Molecular Docking in the Study of Ligand-Protein Recognition: An Overview

Abstract

Molecular docking is a bioinformatics-based theoretical simulation strategy. It is employed to study ligand-protein interaction profiles and predict their binding conformers and affinity through computational tools. Since the 1980s, computational tools have been used in the drug discovery process. The initial molecular modeling approaches available at the time focused on a rigid view of the ligand-protein interaction due to the limited computational capabilities. The advancement of hardware technology has made it possible to simulate the dynamic character of the ligand-protein interactions throughout time. The current chapter deals with an outline of the progression of structure-based drug discovery methodologies in the investigation of the ligand-protein interaction profiles from static to improved molecular docking strategies.

1. Introduction

Docking tools have simplified the study of interactions between drug molecules and receptor proteins, DNA, or biological molecules [1]. These interactions take place covalently. Furthermore, critical molecular mechanisms, ligand binding approaches, and factors influencing the ligand-protein interaction profile can be estimated with the help of the docking results [2, 3]. Docking suites can be used to calculate the binding energies associated with the most stable conformation of drug-receptor interactions (Figure 1) [4, 5].

Figure 1.

General modes of molecular docking simulation.

2. Types of docking

In 1982, Kuntz et al. developed the first molecular docking algorithm through the estimation of the released binding energy [6, 7]. Docking evaluations are performed to regulate the interaction profile between the ligand and target and to search for the most suitable conformation of the ligand in the complex. Empirical scoring functions are also explored, which transform binding energy into the docking score [8]. There are numerous free online tools available to generate 3D ligand and target interaction profiles, such as Biovia DSV, Pymol, Chimera, Rasmol, SwissPDB viewer, etc. Docking is broadly classified into three classes, discussed below:

2.1 Flexible docking

In flexible docking, the side chains of the protein and ligand are kept flexible. The general principle of flexible docking is based on the induced-fit hypothesis offered by Daniel Koshland in 1958 [9]. As a result, it is also known as “induced-fit docking,” in which the binding energies of various conformations of the proposed ligand are calculated at protein or receptor pockets [10, 11]. Furthermore, the target chain should be flexible enough to combine with the conformational modifications of the receptor and ligand. Various altered possible conformations of the ligand can be predicted, which makes it the most accepted and accurate technique, but it is time-consuming and costly at the same time [12].

2.2 Semi-flexible docking

In this approach, the ligand molecule is the only flexible element while the protein is rigid [13]. In addition to the six translational and rotational degrees of freedom, the conformational degrees of freedom of the ligand are also tested [14]. These approaches assume that a protein’s fixed conformation is capable of recognizing the ligands to be docked. As previously stated, this assumption is not always validated [15].

2.3 Rigid docking

In rigid docking, the main geometry of the target and ligand is retained and frozen during docking analysis [16]. The basis of this type of docking analysis is the ‘Lock and Key’ hypothesis, proposed by Emil Fischer in 1894 [17]. Thus, it is defined as lock and key docking, which also leads to several problems. The analysis of ligand-target docking is very significant for observing drug-target interaction, but a problem is associated with it when the ligand is docked at the pocket site of a receptor protein. Due to the rigid structure of both, observation of interactions becomes very challenging and the most suitable confirmation of ligand is not easily obtained [18]. Sometimes ligands do not enter the pocket site of a protein, leading to weak interactions that are not enough to show satisfactory results. Internal flexibility is necessary for good docking interaction. In various cases, the structural modifications that are essential for binding are negligible in rigid docking. Rigid docking is only enough to observe the interaction [11]. Some other benefits of rigid docking are its simplicity and a short period of run time.

