Chemometrics on its way towards personalized health care

Abstract

Nowadays, patients can be examined in unprecedented detail at molecular level. It is expected that this will allow for personalized health care, where diagnosis and treatment are catered towards the unique molecular characteristics of an individual patient. However, currently complexity and size of the data are a major obstacle for analysis. The main contributions of this thesis consist of a careful evaluation of current strategies to remove artefacts from data and the development of novel data analysis methods for personalized health care. These methods can be used to detect differences between groups of patients to obtain a better understanding of biomolecular processes, or to inspect individual patients and (with knowledge of the biological processes) diagnose a disease. The new models were successfully used to investigate effects of diets on metabolism of patients with metabolic syndrome. Additionally, the methods were used to inspect the health status of individual patients diagnosing several rare metabolic diseases.