diff --git a/lib/galaxy/datatypes/util/gff_util.py b/lib/galaxy/datatypes/util/gff_util.py
index 996e97028fe..894bb0b7e3b 100644
--- a/lib/galaxy/datatypes/util/gff_util.py
+++ b/lib/galaxy/datatypes/util/gff_util.py
@@ -12,8 +12,18 @@ class GFFInterval( GenomicInterval ):
     """
     def __init__( self, reader, fields, chrom_col, feature_col, start_col, end_col, \
                   strand_col, score_col, default_strand, fix_strand=False, raw_line='' ):
+        # HACK: GFF format allows '.' for strand but GenomicInterval does not. To get around this,
+        # temporarily set strand and then unset after initing GenomicInterval.
+        unknown_strand = False
+        if not fix_strand and fields[ strand_col ] == '.':
+            unknown_strand = True
+            fields[ strand_col ] = '+'
         GenomicInterval.__init__( self, reader, fields, chrom_col, start_col, end_col, strand_col, \
                                   default_strand, fix_strand=fix_strand )
+        if unknown_strand:
+            self.strand = '.'
+            self.fields[ strand_col ] = '.'
+                                  
         # Handle feature, score column.
         self.feature_col = feature_col
         if self.feature_col >= self.nfields:
@@ -40,13 +50,10 @@ class GFFFeature( GFFInterval ):
         self.intervals = intervals
         # Use intervals to set feature attributes.
         for interval in self.intervals:
-            # Error checking.
+            # Error checking. NOTE: intervals need not share the same strand.
             if interval.chrom != self.chrom:
-                raise ValueError( "interval chrom does not match self chrom: %i != %i" % \
+                raise ValueError( "interval chrom does not match self chrom: %s != %s" % \
                                   ( interval.chrom, self.chrom ) )
-            if interval.strand != self.strand:
-                raise ValueError( "interval strand does not match self strand: %s != %s" % \
-                                  ( interval.strand, self.strand ) )
             # Set start, end of interval.
             if interval.start < self.start:
                 self.start = interval.start
@@ -140,7 +147,7 @@ class GFFReaderWrapper( NiceReaderWrapper ):
             # For debugging, uncomment this to propogate parsing exceptions up. 
             # I.e. the underlying reason for an unexpected StopIteration exception 
             # can be found by uncommenting this. 
-            # raise e
+            #raise e
                
         #
         # Get next GFFFeature
@@ -163,7 +170,7 @@ class GFFReaderWrapper( NiceReaderWrapper ):
     
         # Initialize feature name from seed.
         feature_group = self.seed_interval.attributes.get( 'group', None ) # For GFF
-        feature_id = self.seed_interval.attributes.get( 'id', None ) # For GFF3
+        feature_id = self.seed_interval.attributes.get( 'ID', None ) # For GFF3
         feature_gene_id = self.seed_interval.attributes.get( 'gene_id', None ) # For GTF
         feature_transcript_id = self.seed_interval.attributes.get( 'transcript_id', None ) # For GTF
 
@@ -183,11 +190,14 @@ class GFFReaderWrapper( NiceReaderWrapper ):
             
             # If interval not associated with feature, break.
             group = interval.attributes.get( 'group', None )
+            # GFF test:
             if group and feature_group != group:
                 break
-            id = interval.attributes.get( 'id', None )
-            if id and feature_id != id:
+            # GFF3 test:
+            parent = interval.attributes.get( 'Parent', None )
+            if feature_id and feature_id != parent:
                 break
+            # GTF test:
             gene_id = interval.attributes.get( 'gene_id', None )
             transcript_id = interval.attributes.get( 'transcript_id', None )
             if ( transcript_id and transcript_id != feature_transcript_id ) or \
diff --git a/tools/filters/gff/gff_filter_by_feature_count.py b/tools/filters/gff/gff_filter_by_feature_count.py
index 768a89c4e9d..6dd648fe4de 100644
--- a/tools/filters/gff/gff_filter_by_feature_count.py
+++ b/tools/filters/gff/gff_filter_by_feature_count.py
@@ -8,8 +8,7 @@ Usage:
 import sys
 from galaxy import eggs
 from galaxy.datatypes.util.gff_util import GFFReaderWrapper
-
-assert sys.version_info[:2] >= ( 2, 4 )
+from bx.intervals.io import GenomicInterval
 
 # Valid operators, ordered so that complex operators (e.g. '>=') are
 # recognized before simple operators (e.g. '>')
@@ -62,7 +61,9 @@ def __main__():
     skipped_lines = 0
     first_skipped_line = 0
     out = open( output_name, 'w' )
-    for i, feature in enumerate( GFFReaderWrapper( open( input_name ), fix_strand=True ) ):
+    for i, feature in enumerate( GFFReaderWrapper( open( input_name ) ) ):
+        if not isinstance( feature, GenomicInterval ):
+            continue
         count = 0
         for interval in feature.intervals:
             if interval.feature == feature_name:
@@ -73,6 +74,9 @@ def __main__():
                 out.write( "\t".join(interval.fields) + '\n' )
             kept_features += 1
 
+    # Needed because i is 0-based but want to display stats using 1-based.
+    i += 1
+
     # Clean up.
     out.close()
     info_msg = "%i of %i features kept (%.2f%%) using condition %s.  " % \
diff --git a/tools/filters/gff/gff_filter_by_feature_count.xml b/tools/filters/gff/gff_filter_by_feature_count.xml
index a1fde6dea85..d571f1f9549 100644
--- a/tools/filters/gff/gff_filter_by_feature_count.xml
+++ b/tools/filters/gff/gff_filter_by_feature_count.xml
@@ -20,12 +20,20 @@
     <data format="input" name="out_file1" metadata_source="input_file1"/>
   </outputs>
   <tests>
-    <test>
-        <param name="input_file1" value="gops_subtract_in1.gff"/>
-        <param name="feature_name" value="exon"/>
-        <param name="cond" value=">1"/>
-        <output name="out_file1" file="gff_filter_by_feature_count_out1.gff"/>
-    </test>
+      <!-- Test GTF filtering. -->
+      <test>
+          <param name="input_file1" value="gops_subtract_in1.gff"/>
+          <param name="feature_name" value="exon"/>
+          <param name="cond" value=">1"/>
+          <output name="out_file1" file="gff_filter_by_feature_count_out1.gff"/>
+      </test>
+      <!-- Test GFF3 filtering. -->
+      <test>
+          <param name="input_file1" value="5.gff3"/>
+          <param name="feature_name" value="HSP"/>
+          <param name="cond" value=">=5"/>
+          <output name="out_file1" file="gff_filter_by_feature_count_out2.gff"/>
+      </test>
   </tests>
 
   <help>