Swiss Dock 2025: Docking approaches comparison between attracting cavities and autodockvina

Corresponding author: Kanala Somasekhar Reddy Copyright © 2025 Author(s) retain the copyright of this article. This article is published under the terms of the Creative Commons Attribution License 4.0. Swiss Dock 2025: Docking approaches comparison between attracting cavities and autodockvina Joshua Anthony Vomo,Papireddypalli Sushma, Bukke Geethanjali, B Dhanush and Kanala Somasekhar Reddy * Department of Pharmacology, Raghavendra Institute of Pharmaceutical Education and Research, K.R. Palli Cross, Chiyyedu (Post), Ananthapuarmu, Andhra Pradesh-515721-India. World Journal of Biology Pharmacy and Health Sciences, 2025, 22(02), 120-125 Publication history: Received on 21 March 2025; revised on 27 April 2025; accepted on 30 April 2025 Article DOI: https://doi.org/10.30574/wjbphs.2025.22.2.0452 Abstract Molecular docking is a crucial computational tool in drug discovery enabling the prediction of ligand-target interactions and accelerating the screening of potential drug candidates. This study compares twowidely used docking algorithms AutodockVina and Attracting Cavities to evaluate their efficiency, accuracy and applicability in molecular docking studies. AutodockVina is known for its speed and computational efficiency making it suitable for high-throughput screening. In contrast, Attracting Cavities provides more accurate binding predictions particularly for covalent interactions but it requires significantly more computational time. The studies suggest that AutodockVina is preferable for rapid screening whereas Attracting Cavities is advantageous for detailed validation studies. Despite these advancements, molecular docking faces challenges in flexibility modelling, scoring function accuracy and solvation effects. Future directions involve integrating machine learning and quantum chemistry to improve predictive accuracy. This study highlights the importance of selecting docking algorithms based on study-specific requirements for drug discovery. Keywords: Algorithms; Molecular Docking; Autodockvina; Attracting Cavities and Swiss dock 1. Introduction Molecular docking is a computational technique used in the prediction of the preferred orientation of ligands when binding to a target macromolecule such as protein or DNA[1].Molecular docking is very significant in the drug discovery process as it helps to reduce the time required to screen different compounds that can be potential drug candidates[2]. Molecular docking helps to understand the interactions between ligands and targets thus facilitating the identification of potential drug candidates[3]. Molecular docking is widely used to predict ligand–target interactions, delineate structure-activity relationships and assist in the design of novel drugs without prior knowledge of the chemical structure of other target modulators[4]. By 2025, technology has improved significantly especially in molecular docking approaches[4]. Thus, it is important to select appropriate docking algorithms for successful drug discovery[5]. Various algorithms and approaches offer varying levels of speed, accuracy and computational efficiency[5]. Autodockvina and attacking cavities are widely used in molecular docking[6]. Studies show that molecular docking for prediction of the ligand-target interaction with Autodockvina offers faster predictions and attracts cavities providing more accurate results[7].Comparative assessments of molecular docking algorithms are important to highlight the importance of the selection algorithm that aligns with the specific needs of the study including accuracy in posing and computational efficiency. This article aimed to compare autodockvina and the attractive cavity approaches used in docking. World Journal of Biology Pharmacy and Health Sciences, 2025, 22(02), 120-125 121 2. Molecular docking and its role in drug discovery Molecular docking is an essential technique in drug discovery and is utilized to estimate the optimal alignment of one molecule with another ensuring a stable complex formation[8]. This process is crucial for drug discovery and understanding of molecular interactions[8]. These studies suggest that molecular