Skip to main content
European Commission logo print header

Domain integrity verification of alternative splicing using statistical methods and molecular modeling

Final Activity Report Summary - DIVAS (Domain Integrity Verification of Alternative Splicing using statistical methods and molecular modeling)

Using bioinformatics we studied alternative splicing in the human proteome available in the Swissprot database. We analysed 20538 isoforms, i.e. alternative splicing variants of 7101 human genes currently in Swissprot with at least two tentative splice variants. Using statistics, the Pfam domain database, 3D protein structures in PDB, surface energy calculations (with special respect to exposed hydrophobicity) and intrinsic protein disorder predictions we were able to explain all the known existing splice variants currently with experimental evidence in Swissprot (501 splice variants).

Expanding this data set to include all the named variants in Swissprot we gained a lot of insight into the nature of alternative splicing and concluded that it is basically governed by two rules: 1. Validated alternative splicing (with a surviving protein isoform) occurs mostly in disordered regions and tends to avoid globular domains; 2. If the splicing does truncate a globular domain it is usually marginal (less than 10% of the domain is removed by the truncation event) or the resulting sub-domain has the characteristics of an intact domain regarding the newly exposed hydrophobic surface area.