Background: The human mitochondrial genome is transcribed as long strands of RNA containing multiple genes, which require post-transcriptional cleavage and processing to release functional gene products that play vital roles in cellular energy production. Despite knowledge implicating mitochondrial post-transcriptional processes in pathologies such as cancer, cardiovascular disease and diabetes, very little is known about the way their function varies on a human population level and what drives changes in these processes to ultimately influence disease risk. Here, we develop a method to detect and quantify mitochondrial RNA cleavage events from standard RNA sequencing data and apply this approach to human whole blood data from > 1000 samples across independent cohorts.

Results: We detect 54 putative mitochondrial RNA cleavage sites that not only map to known gene boundaries, short RNA ends and RNA modification sites, but also occur at internal gene positions, suggesting novel mitochondrial RNA cleavage junctions. Inferred RNA cleavage rates correlate with mitochondrial-encoded gene expression across individuals, suggesting an impact on downstream processes. Furthermore, by comparing inferred cleavage rates to nuclear genetic variation and gene expression, we implicate multiple genes in modulating mitochondrial RNA cleavage (e.g. MRPP3, TBRG4 and FASTKD5), including a potentially novel role for RPS19 in influencing cleavage rates at a site near to the MTATP6-COX3 junction that we validate using shRNA knock down data.

Conclusions: We identify novel cleavage junctions associated with mitochondrial RNA processing, as well as genes newly implicated in these processes, and detect the potential impact of variation in cleavage rates on downstream phenotypes and disease processes. These results highlight the complexity of the mitochondrial transcriptome and point to novel mechanisms through which nuclear-encoded genes can potentially influence key mitochondrial processes.