Breakthroughs in computational biology drive new frontiers in drug discovery and protein design. The DeepMice model introduces multilevel mapping to optimize protein-ligand docking, improving virtual screening accuracy. Concurrently, physic-based algorithms expand custom biomolecule design by controlling intrinsically disordered protein ensembles. These advances promise to accelerate rational drug design and biomolecular engineering, integrating AI with molecular insights to tackle complex therapeutic targets.