Machine learning prediction of the potential pesticide applicability of three dihydroquinoline derivatives: Syntheses, crystal structures and physical properties Wesley F. Vaz a, b , Giulio D.C. D’Oliveira c , Caridad N. Perez c , Bruno J. Neves c , Hamilton B. Napolitano a, d, * a Universidade Estadual de Goi as, 75132-400, An apolis, GO, Brazil b Instituto Federal de Educaç~ ao, Ci^ encia e Tecnologia de Mato Grosso, 78455-000, Lucas do Rio Verde, MT, Brazil c Universidade Federal de Goi as, 74690-900, Goi^ ania, GO, Brazil d Centro Universit ario de An apolis, 75075-010, An apolis, GO, Brazil article info Article history: Received 15 October 2019 Received in revised form 7 January 2020 Accepted 13 January 2020 Available online 22 January 2020 Keywords: Machine learning model Tobacco mosaic virus Fusarium oxysporum dihydroquinoline Crystal structure abstract Increasingly machine learning processes have been applied in the search and development of com- pounds that may have specific physicochemical properties to the desired application. This article de- scribes how a machine learning model led us to the synthesis of three dihydroquinoline derivatives with potential application as a pesticide. The synthesized compounds were predicted to be active against the Tobacco mosaic virus (y 90%) and Fusarium oxysporum (y 78%). Regarding a correlation between the pesticide activity and the molecular structure, the new dihydroquinoline derivatives were structurally characterized using spectroscopic techniques and single crystal X-ray diffraction. They crystallized into orthorhombic (I) and monoclinic (II and III) crystal systems with supramolecular arrangements main- tained primarily by non-classical CeH/O hydrogen bonds, which form dimers and chains in their molecular packaging. Frontier molecular orbitals and molecular electrostatic potential maps were un- dertaken using density functional theory in order to study the electronic properties of the observed molecular conformations. Finally, the developed approach is a useful tool on new pesticide investigation when experimental toxicity data are not available. © 2020 Elsevier B.V. All rights reserved. 1. Introduction Pesticides include herbicides, insecticides, fungicides, fumi- gants, and rodenticides and offer significant benefits in public health and food production. Unlike most other significant chem- icals, pesticides are designed to impact living systems [1]. As a result, there is a worry regarding the environmental and human consequences of extensive pesticide use [1 ,2]. Several harmful ef- fects on the overuse of pesticides include (1) potential destruction of biodiversity; (2) problems on environmental sustainability and global stability [2]; and (3) potential contamination on food, water and the environment [3]. Therefore, it is important to develop new compounds less harmful to living systems and machine learning is a powerful methodology regarding its wide use of molecular en- gineering applied to design compounds with desired properties [4]. Machine learning models have long been employed for drug dis- covery utilizing docking studies, virtual screening, molecule syn- thesis, small molecule physicochemical properties, solubility and beyond [5]. Moreover, machine learning methods have also been used to discover solutions to environmental issues, such as municipal solid waste management [6], soil/compost properties with bioavailability and risk assessment [7], to model the bioavailability of contaminants [8], and to obtain new antifungal and antiviral agents in agriculture [9e12]. To propose new compounds with biological potential, we developed a machine learning model from a training set of mole- cules with known activity. In this case, was assumed that the po- tential pesticide is associated with the chemical structure. The obtained model shows that alpha-beta unsaturated systems, sul- fonamide group, and the nitro group are frequently founded in * Corresponding author. Universidade Estadual de Goi as, 75132-400, An apolis, GO, Brazil. E-mail address: hamilton@ueg.br (H.B. Napolitano). Contents lists available at ScienceDirect Journal of Molecular Structure journal homepage: http://www.elsevier.com/locate/molstruc https://doi.org/10.1016/j.molstruc.2020.127732 0022-2860/© 2020 Elsevier B.V. All rights reserved. Journal of Molecular Structure 1206 (2020) 127732