The life science domain in a high value research area, both in terms of the benefits in increased knowledge and in societal impact. Much of the research funding has focused on wet lab based approaches to increase visibility into biological processes and producing maximal relevant information on which to make decisions. Given the complexity of biological functions, in many cases this has led to an information overload. Researchers are now able to routinely generate and access petabytes of data as a result of high throughput experiments, and this capability is growing. This data can be difficult to interpret and intractable for manual evaluations, proffering the need for powerful and accurate bioinformatics tools so that researchers and practitioners can actually make use of the information being generated in a practical sense. Artificial Neural Networks are a machine learning approach which has gained much traction in the field of bioinformatics, as they offer the required high throughput processing for large datasets, while providing powerful generalization, fault tolerance, and robustness to noise, making them appealing for application to life science problems. Major contributions of this thesis include literature reviews that demonstrate the use, effectiveness and limitations of key machine learning technologies in life science, and the development of two novel neuroevolution approaches (MFF-NEAT and RBF-CGP-ANN) which were developed recognizing needs of life sciences, and addressing issues inherent in the application of artificial neural networks to bioinformatics problems. Comprehensive experiments were conducted to gauge the effectiveness of these new tools on life science problems, including breast cancer diagnosis, heart disease, mass spectral datasets, and determining the specificity of HIV-1 protease. The results achieved are discussed, and it is demonstrated that these new tools have the potential to outperform more typical ANN based approaches on specific tasks - (Abstract)

Thesis Cork Institute of Technology, 2017.

Bibliography: (pages 312-351)