The field of molecular interaction analysis is witnessing significant advancements, particularly in the areas of DNA-encoded library screening, molecular relational learning, and gene-metabolite association prediction. Innovations in denoising frameworks for DEL screening are enhancing the accuracy of binding affinity predictions, addressing the noise issues inherent in read counts. This is crucial for the efficient exploration of chemical spaces and the identification of potential binding motifs. In molecular relational learning, the development of flexible toolkits like FlexMol is streamlining the benchmarking process, enabling a more comprehensive and fair comparison of diverse model architectures. This is pivotal for advancing our understanding of molecular interactions in drug discovery. Additionally, the integration of interactive knowledge transfer mechanisms in gene-metabolite association prediction is revolutionizing metabolic engineering by automating candidate gene discovery and improving prediction accuracy across different microorganisms. These developments collectively push the boundaries of current methodologies, offering more precise and efficient tools for future research in these domains.