docking plays a major role in structure-based drug design allowing the prediction of ligand conformations and binding affinities[9]. Molecular docking parameters for accuracy and efficiency include correct pose prediction affinity scoring and the integration of computational advancements such as machine learning and protein flexibility considerations[10]. Molecular docking is an important tool in structure-based drug design, as it predicts the binding conformation of smallmolecule ligands to target binding sites[11]. This prediction aids in the rational design of drugs and elucidates the fundamental biochemical processes[12]. This technique is widely used to explore ligand conformations within macromolecular targets, estimate ligand-receptor binding free energy and identify novel compounds of therapeutic interest[13]. It also assists in visualizing the 3D structures of molecules which is essential for understanding the interactions at the molecular level[14]. 3. Docking algorithms Docking algorithms are essential tools in molecular modelling, particularly in drug discovery in whichthey estimate the optimal orientation of a ligand when it binds to a protein target[15]. The primary goal of docking is to accurately predict the binding mode and affinity of a ligand for its target protein which is crucial for identifying potential drug candidates[16]. Docking algorithms generally operate by exploring the conformational space between the ligand and protein to identify the optimal binding pose[17]. It consists of two key elements: search algorithms and scoring functions[18]. Search algorithms explore the possible orientations and conformations of the ligand within the binding site while scoring functions evaluate these positions to predict binding affinity[18]. This process often involves the use of force fields, energy calculations and solvent models to simulate the interaction between the ligand and protein[19]. Docking approaches can be broadly classified based on the flexibility of the molecules involved and the methods used such as rigid, flexible, fragment-based and consensus docking[20]. Table 1 Classification of docking approaches Classification Description Rigid Docking Both the ligand and the protein were treated as rigid bodies. This approach is computationally efficient but may not accurately capture the true binding mode due to the lack of flexibility. Flexible Docking The ligand, protein, or both are flexible. This approach is more accurate as it accounts for conformational changes upon binding but it is computationally more demanding. Fragment-Based Docking This method involves docking small fragments of the ligand separately and then assembling them at the binding site. It is particularly useful for large and complex ligands. Consensus Docking Recent advances incorporate machine learning algorithms to improve the prediction of binding poses and affinities leveraging large datasets to refine docking predictions. 4. Attracting cavities The Attracting Cavities algorithm is a covalent docking procedure designed to predict interactions between covalent drugs and their biological targets[21].It resembles the two-step process of covalent ligand binding where the ligand initially interacts with the protein cavity via non-covalent forces followed by the establishment of a covalent bond through a chemical reaction[21]. This approach allows for a more accurate prediction of covalent interactions than traditional docking methods[21]. The Attracting Cavities algorithm has been applied to various drug discovery contextsparticularly in the study of covalent inhibitors[22]. For instance, it has been used to assess the binding of covalent complexes to the SARS-CoV-2 World Journal of Biology Pharmacy and Health Sciences, 2025, 22(02), 120-125 122 main protease demonstrating its utility in antiviral drug discovery. In this challenging test set, Attracting Cavities achieved success rates of 58% for re-docking and 28% for cross-docking highlighting its potential for identifying effective inhibitors against viral targets[23]. Additionally, its integration into Swiss Dock facilitates its application in broader drug discovery efforts by providing a streamlined process for docking predictions[24]. The