PhD defence by Joshua Rubin

On Friday 14 February, Joshua Rubin will defend his PhD thesis "Modern and Ancient Hominin Panmitogenomics".

Time: 13:00
Place: Building 101, meeting room 1 & zoom: https://dtudk.zoom.us/meeting/register/PmoNVs-oTZG5-aTS8HU9JA
Please be aware that the PhD defense may be recorded - This will also be informed at the beginning of the PhD defense.

Supervisor: Professor Gabriel Renaud
Co-supervisor: Associate Professor Shyam Gopalakrishnan, University of Copenhagen

Assessment committee:
Associate Professor Henrik Nielsen, DTU Health Tech
Assistant Professor Kristoffer Sahlin, Stockholm University
Associate Professor Vania Alves E Silva Pereira, University of Copenhagen

Chairperson:
Professor Anders Gorm Pedersen, DTU Health Tech

Abstract:
Due to the uniparental inheritance pattern, high copy number, and other properties of mitochondrial DNA, ancient DNA (aDNA) researchers place a premium on the mitochondria when it comes to shedding light on human history through analysis of next-generation sequencing data. However, ancient DNA analysis is more complex than modern DNA (particularly for hominin samples, where contamination risk is high), and methods of inference on mitochondrial aDNA samples have been limited in addressing these complexities. In this thesis we make headway on improving inference methods for mitochondrial aDNA samples of hominin origination, in a number of ways. First, we improve the task of haplogroup assignment (i.e. assigning a single-source sample to its population of origin) by introducing novel methods making use of a graph-theoretic reference structure (pangenomics) in conjunction with Bayesian inference techniques to ensure that the methodology has a solid mathematical foundation, and to further mitigate ethnicity-related biases. Second, we tackle the issue of chemical damage leading to forgone alignments of aDNA to the pangenome by employing a bespoke DNA sequence transformation designed to be unchanged by post-mortem chemical damage. Finally, we make headway on mixture samples comprising multiple mitochondrial sources (be they ancient endogenous material or modern contamination) by appropriately generalizing our Bayesian inference model, to be able to estimate not only the placement of each contributing source but also the mixture proportions. The deliverables take the form of software tools written in C++ which can be used by the community. Taken as a whole, these projects allow researchers to gain better insight into our collective human past.

A copy of the PhD thesis is available for reading at the department