Allosteric modulators have the potential to fine-tune protein functional activity. Therefore, the targeting of allosteric sites, as a strategy in drug design, is gaining increasing attention. Currently, it is not trivial to find and characterize new allosteric sites by experimental approaches. Alternatively, computational approaches are useful in helping researchers analyze and select potential allosteric sites for drug
discovery. Here, we review state-of-the-art computational approaches directed at predicting putative allosteric sites in proteins, along with examples of successes in identifying allosteric sites utilizing these methods. We also discuss the challenges in developing reliable methods for predicting allosteric sites and tactics to resolve demanding tasks.