Development of an Advanced Molecular Profiling Pipeline for Human Population Screening

Abstract

The interaction between a human’s genes and their environment is dynamic, producing phenotypes that are subject to variance among individuals and across time. Metabolic interpretation of phenotypes, including the elucidation of underlying biochemical causes and effects for physiological or pathological processes, allows for the potential discovery of biomarkers and diagnostics which are important in understanding human health and disease. The study of large cohorts has been pursued in hopes of gaining sufficient statistical power to observe subtle biochemical processes relevant to human phenotypes. In order to minimise the effects of analytical variance in metabolic profiling and maximise extractable information, it is necessary to develop a refined analytical approach to large scale metabolic profiling that allows for efficient and high quality collection of data, facilitating analysis on a scale appropriate for molecular epidemiology applications. The analytical methods used for the multidimensional separation and detection of metabolic content from complex biofluids must be made fit for this purpose, deriving data with unprecedented reproducibility for direct comparison of metabolic profiles across thousands of individuals. Furthermore, computational methods must be established for collating this data into a form that is suitable for analysis and interpretation without compromising the quality achieved in the raw data. These developments together constitute a pipeline for large scale analysis, the components of which are explored and refined herein with a common thread of improving laboratory efficiency and measurement precision. Complimentary chromatographic methods are developed and implemented in the separation of human urine samples, and further mated to separation and detection by mass spectrometry to provide information rich metabolic maps. This system is optimised to derive precision from sustained analysis, with emphasis on minimisation of sample batching thereby allowing the development of metabolite collation tools that leverage the chromatographic reproducibility. Finally, the challenge of metabolite identification in molecular profiling is conceptually addressed in a manner that does not preclude the further reinvention of the analytical approaches established within this thesis. In summary, the thesis offers a novel and practical analytical pipeline suitable for achieving high quality population phenotyping and metabolome wide association studies.