Whole exome sequencing is the preferred strategy to identify the genetic defect in patients with a probable or possible mitochondrial cause

Mitochondrial disorders, characterized by clinical symptoms and/or OXPHOS deficiencies, are caused by pathogenic variants in mitochondrial genes. However, pathogenic variants in some of these genes can lead to clinical manifestations which overlap with other neuromuscular diseases, which can be caused by pathogenic variants in non-mitochondrial genes as well. Mitochondrial pathogenic variants can be found in the mitochondrial DNA (mtDNA) or in any of the 1,500 nuclear genes with a mitochondrial function. We have performed a two-step next-generation sequencing approach in a cohort of 117 patients, mostly children, in whom a mitochondrial disease-cause could likely or possibly explain the phenotype. A total of 86 patients had a mitochondrial disorder, according to established clinical and biochemical criteria. The other 31 patients had neuromuscular symptoms, where in a minority a mitochondrial genetic cause is present, but a non-mitochondrial genetic cause is more likely. All patients were screened for pathogenic variants in the mtDNA and, if excluded, analyzed by whole exome sequencing (WES). Variants were filtered for being pathogenic and compatible with an autosomal or X-linked recessive mode of inheritance in families with multiple affected siblings and/or consanguineous parents. Non-consanguineous families with a single patient were additionally screened for autosomal and X-linked dominant mutations in a predefined gene-set. We identified causative pathogenic variants in the mtDNA in 20% of the patient-cohort, and in nuclear genes in 49%, implying an overall yield of 68%. We identified pathogenic variants in mitochondrial and non-mitochondrial genes in both groups with, obviously, a higher number of mitochondrial genes affected in mitochondrial disease patients. Furthermore, we show that 31% of the disease-causing genes in the mitochondrial patient group were not included in the MitoCarta database, and therefore would have been missed with MitoCarta based gene-panels. We conclude that WES is preferable to ...