Fuzzy Tricentric Pharmacophore Fingerprints. 1. Topological Fuzzy Pharmacophore Triplets and Adapted Molecular Similarity Scoring Schemes

Abstract

Topological fuzzy pharmacophore triplets (2D-FPT), using the number of interposed bonds to measure separation between the atoms representing pharmacophore types, were employed to establish and validate quantitative structure-activity relationships (QSAR). Thirteen data sets for which state-of-the-art QSAR models were reported in literature were revisited in order to benchmark 2D-FPT biological activity-explaining propensities. Linear and nonlinear QSAR models were constructed for each compound series (following the original author's splitting into training/validation subsets) with three different 2D-FPT versions, using the genetic algorithm-driven Stochastic QSAR sampler (SQS) to pick relevant triplets and fit their coefficients. 2D-FPT QSARs are computationally cheap, interpretable, and perform well in benchmarking. In a majority of cases (10/13), default 2D-FPT models validated better than or as well as the best among those reported, including 3D overlay-dependent approaches. Most of the analogues series, either unaffected by protonation equilibria or unambiguously adopting expected protonation states, were equally well described by rule-or pK a -based pharmacophore flagging. Thermolysin inhibitors represent a notable exception: pK a -based flagging boosts model quality, althoughssurprisinglysnot due to proteolytic equilibrium effects. The optimal degree of 2D-FPT fuzziness is compound set dependent. This work further confirmed the higher robustness of nonlinear over linear SQS models. In spite of the wealth of studied sets, benchmarking is nevertheless flawed by low intraset diversity: a whole series of thereby caused artifacts were evidenced, implicitly raising questions about the way QSAR studies are conducted nowadays. An in-depth investigation of thrombin inhibition models revealed that some of the selected triplets make sense (one of these stands for a topological pharmacophore covering the P 1 and P 2 binding pockets). Nevertheless, equations were either unable to predict the activity of the structurally different ligands or tended to indiscriminately predict any compound outside the training family to be active. 2D-FPT QSARs do however not depend on any common scaffold required for molecule superimposition and may in principle be trained on hand of diverse sets, which is a must in order to obtain widely applicable models. Adding (assumed) inactives of various families for training enabled discovery of models that specifically recognize the structurally different actives.

ISIDA Property‐Labelled Fragment Descriptors

“…Next, the atomic pharmacophoric types are attributed by the pharmacophore mapping tool PMapper, according to the following custom rules: Note that, like in previously reported work, [34] the pharmacophore flagging rules are specifically applied to automatically generated microspecies in which formal charges had been explicitly assigned to cations and anions, and therefore differ from the ones expected to apply to default neutral representations of compounds.…”

“…By contrast, the rather obscure ECFP numbering scheme requires some quite tedious decoding. The importance of rational, pH-sensitive pharmacophore flagging has been extensively discussed in the context of fuzzy pharmacophoric triplets [34,51] and will not be revisited here. Also, unlike in ECFPs, augmented atom radii are user-defined, including the lower bound, and one can exclude eventually unwanted small fragments.…”

“…Furthermore, in order to ensure for a chemically relevant colouring, pH-dependent typing [34] is performed. Except for symbol-based colouring, pharmacophore and electrostatic property-based flags may differ, and will be rendered as a function of the actual protonation scheme of the compound at given pH.…”

“…Therefore, we generically refer to our descriptors as ISIDA property-labelled fragment descriptors (IPLFs). The list of fragments seen to occur in a molecule (respectively the combined list of all fragments found in all its microspecies being populated at the given pH, for pH-dependent flagging [34] ) and their cumulated population levels define the molecular fingerprint, to be used in similarity scoring and QSAR studies.…”

Pharmacophores in Drug Research

“…The concept of fuzzy 3D pharmacophores has been successfully exemplified in a study aimed at the discovery of novel ligands for RNA targets. [39] Translating the approach of fuzzy pharmacophores for the calculation of 2D fingerprints has been proposed by Horvath et al [40] introducing a novel molecular description, the topological (2D) fuzzy pharmacophore triplets: 2D-FPT. This method is using the number of interposed bonds as the measure of separation between the atoms representing pharmacophore types (hydrophobic, aromatic, hydrogenbond donor and acceptor, cation, and anion).…”