3. Docking interactions

Docking is performed to establish the most suitable interaction profile for a ligand inside the target protein. It is also employed to estimate the energy evolved during the interaction between the ligand and protein [19]. Various forces influence docking interactions. The total energy released during these interactions is calculated through the empirical formula and displayed in the form of total binding energy [11, 18]. Based on the different forces, docking interactions are categorized as electro-dynamic forces (like van der Waals), electrostatic forces (charge-charge, dipole-dipole, and charge-dipole), steric forces (observed between closer molecules and influence the reactivity as well as the chemical reactivity), solvent-related forces (occurring due to interaction among the solvent and protein/ligand) and conformational modifications in the ligand) [20].

4. Types of energies

The preliminary objective of docking analysis is to obtain the best conformation of the drug during the drug-receptor interactions in support of the lowest binding free energy [21]. Molecular docking tools frequently calculate the scoring functions to evaluate the binding energies of drug-receptor interactions [11]. The resultant binding energy (ΔG bind) is calculated in the form of a combination of different energies such as H-bond, torsional free, electrostatic, unbound system’s desolation, total internal, dispersion, and repulsion, etc. The dissociation constant (K_d) is used to signify the binding energy in terms of Gibbs’ free energy (ΔG) [22]. The predication of drug-receptor binding depends on the some factors such as intermolecular interactions, desolation, and entropic effects. Upon increasing the estimation of the physio-chemical parameters, the accuracy of the scoring function is also increased [23].

An example of a scoring function is as follows:

The empirical scoring function of any docking program.

$$\text{Fitness}=\mathrm{vdW}+\mathrm{H}\text{bond}+\text{Elec}.$$

Binding Energy.

$${\mathrm{\Delta G}}_{\text{bind}}={\mathrm{\Delta G}}_{\mathrm{vdw}}+{\mathrm{\Delta G}}_{\text{hbond}}+{\mathrm{\Delta G}}_{\text{elect}}+{\mathrm{\Delta G}}_{\text{conform}}+{\mathrm{\Delta G}}_{\mathrm{tor}}+{\mathrm{\Delta G}}_{\mathrm{sol}}$$E1

5. Docking algorithms

The docking algorithms display a new dimension to evaluating the interaction profile of the ligand-receptor complex [24]. It calculates all possible conformations of the ligand under investigation during the interaction with the receptor. It also delivers the most suitable conformational pose with minimum binding energy [24, 25]. The most common algorithms apply for various docking evaluations (Flexible, Semi-flexible, and Rigid Docking) are (Figure 2):

Figure 2.

Various docking algorithms.

5.1 Flexible docking with single protein conformation

5.1.1 Side-chain flexibility docking

The side-chain flexibility docking approach introduces different conformations for various protein side-chains [26]. This is usually accomplished by utilizing rotamer library databases. Various docking approaches like GOLD use their search engine to sample some degrees of freedom. Large conformational fluctuations of the protein are ignored by these approaches due to side-chain flexibility [27].

5.2 Soft docking

In 1991, Jiang and Kim first described the soft docking strategy, which is based on the understanding of protein flexibility [28]. The VdW revulsion is also working in force field scoring functions because it reduces collisions and allows for more compact ligand-protein packing. In this method, an induced fit is recreated. As a drawback, this method can only simulate faint protein motions, which can lead to erroneous poses [24].

5.3 Flexible docking with multiple protein conformations

For the same target, multiple experimental structures may be offered [29]. Furthermore, computational approaches such as Monte Carlo or Molecular Dynamics simulations can be used to obtain an ensemble of protein conformations [30]. The goal behind multiple protein conformation docking is to consider all of the potential configurations by employing various strategies:

5.3.1 Individual conformations

The target structures are viewed as conformations that could be attached to the ligand. Therefore, several docking scores are undertaken, assessing the ligands on all of the target conformations [31]. Furthermore, to filter the structures, an initial standard to evaluate the presentation of distinct target structures in a docking investigation was also performed in individual conformations [32, 33].

5.3.2 United description of the protein

The structures are utilized to build the best-performing “chimaeras” protein instead of collapsing into an average grid [34]. Like FlexE, it selects structurally conserved areas from the ensemble’s structures to build a rigid configuration. This section is attached to the ensemble’s flexible portions in a combinatorial method, resulting in a pool of “chimaeras” that can be docked [35].

