Linking Artificial Intelligence and Medical Data to Revolutionize Targeted Therapy

<p dir="ltr">The drug discovery process is fundamentally challenged by its complexity and the extensive time and financial resources it demands. Spanning over a decade, this process is marked by high costs and a high rate of failure, primarily due to the complex biological systems involved and the difficulty in identifying effective therapeutic targets. To address these issues and enhance the drug discovery pipeline, particularly in the target discovery phase, this thesis explores the integration of Artificial Intelligence (AI). AI’s capability to handle large datasets and derive predictive insights offers a promising avenue to accelerate and refine the identification and validation of drug targets.</p><p dir="ltr">This thesis introduces several novel AI-driven approaches that address three key challenges within the target discovery phase: (i) enhancing the accuracy and efficiency of identifying potential drug targets through advanced AI algorithms; (ii) overcoming the complexity of biological interactions that inhibit effective drug target validation; and (iii) reducing the time and cost associated with these phases through automation.</p><p dir="ltr">The novelty in this research lies in the development of AI methodologies that:</p><p dir="ltr">Multi-modal AI Integration: Utilize AI to perform a more comprehensive analysis of genomic, transcriptomic, and proteomic data, integrating multi-omic datasets to uncover deeper biological insights and identify potential drug targets with greater precision than traditional methods. Graph-based Predictive Modelling: Apply novel graph neural networks and other predictive models that account for complex biological interactions, enabling more accurate validation of drug targets by modeling the dynamic relationships between genes, proteins, and pathways. AI-Driven Workflow Automation: Implement AI to automate traditionally labor-intensive steps in target validation, significantly speeding up the process and reducing operational costs. The experimental results demonstrate that these novel AI-enhanced approaches not only accelerate the target discovery phase but also improve the precision and reliability of drug target identification and validation, outperforming traditional techniques in both speed and accuracy. The findings of this thesis present a significant advancement in the drug discovery process and have the potential to streamline various phases of pharmaceutical research, contributing to more effective and efficient drug development practices.</p>