Attracting Cavities algorithm has notable advantages and limitations[25]. Its accuracy stands out, demonstrating superior prediction of covalent interactions compared to docking codes such as GOLD and Auto Dock achieving a 78% success rate with a root mean square deviation (rmsd) ≤ 2 Å, outperforming GOLD (66%) and Auto Dock (35%)[26]. Furthermore, Attracting Cavities integration of Attracting Cavities into platforms such as Swiss Dock enhances its accessibility and usability for researchers offering both a user-friendly interface and command-line access for efficient docking predictions[27]. However, the complexity of the algorithm may require more computational resources than simpler docking methods potentially limiting its use in high-throughput screening scenarios. Additionally, while Attracting Cavities excels in covalent docking its performance in non-covalent scenarios or with highly solvent-exposed ligands may not be as robust, as indicated by the lower success rates in cross-docking tests[27]. 5. Autodockvina AutoDockVina is a molecular docking technique used to predict how proteins and ligands interact by determining their binding modes[28]. This is an improvement over its predecessor, AutoDock 4, offering enhanced speed and accuracy[28].Vina achieved a speed-up of approximately two orders of magnitude compared to AutoDock 4 primarily because of its efficient optimization and multithreading capabilities[28]. In addition,Vina autonomously generates grid maps and clusters results, enhancing user convenience[29]. Compared with AutoDock, Vina is noted for its superior docking power ranking in the top quarter of the methods tested in the CASF-2013 benchmark[30]. AutoDockVina utilizes an empirical scoring function to assess binding affinities which has been enhanced through multiple studies to optimize its accuracy and effectiveness[31]. For instance, the Vinardo scoring function based on Vina enhances docking and virtual screening capabilities by optimizing the correlation between predicted and experimental binding affinities[32]. Vina uses an iterated local search global optimizer which significantly improves the speed and accuracy of binding mode predictions compared with AutoDock[32]. The software also supports the AutoDock4.2 scoring function and allows simultaneous docking of multiple ligands[33] AutoDockVina represents a significant advancement over AutoDock 4 offering improved speed and accuracy for molecular docking. Its strengths include speed and efficiency which are significantly faster owing to multithreading and efficient optimization[34]. Vina provides more accurate binding pose predictions and is more user-friendly with automatic grid map calculations and result clustering[35]. Flexibility is also a key strength supporting enhancements such as the Vinardo scoring function and halogen bonding parameters in VinaXB to improve docking accuracy[36]. However, Vina has weaknesses notably a lower correlation coefficient for binding affinity predictions, which can affect ligand ranking[37]. Earlier versions had limited feature support such as for macrocycles or explicit water molecules, although later updates have addressed this[37]. However, efficient large-scale virtual screening with Vina can still be resource-intensive, although GPU acceleration methods such as Vina GPU mitigate this[38]. In general, its strengths lie in its efficiency and user-friendliness while its weaknesses include challenges in binding affinity correlation and resource demands for large-scale screening[38]. 6. Limitations and future direction Molecular docking is a computational technique used to predict the interaction between two molecules, often applied in drug discovery[39]. Despite its utility there are several challenges in current docking algorithms that affect accuracy and efficiency[40]. One major challenge is conformational flexibility as both ligands and proteins undergo structural changes that need to be accurately modelled[41]. Scoring functions which predict binding affinities often lack precision due to oversimplified models and incomplete molecular structures[42]. Additionally, docking methods struggle with protein-protein interactions particularly when significant