5.3.3 Average grid

The ensemble’s structures are combined to form a typical solitary grid [36].

5.4 Semi-flexible docking algorithm with simulation approaches

A well-known model of this class is molecular dynamics. This approach defines a system’s temporal evolution [37]. The molecular dynamics unit provides a more detailed explanation [38]. Energy-saving strategies are also included in this category, but these strategies are rarely utilized as standalone search engines [39]. Energy minimization is a local optimization approach for obtaining a system with certain potential energy [40].

5.5 Semi-flexible docking algorithm with stochastic methods

In this approach, the values of the degrees of freedom of a system are changed randomly rather than systematically like in stochastic algorithms [41]. The speed of these procedures is beneficial, as they might potentially locate the best answer very quickly. The main disadvantage of this approach is that it does not confirm a comprehensive investigation of the conformational space, which denotes the actual solution, which may be overlooked. Increase the number of iterations of the method to partially solve the lack of convergence. The following are the most well-known stochastic algorithms [42]:

5.5.1 Swarm optimization (SO) methods

Several swarm optimization approaches are based on the behavior of swarms [43]. The knowledge supplied by previously sampling good poses guides the sample of a ligand’s degrees of freedom. PLANTS use an Ant Colony Optimization (ACO) algorithm, which simulates the behavior of ants, and uses pheromones to find the quickest way to a food position [44]. Each degree of freedom is coupled with a pheromone in this system. Successful ants contribute to pheromone deposition, while virtual ants choose conformations based on pheromone values.

5.5.2 Evolutionary algorithms (EA)

The most prominent evolutionary algorithms are genetic algorithms (GAs), which are based on the idea of biological evolution [45]. The genes, chromosomes, mutations, and crossover concepts are all taken from biology. Genes are represented in the form of the degrees of freedom as well as ligand conformation, which is defined by a chromosome that is awarded a fitness score [46]. Within a population of chromosomes, mutations and crossovers occur, and the chromosomes with greater fitness survive and replace the ones with lower fitness. rDock, PSI-DOCk, AutoDock, and GOLD are the most well-known instances [46, 47, 48, 49, 50].

5.5.3 Tabu search methods

Tabu search strategies are used to avoid exploring zones of the conformational/positional space that have already been explored. At each cycle, random alterations are made to the ligand’s degrees of freedom. The previously sampled conformations are recorded, and a new stance is allowed only if it is distinct from any previously investigated pose. This category includes programs like PRO LEADS and PSI-DOCK [47, 51, 52, 53, 54].

5.5.4 Monte Carlo (MC) methods

The Metropolis Monte Carlo algorithm, which presents a recognized measure in the development of docking exploration, is the basis for Monte Carlo approaches [55]. Each repetition of the algorithm involves a casual adjustment of the degrees of freedom of the ligand. The Metropolis algorithm in its basic form, although it is implemented in a variety of ways in docking software, AutoDock Vina, MCDOCK, QXP, ICM, and AutoDock [30, 42, 44, 56].

5.6 Semi-flexible docking algorithm with efficient exploration techniques

In an efficient exploration, a collection of findings is associated with each degree of freedom, and all the values of each coordinate are examined in a combinatorial manner [56]. These approaches are classified into the following categories:

5.6.1 Conformational ensemble

Rigid docking approaches can easily be supplemented with a certain amount of flexibility. If an ensemble of previously produced ligand conformers is docked to the target using a conformational variation approach on the ligand complement, an example is MS-DOCK [57].

5.6.2 Fragmentation

DesJarlais et al. in 1986 described an approach to fragmentation of the ligand. The first application of ligand flexibility in docking was the hard docking of fragments into the reaction site and the subsequent connecting of the fragments [58]. Partial flexibility is achieved at the junctions among the fragments in this manner. Additional approaches, known as incremental building, initially dock one fragment and then add the rest, one by one. FlexX [59] and Hammerhead [60] are two approaches that use fragmentation [61].

