In-silico Molecular Docking: Shifting Paradigms in Pesticide Discovery

Food and nutritional security are of utmost importance for the burgeoning population in the country. On an average 15-20% of potential crop production is lost due to insects, pests, weeds, diseases, nematodes, rodents etc., thus plant protection efforts aim at minimizing crop losses. There are many techniques and technologies for insect-pest control including biological control, transgenic plants, cultural control, mechanical control, physical control and increasingly biopesticides1, but for many crop-pest-geography scenarios insecticides remain a critical component.

Globally, insects may be destroying an estimated 18-20% of the annual crop production (estimated value=>US$470 billion).2 Innovation of insect pest control tools has been a critical need for centuries and continues with an expanding global population and the longstanding threats from insect and insect-borne diseases3. Amongst different measures, chemicals quickly gained great popularity as an efficient, labour-saving and economic tool in pest management inmost agricultural sectors.4 In other words, the most frequent method of managing pests and diseases inmost agricultural sectors is through the application of pesticides.5 Strategic resources like pesticides are essential to the security of the country'sfood supply. Global figures show that after using pesticides, 35% of cash crops are lost annually; if pesticides are discontinued, this loss climbs to 70%.6 In addition to saving labor, lowering the price of agricultural products and increasing economic efficiency, the use of pesticides is crucial for several processes related to plant growth, regulation, harvesting, storage, transportation and processing.7 The efficacy of pesticide development has risen significantly with the adoption of computer technologies.8-9 One of the most representative computer techniques, molecular docking technology, can improve our capacity to address issues like pesticide molecular target identification, pesticide molecular design, pesticide resistance prediction, toxicological analysis and environmental safety risk assessment.10-14 During the early stages of pesticide creation, traditional methods such as similar synthesis, random screening and natural active agent simulation played a significant role.15-18 For example, the herbicides alachlor19, nitrofen20 and triadimefon21 were discovered as pesticides by random screening approaches. However, the limitations of using traditional methods to create pesticides are high blindness, low success rates, and prolonged development cycles, all of which severely restrict the amount of research and development that can be done on pesticides.

FIGURE 1: Cost of agrochemical development (dashed line) versus screening success (solid line)

Screening success ratio=1/number of compounds that need to be screened for each product found. Data adapted, in part, from other studies.22 The effective development of a new pesticide necessitates the synthesis and screening of over 159,000 chemicals at a cost of about $286 million, with an average period of 11.5 years from first synthesis to market introduction, according to internationally accepted statistics.22 Additionally, weed and pest resistance are becoming exacerbated due to increased chemical use. The creation of new chemical pesticides is vital because of the need to solve important issues including pesticide residues and the harmful effects of pesticides on non-target organisms.23-34 One of the key instruments for pesticide research and development, virtual screening technology with molecular docking at its core can compensate for the lack of traditional pesticide creation methods by substantially raising the screening success rate for pesticide lead compounds. For instance, Vaidya and associates35 screened the abscisic acid receptor agonist Opabactin from the ZINC database using the GLIDE docking approach. Another significant use of molecular docking is reverse docking, which is useful for screening chemicals for possible targets for protein pesticides. To some extent, the toxicity of pesticides can be mitigated in the early stages of pesticide production by using reverse docking to identify probable targets of first-to-compound chemicals. By examining the interaction between small-molecule ligands and receptor biomolecules, a theoretical technique called molecular docking is utilized to investigate the interaction between proteins and ligands. It can predict the binding mechanism and affinity strength.36-38 Thus, molecular docking has also been applied to the study of pesticide resistance mechanisms and the environmental detection of pesticides and their metabolites39-41. Molecular docking has investigated the use of several machine learning (ML) techniques within the past decade42-43. The most common method entails creating scoring functions to estimate a protein-ligand complex'sbinding affinity. These estimations are then applied to separate various chemicals and binding positions to identify genuine binders and estimate their binding mode. Because molecules are naturally represented as graphs (a collection of nodes or atoms connected by edges or bonds), a deep neural network-based method called deep graph learning can learn from graph-structured data-has been used more and more in this research.44-45 A so-called deep docking approach was recently proposed by Gentile et al.46 to expedite the virtual screening of large databases. This deep learning model uses docking and is based on a multilayer feed-forward neural network. Its goal is to correlate molecular fingerprints with docking scores of molecules. With the help of this technique, Tang et al.47 were able to speedup docking-based virtual screening and find a novel A2AR antagonist for extremely large molecular libraries. Tang et al.47 found a novel A2AR antagonist for enormous chemical libraries by using this strategy to speedup docking-based virtual screening. FIGURE 2: Workflow of virtual screening for different compounds adapted from other studies48 Molecular docking technology has emerged as a powerful and increasingly popular tool in pesticide development. However, alongside its numerous advantages, it also has certain limitations. These drawbacks are highlighted in various pesticide research articles; for instance, Chen encountered difficulties in obtaining virtual screening results using a single screening method.49 Additionally, the molecular docking program itself has inherent issues, such as the flexibility of the target protein and the accuracy of the scoring function. Although there have been significant advancements in improving the scoring function, accurately and quickly predicting receptor-ligand interactions continues to be a major challenge.50-51 Therefore, although docking experiments have made valuable contributions to our understanding of target-ligand interactions in drug discovery projects, their results should be viewed as preliminary and as a foundation for more comprehensive and accurate analyses.52 This article reviews the fundamental principles of molecular docking, available docking software, and pesticide-related databases, along with the challenges associated with molecular docking. We provide a summary of how this method is applied in pesticide development, discuss the issues encountered in its use, and explore the prospects for molecular docking in the field of pesticides. Additionally, we aim to offer a theoretical basis to support the development and application of new pesticides.