On the comparison of sets of alternative transcripts

Alternative splicing is pervasive among complex eukaryote species. For some genes shared by numerous species, dozens of alternative transcripts are already annotated in databases. Most recent studies compare and catalog alternate splicing events within or across species, but there is an urgent need to be able to compare sets of whole transcripts both manually and automatically. In this paper, we propose a general framework to compare sets of transcripts that are transcribed from orthologous loci of several species. The model is based on the construction of a common reference sequence, and on annotations that allow the reconstruction of ancestral sequences, the identification of conserved events, and the inference of gains and losses of donor/acceptors sites, exons, introns and transcripts. Our representation of sets of transcripts is straightforward, and readable by both humans and computers. On the other hand, the model has a precise, formal specification that insures its coherence, consistency and scalability. We give several examples, among them a comparison of 24 Smox gene transcripts across five species.