CPAT

Using RNA-seq, tens of thousands of novel transcripts and isoforms have been identified (Djebali, et al Nature, 2012 , Carbili et al, Gene & Development, 2011) The discovery of these hidden transcriptome rejuvenate the need of distinguishing coding and noncoding RNA. However, Most previous coding potential prediction methods heavily rely on alignment, either pairwise alignment to search for protein evidence or multiple alignments to calculate phylogenetic conservation score (such as CPC , PhyloCSF and RNACode ). This is because most previously identified transcripts including protein coding RNA and short, housekeeping/regulatory RNAs such as snRNAs, snoRNA and tRNA are highly conserved. While still very useful, these approaches have several limitations:

Most lncRNAs are less conserved and tend to be lineage specific which greatly limit the discrimination power of alignment-based methods. For example, of 550 lncRNAs detected from zebrafish, only 29 of them had detectable sequence