Development of a methodology of computational intelligence for robust composite biomarker discovery: targeting breakthrough in the therapeutic management of melanoma
Ανάπτυξη μιας μεθοδολογίας υπολογιστικής νοημοσύνης για την ανακάλυψη σύνθετων εύρωστων βιοδεικτών: στοχεύοντας στη θεραπευτική αντιμετώπιση του μελανώματος
KeywordsBioinformatics ; Composite biomarkers ; Melanoma ; Next generation sequencing ; Genomics ; Dermoscopy ; Skin imaging ; Data integration
Cancer is a complex and intricate disease, and the scientific community has been struggling for decades to identify any feebleness or rudimentary characteristics to discover effective treatments. Next generation sequencing technologies have eased the way for the systematic discovery of diagnostic biomarkers for cancers and other pathologies. Melanoma continues to be a rare form of skin cancer but causes the majority of skin cancer related deaths. For many years research has focused on the investigation of the pathogenesis leading to melanoma, with the aim of better understanding its complexity and the potential advancement of therapeutic strategies. In this PhD thesis a computational model for the integrated analysis of multi-source cancer datasets is proposed, using cutaneous melanoma as disease-model, in order to identify robust composite biomarkers that allow the classification between healthy and disease state. Along this road, for the first time primary cutaneous melanoma biopsies from Greek patients were subjected to whole exome sequencing and were analysed in order to derive their mutational profile landscape. Moreover, in the context of big data analytical methodologies, integration of exome sequencing and transcriptomic data was performed, in an attempt to achieve a multi-layered analysis and infer a tentative disease network for primary melanoma pathogenesis, offering deeper insight in the underlying mechanisms affected by melanoma and potentially contributing to the valuable effective epidemiological characterisation of this disease. This study exhibits a modular and distributed workflow that can integrate heterogeneous, multidimensional (omics, histological images and clinical) data for the multi-angled portrayal and classification of melanoma patients. All the proposed methodologies achieve satisfying performance through the proposed framework. The specific architecture aspires to lower the barrier for the introduction of personalised therapeutic approaches, towards precision medicine.