Added --gff3-genes-as-transcripts option for virus genomes

rsem-gff3-to-gtf

@@ -153,7 +153,7 @@ class Transcript:
 		self.index = 0
 		return self
 
-	def next(self):
+	def __next__(self):
 		if self.index == len(self.results):
 			raise StopIteration
 		interval = self.results[self.index]
@@ -217,6 +217,7 @@ def flush_out(fout):
 parser = HelpOnErrorParser(formatter_class = argparse.ArgumentDefaultsHelpFormatter, description = "Convert GFF3 files to GTF files.")
 parser.add_argument("input_GFF3_file", help = "Input GFF3 file.")
 parser.add_argument("output_GTF_file", help = "Output GTF file.")
+parser.add_argument("--make-genes-as-transcripts", help = "GFF3 file does not contain transcripts, make each gene as a transcript.", action = "store_true")
 parser.add_argument("--RNA-patterns", help = "Types of RNAs to be extracted, e.g. mRNA,rRNA", metavar = "<patterns>")
 parser.add_argument("--extract-sequences", help = "If GFF3 file contains reference sequences, extract them to the specified file", metavar = "<output.fa>")
 args = parser.parse_args()
@@ -274,11 +275,16 @@ with open(args.input_GFF3_file) as fin:
 			my_assert(gid not in gid2gname, 
 				"Gene {0} appears multiple times! Last occurrence is at line {1}:\n{2}".format(gid, feature.line_no, feature.line))
 			gid2gname[gid] = feature.getAttribute("Name")
-		elif feature.feature_type == "transcript":
+
+			if args.make_genes_as_transcripts:
+				feature.feature_type = feature.original_type = "transcript"
+				feature.attribute_dict["Parent"] = [feature.attribute_dict["ID"]]
+
+		if feature.feature_type == "transcript":
 			transcript = getTranscript(feature.getAttribute("ID", True), feature)
 			transcript.setTranscript(feature)
-		else:
-			assert feature.feature_type == "exon"
+
+		if feature.feature_type == "exon":
 			for parent in feature.getAttribute("Parent", True):
 				transcript = getTranscript(parent, feature)
 				transcript.addExon(feature)

rsem-prepare-reference

@@ -18,6 +18,7 @@ my $status;
 my $gtfF = "";
 my $gff3F = "";
 my $gff3_RNA_patterns = "";
+my $gff3_genes_as_transcripts = 0;
 my $gtf_sources = "None";
 my $mappingF = "";
 my $polyAChoice = 1; # 0, --polyA, add polyA tails for all isoforms; 1, default, no polyA tails; 2, --no-polyA-subset
@@ -43,6 +44,7 @@ my $star_sjdboverhang = 100;
 GetOptions("gtf=s" => \$gtfF,
 	   "gff3=s" => \$gff3F,
 	   "gff3-RNA-patterns=s" => \$gff3_RNA_patterns,
+     "gff3-genes-as-transcripts" => \$gff3_genes_as_transcripts,
 	   "trusted-sources=s" => \$gtf_sources,
 	   "transcript-to-gene-map=s" => \$mappingF,
 	   "allele-to-gene-map=s" => \$alleleMappingF,
@@ -116,6 +118,9 @@ if ($gff3F ne "") {
     if ($gff3_RNA_patterns ne "") {
 	$command .= " --RNA-patterns $gff3_RNA_patterns";
     }
+    if ($gff3_genes_as_transcripts) {
+      $command .= " --make-genes-as-transcripts";
+    }
     $command .= " $gff3F $gtfF";
     &runCommand($command)
 }
@@ -260,6 +265,10 @@ The annotation file is in GFF3 format instead of GTF format. RSEM will first con
 
 <pattern> is a comma-separated list of transcript categories, e.g. "mRNA,rRNA". Only transcripts that match the <pattern> will be extracted. (Default: "mRNA")
 
+=item B<--gff3-genes-as-transcripts>
+
+This option is designed for untypical organisms, such as viruses, whose GFF3 files only contain genes. RSEM will assume each gene as a unique transcript when it converts the GFF3 file into GTF format.
+
 =item B<--trusted-sources> <sources>
 
 <sources> is a comma-separated list of trusted sources, e.g. "ENSEMBL,HAVANA". Only transcripts coming from these sources will be extracted. If this option is off, all sources are accepted. (Default: off)