Advanced bioinformatics tools and artificial intelligence in genomic data analysis: application in cancer diagnosis

Abstract

Throughout the past decades, proteomics and transcriptomics have contributed to a large extent in unraveling the complex molecular mechanisms of cancer. The widespread acceptance of next-generation sequencing and high-throughput technologies has transformed our ability to study large-scale biological data, yet several challenges still remain — from heterogeneity, dimensionality of data to the multimodal information integration challenge. Addressing these challenges is critical in making enhanced cancer diagnosis, monitoring, and treatment a reality through the precision medicine paradigm. Parallel to that, artificial intelligence more specifically machine learning and deep learning — has emerged as a remarkable facilitator in biomedical research capable of extracting useful patterns from complex data structures. This thesis follows the integration of transcriptomic profiles with image-based data derived from protein interaction networks, represented as protein graph images. Through these various data sources, we aim to improve the accuracy and interpretability of cancer classification models.From actual datasets in The Cancer Genome Atlas (TCGA), we constructed and validated machine learning models with MLP, SVM, and KNN classifiers to detect cancer-relevant molecular signatures. Our integrated approach demonstrates the efficacy of applying omics data and image representation in early detection. The work contributes to the new field of computational oncology and provides a valuable starting point for researchers designing multi-modal, AI-enabled tools for personalized cancer diagnostics.