Our AI approach can assess the pathogenicity of digenic variant combinations with respect to a patient's specific phenotypes.
Beyond monogenic interpretation
By taking into consideration the additive effect of variant combinations and genetic modifiers we aim to explain the missing heritability and variability of phenotypic outcomes for hundreds of undiagnosed rare diseases.
Our new technology applies a sophisticated artificial intelligence algorithm that can identify combinations of genomic variants among millions of possibilities, rather than considering one variant at a time.
The tool, aptly named DIVAs, is able to evaluate the pathogenicity of combinations of variants present on two different genes, thus supporting pure digenic, modifiers-mediated or dual diagnosis hypothesis.
DIVAs is a machine-learning (ML) model exploiting different features to describe each mutated digenic combination capturing gene-gene interaction, single variant impact, a priori genes properties and genes association to HPO-based phenotypes. Family analysis can also be performed.