Filtering promiscuous compounds in early drug discovery: is it a good idea?

The use of computational filters for excluding supposedly nonspecific and promiscuous compounds from chemical libraries is a controversial issue, because many drugs used in clinics today would never reach the market if these filters were applied. In part, this conflict could be caused by the paradigm: one-drug-one-target, even though it is widely agreed that drug action is a result of a complex network of biomolecular interactions. Therefore, the so-called pan assay interference compounds (PAINS) or promiscuous compounds could be in fact assay artifacts, false positives or, simply, bright chemical matter (BCM) composed of privileged scaffolds, as we propose here. Despite apparent promiscuity, BCM can be tailored into new and safe drugs after overcoming selectivity criteria.