Analog Accessibility\nScore (AAscore) for Rational\nCompound Selection

Various <i>in silico</i> scores have been proposed\nto\nobjectively assess the characteristics and properties of a compound.\nHowever, there is still no score that represents the analog accessibility\nof a compound. Such a score would be valuable for selecting compounds\nproposed by virtual screening or for prioritizing hit compounds for\nthe hit-to-lead phase. This study proposes an analog accessibility\nscore (AAscore), where retrosynthesis prediction and forward product\nprediction models were utilized to generate virtual analogs. The AAscore\nis defined as the number of unique analogs and virtual synthetic routes.\nTo evaluate the AAscore in terms of the number of actually synthesized\nanalog compounds, analog compounds were prepared by using the compound-core\nrelationship (CCR) method. It was found that the AAscore was little\ncorrelated with the number of CCR-based analogs. Furthermore, AAscores\nwere found to be significantly influenced by the number of extracted\ncandidate reactants from a reactant database. A case study targeting\ncompounds active against carbonic anhydrase 2 showed that the AAscore\ncould identify compounds that were synthesized into analogs.