Principles for developing core genome multilocus sequence typing systems

Core genome multilocus sequence typing is a next generation typing method for the long-term tracking of pathogenic bacteria. Although such methods provide the very high discriminatory power required by public health agencies, they are prone to difficulties relating to data loss intrinsic to current DNA sequencing technologies. This thesis describes a framework for developing conservative, but powerful core genome multilocus sequencing systems. To this end, I developed a prototype scheme for Campylobacter jejuni consisting of 697 core genome loci identified through the analysis of 5,693 C. jejuni whole genome sequences. I surveyed the extent of missing data in the dataset, and studied optimizing number of genes to include in such a scheme. Using the information learned in the survey of missing data, I developed a system for predicting unknown alleles from core genome typing data. The principles learned through my research can be applied to develop robust methods of pathogen surveillance.