An alignment‐free methodology for modelling field‐based 3D‐structure activity relationships using inductive logic programming

Abstract

Traditional 3D-quantitative structure-activity relationship (QSAR)/structure-activity relationship (SAR) methodologies are sensitive to the quality of an alignment step which is required to make molecular structures comparable. Even though many methods have been proposed to solve this problem, they often result in a loss of model interpretability. The requirement of alignment is a restriction imposed by traditional regression methods due to their failure to represent relations between data objects directly. Inductive logic programming (ILP) is a class of machine-learning methods able to describe relational data directly. We propose a new methodology which is aimed at using the richness in molecular interaction fields (MIFs) without being restricted by any alignment procedure. A set of MIFs is computed and further compressed by finding their minima corresponding to the sites of strongest interaction between a molecule and the applied test probe. ILP uses these minima to build easily interpretable rules about activity expressed as pharmacophore rules in the powerful language of first-order logic. We use a set of previously published inhibitors of factor Xa of the benzamidine family to discuss the problems, requirements and advantages of the new methodology.

An Evolutionary Algorithm for de Novo Optimization of Functional Transition Metal Compounds

“…An alignment step, such as used in the comparative molecular field analysis (CoMFA), would make the optimization less efficient. However, there are alignment-free 3D QSAR methods that might be suitable for this purpose, such as grid-independent descriptors (GRIND), , VolSurf, and those based on inductive logic programming. − …”

Machine Learning Methods for Property Prediction in Chemoinformatics:Quo Vadis?

“…The main advantage of ILP stems from its ability to provide relational learning and therefore to treat structured input data of any complexity, including molecular graphs. ILP has successfully been applied to mutagenicity , and toxicity prediction, pharmacophore discovery, classification of bioactive chemical compounds, scaffold hopping in drug discovery, building ordinary , and field-based 3D QSAR models, etc.…”

“…Modern machine learning methods offer a unique opportunity to work directly with the connectivity matrix (entry 8). One can also mention different graph mining approaches, ,,, the use of graph kernels, ,,, the Inductive Logic Programming (ILP), , and its application to chemoinformatics. − , …”

“…One can also mention different graph mining approaches, 95,213,216,219 the use of graph kernels, 33,34,37,40 the Inductive Logic Programming (ILP), 141,142 and its application to chemoinformatics. [145][146][147][148][149][150]152 A chemical structure is a complex object which, sometimes, must be represented by several graphs (tautomers) or 3D structures (conformers). The multi-instance learning approach (entry 9) could be very helpful in taking this into account in structure−property modeling.…”