5.6.3 Exhaustive search

Exhaustive exploration is an efficient method in austere intelligence, as it examines all of the ligands’ rotatable bonds systematically. To limit the search space and avoid a combinatorial explosion, several limitations and termination criteria are usually defined. The software Glide’s docking pipeline [62, 63] includes an exhaustive search stage.

6. Some common docking software

6.1 AutoDock

AutoDock is an open-source and automated docking package introduced by the Molecular Graphics Lab, Scripps Research Institute, La Jolla, CA 92037, USA. It is effectively applied to the calculation of the binding sphere of biological macromolecules like proteins and enzymes, as well as ligands (small molecules) [25]. The AutoDock docking suite offers the minimum binding energy of interaction obtained between the ligand and the receptor protein. The binding energy calculation is based on the formula offered in the form of the scoring function. Using the Lamarckian genetic algorithm (LGA), the AutoDock scoring function is established on the AMBER force field as well as through linear regression analysis [64]. It deals with reinforcing docking evaluation for ligands through almost zero to ten flexible bonds. The default settings of AutoDock are tremendously effective and are commonly applied to search for the interaction profile of a drug candidate. Furthermore, it is also extensively used for virtual screening. For each docking, the AutoDock is performed for a considerable duration to provide frequently docked conformations of the ligand concerning a receptor protein [65]. Examples: drug-receptor docking; protein-protein docking; molecule optimization; analysis oscillating from structure-based drug design; validation of the action mechanism of drug molecules; etc.

6.2 Handling tips of AutoDock

AutoDock tools offer multiple approaches for docking simulation, such as alternating from simple docking to advanced docking procedures [66]. The successful run of AutoDock requires four different files, such as ligand coordinates, target coordinates, grid parameters, and docking parameters [67, 68]. These files are prepared with the help of AutoDock Tools (ADT)/MGL Tools and their preparatory procedures are as follows:

6.3 Preparation ligand coordinate file

AutoDock accepts PDB or mol2 files as an input. In the novel compound, the first three-dimensional (3D) structure of the compound is prepared. The two-dimensional (2D) structure of the proposed compound can be prepared with the help of ChemDraw or ChemDoodle (https://web.chemdoodle.com/demos/sketcher/) and saved as a SMILES file. The SMILES file is pasted into the online CORINA Classic service (https://www.mn-am.com/online_demos/corina_demo) to prepare meals or.pdb files, but it needs further structural optimization through a suitable method such as Merck Molecular Force Field (MMFF). On the other hand, for simple preparation to optimize 3D structures, the online molsoft (https://www.molsoft.com/2dto3d.html) is recommended. It can prepare 2D as well as 3D structures in a single place. During the conversion of a 2D structure into 3D, it automatically optimizes the structure through MMFF. It has been found that the most accurate, optimized structure can be offered by DFT, but MMFF is still useful for an organic molecule. If the proposed compound has a known structure, then its crystalline 3D structure can be obtained from PubChem (https://pubchem.ncbi.nlm.nih.gov)and ChemSpider (http://www.chemspider.com/), etc. The coordinate setting of proposed compounds needs the addition of hydrogen atoms that are included in the 3D structure [69]. The proposed compound’s open 3D structure is selected as a ligand in ADT, and the ‘edit’ button is clicked to add polar hydrogens, Gasteiger charge, number of torsions, and detect root. At this moment, the ligand will be visible on the screen in which aromatic carbons appear green and another fragment looks red. Now click ‘ok’ and save it as a ligand pdbqt file.

