Compound pathway model to capture SAR progression: comparison of activity cliff-dependent and -independent pathways

A compound pathway model is introduced to monitor SAR progression in compound data sets. Pathways are formed by sequences of structurally analogous compounds with stepwise increasing potency that ultimately yield highly potent compounds. Hence, the model was designed to mimic compound optimization efforts. Different pathway categories were defined. Pathways originating from any active compound in a data set were systematically identified including compounds forming activity cliffs. The relative frequency of activity cliff-dependent and -independent pathways was determined and compared. In 23 of 39 different compound data sets that qualified for our analysis, significant differences in the relative frequency of activity cliff-dependent and -independent pathways were observed. In 17 of these 23 data sets, activity cliff-dependent pathways occurred with higher relative frequency than cliff-independent pathways. In addition, pathways originating from the majority of activity cliff compounds displayed desired SAR progression, reflecting SAR information gain associated with activity cliffs.