Using artificial neural networks to inform the exploitation of 'omic biomarkers in lung cancer diagnosis

Abstract

In this thesis we utilise a number of ’omics technologies to identify a potential biomarkers for a wide range of respiratory disease, namely chronic obstructive pulmonary disease and lung carcinoma. In addition to this, several bioinformatic and clinical informatic tools were developed to both supplement our efforts, and to aide individuals and institutions that supported the project. Chapter Two details the use of metabolomics to study changes within the molecular properties of various disease categories, to detect the presence of any potentially significant disease biomarkers by using Flow injection electrospray mass spectrometry (FIE-MS). From this, a total of nine mass-to-ion values were able to provide statistically significant separation between small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC) sputum samples. Out of these nine biomarkers, six were tentatively identified as being Phenylacetic acid, L-Fucose, Caprylic acid, Acetic acid, Propionic acid, and Glycine. In Chapter Three we investigate the salivary microbiome to determine it efficacy as a platform for the identification of respiratory disease. We found significant changes in the diversity of the diseased lung. Furthermore, we also found number of statistically significant bacteria including Preveotella; Haemophilus; and Peptostreptococcus. Moreover, the Staphylococcales family of bacteria correlated strongly with the FEV1/FVC of individuals suffering from COPD. In Chapter Four we explore the use of microarray technologies on pooled tumour samples. In this pilot study we identify both lung cancer and tissue-specific genes using high throughput microarray screening. Characterisation of our libraries demonstrated enrichment of both cancer and tissue-specific genes.

Date of Award	2022
Original language	English
Awarding Institution	Aberystwyth University
Supervisor	Luis Mur (Supervisor) & Chuan Lu (Supervisor)