(Bed):
     """Tab delimited data in strict BED format - no non-standard columns allowed"""
     edam_format = "format_3584"