mobility or weak interactions are involved[43]. Another critical issue is properly accounting for solvation effects which significantly impact the accuracy of docking predictions[44]. To address these challenges, several improvements in docking techniques are being explored[45]. The integration of big data into scoring functions can enhance prediction accuracy by incorporating vast biological datasets[46]. Advanced algorithms, including deep learning and machine learning, are being developed to improve pose prediction and affinity scoring[47]. Furthermore, quantum chemistry methods are gaining attention for their ability to provide more precise World Journal of Biology Pharmacy and Health Sciences, 2025, 22(02), 120-125 123 energy calculations, leading to better docking accuracy[48].These improvements have the potential to refine molecular docking techniques, making them more reliable and effective[49]. Emerging trends in molecular docking focus on leveraging advanced computational approaches to enhance accuracy and applicability[50]. Machine learning algorithms are increasingly being used to improve docking efficiency and prediction reliability[50]. Fragment-based approaches are also gaining traction, enabling a more detailed exploration of binding sites[51]. Additionally, there is a growing emphasis on realistic modelling considering protein interactions in vivo within crowded cellular environments[52]. As these advancements continue to evolve they are expected to significantly enhance the role of molecular docking in drug discovery and other biomedical applications[52]. 7. Conclusion Molecular docking plays a pivotal role in drug discovery by predicting ligand-target interactions aiding in structurebased drug design and facilitating the identification of potential drug candidates. This study compared two widely used molecular docking algorithmsAutoDockVina and Attracting Cavities highlighting their respective strengths and limitations. AutoDockVina demonstrated superior speed and efficiency making it an ideal tool for high-throughput screening applications where rapid predictions are essential. On the other hand, Attracting Cavities exhibited higher accuracy in binding score predictions particularly in covalent docking scenarios making it a valuable approach for validating target-ligand interactions.The comparative analysis revealed that while AutoDockVina is more computationally efficient Attracting Cavities offers greater precision in ligand-binding predictions. The case study using the leptin receptor and orlistat as a ligand further reinforced these findings with Attracting Cavities achieving a more accurate binding score, albeit at the cost of significantly increased computational time. These findings indicate that selecting a docking algorithm should depend on the study's specific needs whether emphasizing speed for large-scale screening or accuracy for target validation. Despite the advancements in molecular docking, challenges such as protein flexibility, solvation effects and scoring function limitations persist. Future developments integrating machine learning, quantum chemistry and big data analytics hold promise for enhancing the accuracy and efficiency of docking algorithms. As computational techniques evolve molecular docking will continue to be a cornerstone of drug discovery enabling the development of novel therapeutics with improved precision and reliability. Compliance with ethical standards Acknowledgments I would like to extend my deep sense of gratitude and sincere thanks to Dr. SudheerAkkirajufor his keen interest, inspiring guidance and constant encouragement with my work during all stages. The guidance and support received from all the faculty members and my classmates who contributed to this knowledge for the completion of my article. I am grateful to all of them for their constant support and guidance either directly or indirectly towards the completion of my article. Disclosure of conflict of interest The authors declare no conflict of interest. References [1] Pagadala NS, Syed K, Tuszynski J. Software for molecular docking: a review. Biophys Rev 2017; 9: 91–102. [2] Stanzione F, Giangreco I, Cole JC. Use of molecular docking computational tools