Evaluating phenotype-driven approaches for genetic diagnoses from exomes in a clinical setting

Abstract

Next generation sequencing is transforming clinical medicine and genome research, providing a powerful route to establishing molecular diagnoses for genetic conditions; however, challenges remain given the volume and complexity of genetic variation. A number of methods integrate patient phenotype and genotypic data to prioritise variants as potentially causal. Some methods have a clinical focus while others are more research-oriented. With clinical applications in mind we compare results from alternative methods using 21 exomes for which the disease causal variant has been previously established through traditional clinical evaluation. In this case series we find that the PhenIX program is the most effective ranking the true causal variant at between 1 and 10 in 85% of these cases. This is a significantly higher proportion than the combined results from five alternative methods tested (P=0.003). The next best method is Exomiser (hiPHIVE), in which the causal variant is ranked 1-10 in 25% of cases. The widely different targets of these methods (more clinical focus, considering known Mendelian genes, in PhenIX, versus gene discovery in Exomiser) is perhaps not fully appreciated but may impact strongly on their utility for molecular diagnosis using clinical exome data.

More information

Identifiers

ORCID for Reuben J. Pengelly: orcid.org/0000-0001-7022-645X

ORCID for Sarah Ennis: orcid.org/0000-0003-2648-0869

ORCID for Andrew Collins: orcid.org/0000-0001-7108-0771

Catalogue record

Export record

Altmetrics