The GENCODE gene set gives the most comprehensive overview of alternative splicing than of any gene set available. However improved transcriptome sequencing is being facilitated by high-depth RNA sequencing (Mercer et al. 2011), which reveals that the full extent at which different transcripts are expressed in different cells and tissues and at different developmental stages is still to be fully characterized. The RT-PCR-seq methodology developed within the GENCODE project has identified novel exons within 10% of annotated introns being targeted (Howald et al. 2012). Therefore, even when the first pass of whole-genome manual annotation is finished in 2012, reannotation updates will be required on a large percentage of loci to correctly classify the new alternatively spliced variants being revealed by next-generation transcriptomics. However the major drawback with using next-generation short reads to assemble transcripts de novo is that the correct structure of the transcript is hard to predict, as was investigated in the RNA-seq Genome Annotation Assessment Project (RGASP) (J Harrow, T Steijger, F Kokocinski, JF Abril, C Howald, A Reymond, A Mortazavi, B Wold, T Gingeras, R Guigó, et