in drug discovery. In: Progress in Medicinal Chemistry. Elsevier, pp. 273–343. [3] Sivakumar KC, Haixiao J, Naman CB, et al. Prospects of multitarget drug designing strategies by linking molecular docking and molecular dynamics to explore the protein–ligand recognition process. Drug Development Research 2020; 81: 685–699. [4] Pinzi L, Rastelli G. Molecular Docking: Shifting Paradigms in Drug Discovery. IJMS 2019; 20: 4331. [5] Torres PHM, Sodero ACR, Jofily P, et al. Key Topics in Molecular Docking for Drug Design. IJMS 2019; 20: 4574. [6] Chaskar P, Zoete V, Röhrig UF. On-the-Fly QM/MM Docking with Attracting Cavities. J ChemInf Model 2017; 57: 73–84. World Journal of Biology Pharmacy and Health Sciences, 2025, 22(02), 120-125 124 [7] Patel B, Gelat B, Soni M, et al. Bioinformatics Perspective of Drug Repurposing. CBIO 2024; 19: 295–315. [8] Ferreira L, Dos Santos R, Oliva G, et al. Molecular Docking and Structure-Based Drug Design Strategies. Molecules 2015; 20: 13384–13421. [9] Śledź P, Caflisch A. Protein structure-based drug design: from docking to molecular dynamics. Current Opinion in Structural Biology 2018; 48: 93–102. [10] Vittorio S, Lunghini F, Morerio P, et al. Addressing docking pose selection with structure-based deep learning: Recent advances, challenges and opportunities. Computational and Structural Biotechnology Journal 2024; 23: 2141–2151. [11] Sotriffer C, Klebe G. Identification and mapping of small-molecule binding sites in proteins: computational tools for structure-based drug design. IlFarmaco 2002; 57: 243–251. [12] Niazi SK, Mariam Z. Computer-Aided Drug Design and Drug Discovery: A Prospective Analysis. Pharmaceuticals 2023; 17: 22. [13] Lin X, Li X, Lin X. A Review on Applications of Computational Methods in Drug Screening and Design. Molecules 2020; 25: 1375. [14] Tao A, Huang Y, Shinohara Y, et al. ezCADD: A Rapid 2D/3D Visualization-Enabled Web Modeling Environment for Democratizing Computer-Aided Drug Design. J ChemInf Model 2019; 59: 18–24. [15] Adelusi TI, Oyedele A-QK, Boyenle ID, et al. Molecular modeling in drug discovery. Informatics in Medicine Unlocked 2022; 29: 100880. [16] Sulimov VB, Kutov DC, Sulimov AV. Advances in Docking. CMC 2020; 26: 7555–7580. [17] Zhang X, Shen C, Zhang H, et al. Advancing Ligand Docking through Deep Learning: Challenges and Prospects in Virtual Screening. AccChem Res 2024; 57: 1500–1509. [18] Guedes IA, Pereira FSS, Dardenne LE. Empirical Scoring Functions for Structure-Based Virtual Screening: Applications, Critical Aspects, and Challenges. Front Pharmacol 2018; 9: 1089. [19] Shukla R, Tripathi T. Molecular Dynamics Simulation of Protein and Protein–Ligand Complexes. In: Singh DB (ed) Computer-Aided Drug Design. Singapore: Springer Singapore, pp. 133–161. [20] Dos Santos Maia M, Soares Rodrigues GC, Silva Cavalcanti AB, et al. Consensus Analyses in Molecular Docking Studies Applied to Medicinal Chemistry. MRMC 2020; 20: 1322–1340. [21] Bugnon M, Röhrig UF, Goullieux M, et al. SwissDock 2024: major enhancements for small-molecule docking with Attracting Cavities and AutoDockVina. Nucleic Acids Research 2024; 52: W324–W332. [22] Röhrig UF, Goullieux M, Bugnon M, et al. Attracting Cavities 2.0: Improving the Flexibility and Robustness for Small-Molecule Docking. J ChemInf Model 2023; 63: 3925–3940. [23] Gervasoni S, Vistoli G, Talarico C, et al. A Comprehensive Mapping of the Druggable Cavities within the SARS-CoV2 Therapeutically Relevant Proteins by Combining Pocket and Docking Searches as Implemented in Pockets 2.0. IJMS 2020; 21: 5152. [24] Nivetha T, Monisha M, Paramasivam G. In silicoscreening of inhibitors for MEK1 from alkaloids using swissdock software for human lung carcinoma. Baghdad, Iraq, p. 020199. [25] FeyzaMaden S, Sezer S, EceAcuner S. Fundamentals of Molecular Docking and Comparative Analysis of Protein– Small-Molecule Docking Approaches. In: SalihIstifli E (ed) Biomedical Engineering. IntechOpen. Epub ahead of print 25 January 2023. DOI: 10.5772/intechopen.105815. [26] Goullieux M, Zoete V, Röhrig UF. Two-Step Covalent Docking with Attracting Cavities. J ChemInf Model 2023; 63: 7847–7859. [27] Bitencourt-Ferreira G, De Azevedo WF. Docking with SwissDock. In: De Azevedo WF (ed) Docking Screens