Technical tools for computational drug repositioning

The large amount of data currently available on small molecules and their biological activity at various levels is a fundamental resource both in the search for drug candidates and to better understand the mechanisms behind their effects. Neuropharmacology, often dealing with some of the most complex mechanisms in human pathology, could particularly benefit from the availability of system-wide molecular data. Moreover, the neurochemical space (the space of small molecules that could have neurological activity) has been estimated to possibly include as many as about 6 × 1015 different molecules (Weaver and Weaver, 2011). For such reasons, techniques from statistics and Machine Learning (ML), especially suited to deal with complexity and large samples, have been widely adopted in the modern drug repositioning research (Murphy, 2011; Lavecchia, 2015).