Molecular Informatics in Natural Products Research
Johannes Kirchmair
Abstract
Natural products (NPs) remain the single most prolific source of inspiration in the development of functional small molecules, in particular of drugs.1 Researching NPs, however, is a non-trivial task, beginning with the sourcing, transfer and handling of materials for testing, and extending to the production of extracts, isolation of compounds, biological testing and identification of viable synthetic routes.2, 3 In this context, a lot of hope lies in the use of computational methods to provide guidance to experimentalists, allowing them to direct the available resources to the most promising avenues of research. Recent surveys have found a rich body of literature on the successful and broad application of cheminformatics approaches in NP-based drug discovery.4, 5 The most common applications include (i) data analysis and visualization, (ii) NP dereplication, (iii) quantification of NP-likeness, (iv) generation of synthetically accessible mimics of NPs, (v) analysis and prediction of the structural basis for the interaction of NPs with proteins, (vi) virtual screening for bioactive NPs, (vii) prediction of the macromolecular targets of NPs (target prediction), and (viii) prediction of ADME properties and toxicity. However, the structural complexity of many NPs (3D molecular shape complexity, stereochemistry, ring complexity, etc.) poses significant challenges to computational approaches, and the fact that most in silico approaches are designed for, or trained on, synthetic compounds is all too often overlooked. The challenges and opportunities involved in computer-guided NP research have fueled extensive interest in the development and application of new in silico methods, and we consider it timely to dedicate a special issue of Molecular Informatics to the topic. The contributions of this special issue cover (i) different methods for the analysis and visualization of the chemical space of NPs based on full molecular structures, scaffolds and fragments, (ii) the development and analysis of virtual collections of NPs and NP fragments, (iii) a review of pharmacophore-based approaches and their application to NP-based drug discovery, and (iv) examples of the successful application of target prediction methods and docking in the discovery and research of bioactive NPs. On behalf of all contributors to this special issue, I wish you an enjoyable and interesting read!