Fig. 2

Generation of Nonredundant Contig Set, Canonical Coding Transcript Set, and High Confidence Novel Transcript Set. From the union of all contigs, we generated the nonredundant set of transcripts by iterative pairwise merging of contig set of all tissues; this yielded 68 % reduction of the contig set. a To generate Canonical Coding Transcript Set, we selected the contigs that are annotated with MSA pipeline. The annotated contigs are further filtered for contigs that have a gene symbol. For an individual gene cluster, we chose a transcript with the longest ORF to represent the corresponding gene (Canonical Coding Transcript Set). b For unannoated contigs, we selected for expression level, presence of an ORF with both start and stop codons in the CDS, and a minimum length of 400 nucleotides. We identified 8 high-confidence novel coding transcript candidates for validation