6.4 Preparation of target coordinate file

ADT also requires preparing the coordinates of a biological macromolecule such as a protein or enzyme. The PDB file of the receptor can be downloaded from the Protein Data Bank (www.pdb.org), the Cambridge Crystallographic Database (ccdc.cam.ac.uk), etc. To generate the target coordinate file, all hydrogen atoms, need to be added. The 3D coordinates of the target can be taken from the PDB, and it requires the removal of water, ligands, cofactors, ions, etc. Click on ‘Edit’ to incorporate polar hydrogen, Kollman charge, Marge nonpolar hydrogen, and macromolecules are saved as target pdbqt.

6.5 Preparation grid parameter file

ADT needs a pdbqt file to prepare the grid parameter file (gpf). In a new window to set the grid, click on Grid > Macromolecule > Open and open the target pdbqt file by macromolecule. Similarly, click on Grid > Set map type > Open and open the ligand pdbqt file of the proposed small molecule or ligand, and then set the grid map, grid size, as well as grid center in x, y, and z-direction by clicking on “grid > Grid box”. After that, the output file can be saved as a gpf file.

6.6 Preparation docking parameter file

For the preparation of the docking parameter file (dpf), click on Docking > Macromolecule > Set rigid filament > Open in the ADT window to open the target PDBQT. Similarly, ligand pdbqt can also be opened by clicking on Docking > Ligand > Open. Then, set the algorithm by clicking on Docking > Search Parameters > Genetic algorithm and setting docking parameters. Finally, click on Docking > Output > Lararckian GA and save it as a dpf file. Then ADT is ready to run. Firstly, it runs. It required a proper time and needed the grid parameter file as well as a docking parameter file.

6.7 Analysis of docking result

ADT also offers to evaluate docking interactions and binding energies of a minimum of ten conformations along with a docking inhibition constant (K_a). By selecting Analyze > Docking > Open, you may view the findings by opening the dlg file. A popup will open, click “OK” and then further click Analyze > Conformation > Play > & > Show info.

The AutoDock scoring function can be calculated based on the following formula:

$$\begin{array}{c}\text{Free binding energy}=\text{Final intermolecular energy}+\text{Final total internal energy}\\ +\text{Torsional free energy ‐ Unbound system’s energy.}\end{array}$$

Where, total energy of van der Waal energy, hydrogen bond energy, electrostatic energy and dissolved energy equals to final intermolecular energy.

7. AutoDock Vina

AutoDock Vina was established by Oleg Trott in the Molecular Graphics Lab at the Scripps Research Institute in 2010 [70]. It is a relatively new, freely available tool for molecular docking, drug discovery, and virtual screening. It also offers high performance, multi-core proficiency, greater accuracy, and a simple handling protocol. Vina itself predicts the grid maps and clusters. Vina considerably enhances the accuracy of the interaction mode calculations as associated with AutoDock. Vina has been found to predict more accurate results as compared to other tools [71].

7.1 Handling tips of AutoDock Vina

The input and output files of Vina are pdbqt. It is essential to prepare the ligand as well as the target coordinate file in pdbqt format. Both coordinate files are prepared similarly as in AutoDock. Vina does not require a grid parameter file and a docking parameter file [72]. Additionally, it requires a text configuration file. Complete handling of AutoDock Vina is discussed below.

7.2 Preparation of configuration file

A new window of ADT is opened after the preparation of the ligand and target coordinate file. Click on Grid > Macromolecule > Open and open the target pdbqt file. Click “YES” to save the present modifications in the folder, and then press “OK” to receive them. Sometimes a warning window is also opened if there are minor indiscretions in charge. Ignore it by pressing “OK.”

Then, click Grid > Set map types > Open and open the ligand pdbqt file. The grid map, grid size, and grid center of the analysis space are then described in a new window that is opened by selecting Grid > Grid box. To begin the box built on the ligand, click Center > Center on the ligand. Here, thumbnails are available for the manual changes in the values of grid size and center, along with other options. Press file > close to save the current after adjusting the grid’s size and center. To complete the setup, select Docking > Output > Vina Configuration and click “SAVE” to provide a text configuration file with the default name of config.txt.

