The maternally inherited mitochondrial DNA (mtDNA) is a multi-copy genome that encodes several mitochondrial proteins essential for energy production. Heteroplasmy, the state of possessing multiple variants of mtDNA, is very common in humans. The levels of these mtDNA variants are subject to genetic drift and selection during development and ageing. Disease symptoms will manifest when a detrimental mtDNA variant reaches a high level. To date, over 350 pathogenic mtDNA mutations have been identified as causing mtDNA-associated diseases, for which there are no cures. Given the role that mtDNA plays in disease, there is an increasing pressure to better understand the mechanisms underlying mtDNA selection. Previous studies of heteroplasmy dynamics during development have demonstrated tissue-specificity and age-dependence, but no clear mechanistic explanation has emerged, mainly due to a lack of animal models for systematic and gene-focused studies. Here, I developed a cell culture-based system for investigating mtDNA selection, based on a Drosophila melanogaster line that transmits a temperature-sensitive mtDNA variant alongside a functional mtDNA variant at a stable ratio. I generated multiple cell lines from embryos of the heteroplasmic flies and showed that they could maintain a balance of the two mtDNA variants which I could manipulate. The established heteroplasmic cell lines allowed me to perform high-throughput RNAi and compound screens to identify factors that affect the heteroplasmic ratio. By doing so, I demonstrated that the knockdown of a number of genes and the application of certain small molecules changed the heteroplasmic ratio over a few cell divisions. In particular, I revealed that RNAi-mediated knockdown of Regulatory particle triple-A ATPase 2 or Regulatory particle non-ATPase 11, two conserved components of the proteasome, was sufficient to increase the proportion of detrimental mtDNA. This indicated that 26S proteasome activity is involved in maintaining functional mtDNA levels and has important implications for neurodegenerative diseases, which are often associated with proteasome impairment. In this work, I also investigated mtDNA selection at the organismal level. A deficiency screen using the stable heteroplasmic fly line identified genes such as mtDNA polymerase PolG1 that influence the ratio of mtDNA variants over generations, providing an understanding of germline selection acting at the mitochondrion level mediated by PolG1 availability. I explored the role of individual genes in selective transmission of mtDNA by employing CRISPR/Cas9-mediated mutagenesis and found that whilst PolG1 can influence the ratio, there is no evidence that mtDNA replisome components PolG2 or mtDNA-helicase do so. Taken together, my work demonstrates that the nuclear genome plays a complex role in determining the outcome of mtDNA selection in cultured cells as well as in vivo. Furthermore, the small molecules I identified have the potential to be developed into novel therapeutic methods for lowering the level of detrimental mtDNA in patients. mtDNA selection within our bodies is not an inscrutability: it can be modelled in flies and, hopefully, soon manipulated in patients.