for Drug Discovery. New York, NY: Springer New York, pp. 189–202. [28] Nguyen NT, Nguyen TH, Pham TNH, et al. AutodockVina Adopts More Accurate Binding Poses but Autodock4 Forms Better Binding Affinity. J ChemInf Model 2020; 60: 204–211. [29] Temelkovski D. Conceptual Framework and Methodology for Analysing Previous Molecular Docking Results. Epub ahead of print 2019. DOI: 10.34737/QV97X. World Journal of Biology Pharmacy and Health Sciences, 2025, 22(02), 120-125 125 [30] Lu H, Wei Z, Wang C, et al. Redesigning Vina@QNLM for Ultra-Large-Scale Molecular Docking and Screening on a Sunway Supercomputer. Front Chem 2021; 9: 750325. [31] Li H, Leung K, Wong M, et al. Improving AutoDockVina Using Random Forest: The Growing Accuracy of Binding Affinity Prediction by the Effective Exploitation of Larger Data Sets. Molecular Informatics 2015; 34: 115–126. [32] Vieira TF, Sousa SF. Comparing AutoDock and Vina in Ligand/Decoy Discrimination for Virtual Screening. Applied Sciences 2019; 9: 4538. [33] Hou X, Du J, Zhang J, et al. How to Improve Docking Accuracy of AutoDock4.2: A Case Study Using Different Electrostatic Potentials. J ChemInf Model 2013; 53: 188–200. [34] Zhang B, Li H, Yu K, et al. Molecular docking-based computational platform for high-throughput virtual screening. CCF Trans HPC 2022; 4: 63–74. [35] Mohammad T, Mathur Y, Hassan MI. InstaDock: A single-click graphical user interface for molecular dockingbased virtual high-throughput screening. Briefings in Bioinformatics 2021; 22: bbaa279. [36] Gorgulla C. Recent Developments in Ultralarge and Structure-Based Virtual Screening Approaches. Annu Rev Biomed Data Sci 2023; 6: 229–258. [37] Ain QU, Aleksandrova A, Roessler FD, et al. Machine‐learning scoring functions to improve structure‐based binding affinity prediction and virtual screening. WIREs ComputMolSci 2015; 5: 405–424. [38] Lemli B, Pál S, Salem A, et al. Prioritizing Computational Cocrystal Prediction Methods for Experimental Researchers: A Review to Find Efficient, Cost-Effective, and User-Friendly Approaches. Int J MolSci 2024; 25: 12045. [39] Saikia S, Bordoloi M. Molecular Docking: Challenges, Advances and its Use in Drug Discovery Perspective. CDT 2019; 20: 501–521. [40] Huang S-Y. Comprehensive assessment of flexible-ligand docking algorithms: current effectiveness and challenges. Briefings in Bioinformatics 2018; 19: 982–994. [41] Spyrakis F, BidonChanal A, Barril X, et al. Protein Flexibility and Ligand Recognition: Challenges for Molecular Modeling. CTMC 2011; 11: 192–210. [42] Wang H. Prediction of protein–ligand binding affinity via deep learning models. Briefings in Bioinformatics 2024; 25: bbae081. [43] Yang C, Chen EA, Zhang Y. Protein–Ligand Docking in the Machine-Learning Era. Molecules 2022; 27: 4568. [44] Horvath D. Pharmacophore-Based Virtual Screening. In: Bajorath J (ed) Chemoinformatics and Computational Chemical Biology. Totowa, NJ: Humana Press, pp. 261–298. [45] Yuriev E, Agostino M, Ramsland PA. Challenges and advances in computational docking: 2009 in review. J of Molecular Recognition 2011; 24: 149–164. [46] Li Y, Wu F-X, Ngom A. A review on machine learning principles for multi-view biological data integration. Brief Bioinform 2016; bbw113. [47] Shen C, Ding J, Wang Z, et al. From machine learning to deep learning: Advances in scoring functions for protein– ligand docking. WIREs ComputMolSci 2020; 10: e1429. [48] Stroganov OV, Novikov FN, Stroylov VS, et al. Lead Finder: An Approach To Improve Accuracy of Protein−Ligand Docking, Binding Energy Estimation, and Virtual Screening. J ChemInf Model 2008; 48: 2371–2385. [49] Keith JA, Vassilev-Galindo V, Cheng B, et al. Combining Machine Learning and Computational Chemistry for Predictive Insights Into Chemical Systems. Chem Rev 2021; 121: 9816–9872. [50] Niazi SK, Mariam Z. Recent Advances in Machine-Learning-Based Chemoinformatics: A Comprehensive Review. IJMS 2023; 24: 11488. [51] Knight S, Gianni D, Hendricks A. Fragment-based screening: A new paradigm for ligand and target discovery. SLAS Discovery 2022; 27: 3–7. [52] Grassmann G, Miotto M, Desantis F, et al. Computational Approaches to Predict Protein–Protein Interactions in Crowded Cellular Environments. Chem Rev 2024; 124: 3932–3977.