7.3 Run AutoDock Vina

The default setting of AutoDock Vina is not enough to accurately evaluate the interaction profile and binding energies. Vina offers a factor known as exhaustiveness to adjust the computer-aided strength utilized during a docking analysis. In Vina, the default value of exhaustiveness is 8. For greater accuracy, the default value of exhaustiveness is changed and set to about 24. It will provide more accurate docking findings. The most well-known way to run Vina is via ADT. ADT offers to click “run” to run AutoDock Vina. Open a window to start the route of the Vina executable file by pressing the browse option, and then press the launch button to operate the Vina. The second path is through the command line. Open a terminal window and modify the directory that encompasses the coordinate files as well as the configuration file. The command line is edited to adjust the values of exhaustiveness (like: /Vina--config config.txt-exhaustiveness = 24). This command accepts that the AutoDock Vina executable Vina is also found in a similar directory.

7.4 Analysis of Vina docking result

ADT also offers the ability to visualize the outcomes of docking from AutoDock Vina. Open a new ADT window and select the working directory. Analyze > Docking > Open the AutoDock Vina result and select the output file obtained from step II. Then select the default single molecule with numerous conformations followed by pressing “OK” to visualize the coordinates for all docked outcomes through arrow keys. To visualize the target coordinate file, select Analyze > Macromolecule > Open and open the target pdbqt file. Similarly, open the ligand coordinate file by clicking on File > Read molecule > Open and open the ligand pdbqt file to read the crystallographic location of the ligand. It offers the ability to evaluate the ligand as well as docked conformation. Select Analyze > Docking > Show interactions to examine the ligand-target complex’s interaction profile.

The estimated scoring function of AutoDock Vina is based on the following formula:

$$\text{∆G}\left(\text{binding}\right)=\text{∆G}\left(\mathrm{vdW}\right)+\text{∆G}\left(\mathrm{H}\text{bond}\right)+\text{∆G}\left(\text{Elec}.\right)+\text{∆G}\left(\mathrm{E}\text{desolv}.\right)$$E2

Where ∆G denotes Gibbs’ free energy, ∆G (vdW) denotes van der Waal’s free energy, and ∆G (H bond) denotes hydrogen bond free energy. ∆G (Elec.) stands for electrostatic free energy; ∆G (E dissolv.) stands for dissolved free energy. Torsional free energy is denoted by the symbol ∆G (tors).

8. AutoDock FR

AutoDock FR (ADFR: AutoDock for Flexible Receptors) was developed by Dr. Pradeep Anand Ravindranath in the Integrative Structural and Computational Biology Lab at the Scripps Research Institute in 2015. The ADFR is a newly designed docking tool built for the AutoDock scoring function. The ADFR was deliberately designed to study the interaction of small flexible ligands with the target protein [69, 73]. It offers preparation of side-chains of target proteins flexibly to simulate induced-fit without the knowledge of the side-chain conformational alterations [73]. The ADFR regulates up to 14 targets with side-chain flexibility. The proficient growth rate of docking realization is more than 50%. On the cross, docking is investigated along with up to 12 flexible receptor side-chains. The ADFR displays superior results as compared to AutoDock Vina. Vina requires uncontrolled run time for docking by increasing the number of flexible receptor side chains. On the other hand, ADFR requires linear run time [73].

8.1 Handling tips of AutoDockFR

The input format of ADFR is pdbqt format. ADFR requires the preparation of coordinate files of ligand and target. Coordinate files are prepared with the help of ADT. To perform docking through ADFR also requires the generation of affinity maps and translational points that are probable ligand binding areas. The step-by-step handling protocol of ADFR is discussed below.

