Highlight talk – Theme: Data
Abstract
Current variant repositories report on the relationships between single variants and phenotypes, ignoring that many disorders classically considered monogenic might be better described by more complex inheritance mechanisms. The digenic diseases database DIDA moves beyond this state of the art, allowing geneticist to discover the relationships between variant combinations in different genes and phenotypes. DIDA reveals for instance that certain variant combinations are required simultaneously to induce a disease whereas in others only one induces the disease and the other tunes its severity. We show that predictive models using the annotated DIDA data easily identify the features distinguishing between both types. Moreover, comparisons to neutral variant combinations show also that a highly accurate digenic variant effect predictor may be designed separating disease and non-disease combinations. Together, DIDA and its developments provide a novel route in genetics research, which will be explored thoroughly in the years to come.
Authors
Tom Lenaerts, Universite Libre de Bruxelles, Belgium
Andrea Gazzo, (IB)² – Interuniversity Institute of Bioinformatics in Brussels, Belgium
Dorien Daneels, UZ Brussel, Belgium
Elisa Cilia, Université Libre de Bruxelles, Belgium
Maryse Bonduelle, UZ Brussel, Belgium
Marc Abramowicz, Université Libre de Bruxelles, Belgium
Sonia Van Dooren, UZ Brussel, Belgium
Guillaume Smits, HUDERF, Belgium
Source of publication
2016, Nucleic Acids Research 44 (D1): D900-D907.