Fig. 8

Representation of the virtual screening flow and its two phases (phase 1 and phase 2 screening). Orange arrows indicate the ‘training’ pass using the DUD-E dataset as training set, and blue arrows the ‘screening’ pass with the Enamine HTS library as input set. The generated output consists of a list of spectrophores (corresponding to Enamine molecules) labeled with their most likely pharmacological target. In total, the 93 spectrophores were assigned to 32 different labels (out of the 102 possibilities)