Research mainly focuses on digital platforms for intelligent pharma applications. Prediction of molecular properties, as well as the design of molecules with target properties, are highly important problems, that still need to be addressed. Complex and rich nature of molecules allows us to represent them as sequences, graphs, 3D objects, or high-dimensional descriptors, and to apply numerical methods in order to solve open problems in structural bioinformatics and chemoinformatics. Rapid accumulation of molecular data opened gates for machine learning to be applied for such representations and to derive powerful prediction models that outperform heuristical methods.