Modern and Ancient Hominin Panmitogenomics

Beyond the double-helical DNA in the nucleus, eukaryotes also harbor a separate genome in their mitochondria, the remnant of a primordial bacterium from billions of years ago. For a number of reasons, mitochondrial DNA (mtDNA) is particularly useful for ancient DNA researchers, allowing them to reconstruct population movements over time in great detail. Unfortunately, owing to chemical degradation and the ubiquity of contamination, ancient DNA (aDNA) is difficult to analyze, necessitating detailed mathematical modelling. Given these challenges, my research has focused on developing novel methods that directly address these issues plaguing the field. In this thesis, I outline my contributions to the field in the form of the development of software for improved inference of modern or ancient hominin mtDNA mixture samples. The work involves the innovative use of advanced reference genome data structures inspired by graph theory to more accurately represent genetic variation, and alignment of reads to said graphs in a way that takes into account post-mortem damage, in conjunction with Bayesian methods of inference and biology-driven modelling, to overcome the limitations of existing tools and achieve accurate and confident inference in a fully probabilistic way that is robust even when only trace amounts of DNA are recoverable. The deliverables are three C++ software tools (HaploCart, SAFARI, and TrailMix) which enable researchers to gain better understanding from their samples, and by extension, of our collective human heritage.