Ligand-Based Pharmacophore Modeling Using Novel 3D Pharmacophore Signatures.

Alina Kutlushina,Pavel Polishchuk,Timur Madzhidov,Aigul Khakimova

doi:10.3390/molecules23123094

Abstract

Pharmacophore modeling is a widely used strategy for finding new hit molecules. Since not all protein targets have available 3D structures, ligand-based approaches are still useful. Currently, there are just a few free ligand-based pharmacophore modeling tools, and these have a lot of restrictions, e.g., using a template molecule for alignment. We developed a new approach to 3D pharmacophore representation and matching which does not require pharmacophore alignment. This representation can be used to quickly find identical pharmacophores in a given set. Based on this representation, a 3D pharmacophore ligand-based modeling approach to search for pharmacophores which preferably match active compounds and do not match inactive ones was developed. The approach searches for 3D pharmacophore models starting from 2D structures of available active and inactive compounds. The implemented approach was successfully applied for several retrospective studies. The results were compared to a 2D similarity search, demonstrating some of the advantages of the developed 3D pharmacophore models. Also, the generated 3D pharmacophore models were able to match the 3D poses of known ligands from their protein-ligand complexes, confirming the validity of the models. The developed approach is available as an open-source software tool: http://www.qsar4u.com/pages/pmapper.php and https://github.com/meddwl/psearch.

Highlights

Pharmacophore models have proved to be useful for the selection of focused sets of compounds [1,2,3,4].There are two kinds of pharmacophores: (i) structure-based pharmacophores derived directly from X-ray structures of protein-ligand complexes, and (ii) ligand-based pharmacophores derived from structures of known active compounds
We developed a new approach for 3D pharmacophore representation and matching which does not require alignment of compounds/pharmacophores to find a common pharmacophore
The developed 3D pharmacophore signatures can be used for ligand-based pharmacophore modeling that does not require pre-defined geometry of active compounds to be used as a template, or the explicit alignment of pharmacophores

Summary

Introduction

Pharmacophore models have proved to be useful for the selection of focused sets of compounds [1,2,3,4]. There are two kinds of pharmacophores: (i) structure-based pharmacophores derived directly from X-ray structures of protein-ligand complexes, and (ii) ligand-based pharmacophores derived from structures of known active compounds. There are many ligand-based modeling tools, but almost all of them are commercial (LigandScout, Discovery Studio, MOE, PHASE, etc.) [5]. These programs use different algorithms for common pharmacophore identification based on genetic optimization [6,7,8], clique detection [9]

Objectives

Methods

Results

Conclusion