Statistical modelling of mitochondrial disease

Abstract

Mitochondrial DNA mutations are a major cause of disease in the human population. Understanding the disease associated with these mutations is complicated by heteroplasmy, the mixture of wild-type and mutated mitochondrial DNA. Heteroplasmy can vary between cells, tissues, and organs, and the disease associated individual mutations is hugely varied on account of this. The mitochondrial genome encodes critical proteins of the oxidative phosphorylation system and mutation leads to energy deficits in cells and a wide range of secondary effects. The central and peripheral nervous system are commonly affected in mitochondrial disease and quality of life for patients is severely impaired. Although pathogenic mitochondrial genetic mutations were first identified over twenty five years ago, little progress has been made in understanding the expected progression of disease in patients. The aim of this study was to use statistical modelling to further understanding of disease progression in mitochondrial DNA mutations. The Medical Research Council Mitochondrial Disease Cohort provided the majority of patient data. Patients had been assessed using the Newcastle Mitochondrial Disease Adult Scale, which facilitates quantitative research on mitochondrial disease burden. This project comprises studies of two of the most common mitochondrial DNA mutations. The first study concerns patients with the m.3243A>G mutation, the most common pathogenic point mutation, and considers the effect of age and heteroplasmy on disease progression. Prediction models of both overall disease burden and specific phenotypic features were developed. Important features of the patient cohort were also examined, including heteroplasmy in different tissues and differences in disease expression between sexes. The second study looks at patients with single large-scale mitochondrial DNA mutations. The effect of deletion size, location of the deletion on the genome, and heteroplasmy were investigated, and all three predictors were found to be significant in understanding disease progression.