Introns are non-coding intervals that interrupt the codingsequences of eukaryotic genes. Intron removal is performedby a complicated molecular machinery, called thespliceosome, concomitantly with gene transcription. Intronsand the splicing machinery (or at least their traces) arefound in every sequenced eukaryotic genome. Moreover, manyintrons are found at homologous positions across differentkingdoms, suggesting that some originate in the earliesteukaryotes.Introns are largely devoid of function, yet in humans (andmammals in general), they make up more than~40% of thegenome. The most obvious evolutionary advantage of theinterrupted coding sequences is that they increasefunctional complexity by enabling alternate assemblies.Introns provide a powerful source of variations for naturalselection in many other ways, since splicing is tightlycoupled with transcription and export from the nucleus, andintronic sequences frequently harbor regulatory elements.The proposed project aims at developing bioinformatics toolsand mathematical models that help understanding randomnessand natural selection that shape exon-intron architecture indifferent eukaryotic lineages. In particular, we willinvestigate intron turnover in fast-evolving genes,selective constraints on intron length and mechanisms ofintron gain on a large scale, using annotated whole genomes.In addition to providing new insights into the waysevolution affects gene structure, the developedmethods will be useful to produce better annotations ofcoding regions and functional intronic elements.