Computational approaches have become an integral part of modern drug discovery and medicinal chemistry. These approaches can be roughly classified into data/information mining (or filtering) and modelling/simulation methods. Taken together, they represent an ever growing source of hypotheses used to guide experimental approaches and hence drug discovery decisions. Therefore, it is not only important to optimally understand and apply existing methods, but also invest in the development of new algorithms to further improve our selection of drug candidate. The present contribution will describe a few approaches which have become routine at Novartis.