8.2 Prepare affinity maps and translational points

Open a new ADFR window, select the receptor PDBQT > Open, and upload the target coordinate file in pdbqt format to run the docking analysis. Similarly, the ligand pdbqt file is uploaded by selecting Open under ligand PDBQT. Then press the box entire ligand button to surround the ligand with a docking box or grid box, followed by clicking on the center view of the docking box to center the docking position. In the docking box, along with ligand, amino acid residues can also be labeled by clicking on “show receptor residue labels.” ADFR is the only tool to select the amino acid residue up to 14 at a time with a single click. To select the amino acid residues for docking investigation, click on flexible residues and select the amino acids from the list. The selected side chains of the amino acid are presented as orange balls-sticks and the other portions remain the same. Then click the green checkmark.

For the prediction of binding pockets, click on the ‘compute pockets’ button. Auto Site recognizes multiple pockets in the docking box and selects those at which the actual ligand is found in higher volume. These binding pocket fill-points appear as a green mesh, denoted as translational points. If the binding pocket fills-points button is green, then generating maps is supported. To generate affinity maps, press the Generate maps button and save the maps as a zip file in the working folder.

8.3 Run ADFR

Open the command window, adjust the working directory, and type the following windows command to run the ADFR: “c:\Program Files\MGL Tools 2-latest\adfr.bat” random pdbqt -m generate.zip -r ligand pdbqt -job Name Result --seed −1. To visualize the docking result, a visualization tool like Biovia DSV is used to generate the interaction profile of the ligand-target complex.

9. iGEMDOCK

The iGEMDOCK tool was established by the Institute of Bioinformatics at National Chiao Tung University, Taiwan for docking, drug design, screening, and post-screening analysis. It is an automatic multipurpose graphical package [74]. For docking evaluation on the iGEMDOCK, initially prepare the coordinate files of the ligand as well as the target. Coordinate files are prepared similarly as in AutoDock by adding torsions, bond orders, hydrogen atoms, and charges. These parameters are assigned to both the ligand and the target. The input and output files of the iGEMDOCK are PDB and Mol. IGEMDOCK automatically selects the most suitable conformation of the ligand and gives the total binding energy [74]. The iGEMDOCK scores are calculated using an empirical formula or fitness score, denoted as.

Van der Waal energy + Hydrogen bond energy + Electro-statistic energy equals fitness score.

During the docking evaluation, the estimation of target binding sites and structure optimization are very significant. The hydrogen bonds found in the docked complex strongly impact the scoring function. This possibility reduces the number of suspected H bonds significantly. Additionally, internal H bonds, as well as internal electrostatic interaction, are predicted as sp2-sp2 torsions from the interaction complex. The iGEMDOCK works, since the generic evolutionary method (GA), provides three effective docking methods, viz., standard docking, stable docking, and accurate docking. Accurate docking is a very slow docking protocol and offers a maximum of 80 numbers of runs or generations, 800 population size, 8000 interactions, and 10 numbers of the solution, along with 100-threshold energy. For every single step, torsions, translations, and rotations are verified. For a better result, the hydrophobic, as well as electrostatic preferences are set to 1.00. The iGEMDOCK automatically selects the lowest energy conformation. When the iGEMDOCK calculates unfavorable electrostatic interaction, then a positive energy value is obtained. To rectify this problem, check the docked position and restart; or if the docked pose is closer to the listed ligands, define the RMSD threshold and add an energy penalty (i.e., the 100-energy penalty, 2.00 RMSD threshold, and atom ID (fast) RMSD calculations were set.) In the scoring function, the docking tool resolves and emphasizes the results of its previous search and finds their variations. Then ligand-target docking proceeds and results are obtained in the form of binding affinities (kcal/Mol) and docking run time. The minimum binding energy conformation is automatically selected as the best finding. The overall docking performance of the iGEMDOCK as compared to other docking tools is simple and better.

