Suzanne de Bruijn

303 General discussion and perspectives Examples are mitochondrial inheritance and somatic mosaicism that could lead to mosaic presence of the pathogenic variant in different cellular subtypes. A case study performed by de Kok et al. reported somatic mosaicism for a POU3F4 variant (associated with HL type DFNX2); semiquantitative analysis revealed that the variant was present in approximately 50% of peripheral blood lymphocytes of the affected individual. 120 The authors hypothesized that most likely, the variant occurred in early embryogenesis before occurrence of the otic vesicle. Mitochondrial DNA is not routinely investigated in clinical diagnostics, however, pathogenic variants in mitochondrial genes have been described for both HL and RD. 31,121 Based on current observations, for the many genetically unexplained isolated cases diagnosed with HL and RD, a non-Mendelian inheritance pattern should be considered. To readily implement this in genetic diagnostics, data analysis pipelines and workflows for sequencing data should be adapted. Variants considered too common should no longer be readily disregarded as they could be part of a complex oligogenic or polygenic inheritance pattern. In case of monoallelic pathogenic variants, one should not only focus on the mutated gene, but should also consider pathogenic variants in other disease-associated (e.g. pathogenic FOXI1 variants for monoallelic SLC26A4 patients, chapter 5 ). Digenic inheritance has also been suggested for other HL- or RD-associated genes (e.g. CDH23 / PCDH15 122 and CEP250 / PCARE 123 ) which, however, could not be confirmed due to the lack of additional cases. 124 Last but not least, to make this process feasible, a complete work-up of family history is required, and careful clinical evaluation of all family members (if possible), including unaffected siblings, should be performed to allow to establish phenotype-genotype correlations. As is clear from the above, there is an important need to start looking beyond “standard” Mendelian genetics. If we do this, hopefully, a significant part of genetically unsolved cases will receive a genetic, a non-genetic or combinatory explanation for their phenotype. FINAL REMARKS With the developments discussed in the previous sections (multi-omics approaches, single-cell sequencing technologies and non-Mendelian inheritance patterns), a significant portion of missing heritability could be resolved. However, how realistic is a fast implementation of these technologies in clinical practice? Implementation often requires collection of additional tissue biopsies, (fresh) blood samples, the establishment of new lab workflows and analytic pipelines. 125 Incorporation of the new technologies and approaches in standard procedures will require extreme efforts that are time- and money-consuming. Also, once implemented, the increased amount of data that need