Pharmacophore Modelling: Compound Optimization and Similarity Analysis

In advance of building pharmacophore models, structures of 15 compounds were drawn in the ACD/ChemSketch program (freeware) 2020 1.2 (Hunter, 1997 ) and submitted to geometry optimization in the program BIOVIA Discovery Studio Visualizer (v 17.2.0.16349) (Biovia et al., 2000 (link)). The force field used was the MM+ (Molecular Mechanics), according to the methodological strategy proposed by da Silva Costa et al. (2018) (link); afterwards the structures underwent refinement by using the Dreiding-like force field (Hahn, 1995 (link)).
After optimization of compounds, their structures were inputted in the BIOVIA Discovery Studio Visualizer (v17.2.0.16349) and gathered into a single file (mol2). Then, this file was submitted to the BindingDB webserver (https://www.bindingdb.org/bind/index.jsp) for calculation of similarities values by means of Tanimoto index (TI) (Liu et al., 2007 (link)). TI values (Eq. 1) varies between 0 and 1, representing the overall similarity between two compounds based on their fingerprint bits (molecular fragments), so that the closer to 1, greater the similarity (Gimeno et al., 2019 (link)). $Tanimoto Index=\frac{\mathit{c}}{\left(\mathit{a}+\mathit{b}-\mathit{c}\right)\mathit{ }}$ Where, for two generic compounds A and B: a: number of bits in A; b: number of bits in B; c: number of common bits between A and B.