Logico-Structural Approach to Computer-Assisted Drug Design

Fragment Descriptors in SAR/QSAR/QSPR Studies, Molecular Similarity Analysis and in Virtual Screening

“…The non-quantitative SAR (Structure-Activity Relationships) models developed in the 1970s by Hiller, 37,38 Golender and Rosenblit, 39,40 Piruzyan, Avidon et al, 41 Cramer, 42 Brugger, Stuper and Jurs, 43,44 and Hodes et al 45 were inspired by the, at that time, popular artificial intelligence, expert systems, machine learning and pattern recognition paradigms. In those approaches, chemical structures were described by means of indicators of the presence of structural fragments interpreted as topological (or 2D) pharmacophores (biophores, toxophores, etc.)…”

“…MCSs were first applied to SAR studies in the early 1980s by Rozenblit and Golender in the framework of their logical-combinatorial approach. 40,41,147 Since at that time computer power was limited, the authors suggested the use of reduced graphs (Section 1.3.5) built on pharmacophoric centers. The MCS fragments were subsequently applied to perform a similarity search, 148 to cluster chemical databases 149,150 as well to assess biological activities of organic compounds.…”

Substructural QSAR approaches and topological pharmacophores.

“…To this end a basic library of potentially active centers such as atoms or groups of atoms (functional groups) likely to be involved in drug-receptor interaction via van der Waals or other forces may be set up and used to create the concrete features for each particular problem via a set of rules (a language) in an open-ended way. Cases in point are the SSFN and the DCAM systems that are based on the predefined "descriptor centers" and distances between them (22)(23)(24).…”

“…An extended version of this procedure has successfully been applied to mutagenicity data by Tinker as a predictive test for hazard evaluation (43,44). Philosophically similar is the ORACLE procedure (22,24), which was developed for the purpose of guiding the screening of compounds in a battery of biological tests. The basis of ORACLE is a structure and pharmacological activity file including about 6000 compounds and 55 major types of pharmacological activities.…”

“…The system correctly recognized the presence of earlier established activities for the majority of compounds in the data base, and a num-ber of unknown activities could be predicted for several compounds. Another program of the index type which, however, is already very close to classification techniques is the STRAC procedure designed for lead optimization (22). Again, the compounds of the starting set are divided into the classes "active" and "inactive" and described by topological features.…”