9.1 Handling tips of iGEMDOCK

The iGEMDOCK is a complete package of automated docking and screening. It is a combination of two main parts; the first part predicts the interaction profile among the ligand-target complex in the 3D structure, while the second part predicts the suitable pose of the ligand-target complex along with post-analysis. The docking evaluation with the iGEMDOCK begins with the preparation of ligand and target protein coordinate files. Both coordinate files are prepared like AutoDock. The iGEMDOCK input and output file formats are mol, mol2, and PDB.

9.2 Target binding site preparation

In the iGEMDOCK operator, a distinct binding site of the target protein/enzyme or complete target structure is selected. If the target’s input file contains a natural physiological ligand, it will automatically determine the target’s binding site. To begin docking, upload the target’s coordinate file (PDB) by clicking “Prepare binding site > Browse > Open” in the “Protein-ligand docking/screening” window. To select the binding site of the target, click on “By bounded ligand” and then define the binding site center by selecting the available ligand which you want to study. It also offers to set the binding site radius; by default, its value is 8.0 Å. Uncheck the “Retain reference ligand” box, and then click “OK” to save the defined parameter to the chosen binding site. This will delete the physiological ligand. Select “by a current file” to specify the binding sites of the new target protein.

9.3 Ligand preparation

The iGEMDOCK provides two methods for ligand preparation. To begin, for “single ligand,” upload the ligand coordinate file (single/many) directly by clicking “Prepare compounds > Ligands > Open” and pressing “OK” at the “docking/screening” window. The iGEMDOCK recommends preparing the ligand coordinate file in mol. It does not assign charges and hydrogen to all of the ligand’s atoms. For the “ligand database”, the ligand library is also prepared as mol. To upload the list of compounds, click “Prepare compounds,” then “import list,” “Open,” and “OK.”

9.4 Run iGEMDOCK

Set the output path before the start of docking evaluation. Set the output path by clicking on the “Set output path”. Then choose the desired file and press “OK.”

Set the GA Parameters: iGEMDOCK works based on the generic evolutionary method (GA) for docking performance. It automatically calculates the ligand conformation as well as orientation compared to the interaction site of the target. The following default GA parameters are generally recommended: population size: 200; generations: 70; and a number of solutions: 3.

Advanced Options: The iGEMDOCK offers an advanced option for the adjustment of the scoring function, saving/loading configurations, and generating docking poses. It also allows setting the internal energy of the ligand in docking prediction or the addition of certain molecular filters. It automatically produces a configuration file with the name config.dock file in the directory “/bin/”. Set up all the parameters in the configuration file and run it with the help of command mode.

Start Docking: After the setting of coordinate files of ligand and target along with output path and docking parameters, press “start docking” and observe the status of the job on the screen. After the completion of docking, a default alert is opened. To close it, click “OK,” then press “View docked poses > post-analyze” to visualize the docking poses and the complete binding energy of the docked complex. These docking poses will be saved in the “best_pose” and “fitness.txt” at the output site, respectively (Table 1).

Table 1.

The server/software of the molecular docking analysis.

10. Use of molecular docking

In the last decade, technologies like high-throughput sequencing and X-ray crystallography have been regularly updated. The crystal structures of large numbers of proteins have been defined. Consequently, the structural and functional significance of biological macromolecules (like proteins and enzymes) has been expanded and many novel drug targets also have been identified [75]. Due to the revolution of computational science in various fields of research, the utilization of virtual screening and molecular docking in DDD has been significantly stimulated. The development of a novel drug is time-consuming, costly, and needs more manpower [76]. Currently, computer-aided technology has become a key tool in DDD. Through molecular docking simulation, the analysis of the mutual interaction of drug and receptor becomes very easy along with high accuracy and boosts the drug development procedure by reducing the time [77].

Reverse molecular docking is a particularly fresh and innovative significant of molecular docking. It precedes the library of small molecules as a key structure to execute molecular docking in the spatial or 3D target database and evaluate the conceivable larger entities to conclude the three-dimensional structure and energy of identical assessment. That is to say, it identifies the most suitable target with minimum binding energy. For that reason, the development of reverse molecular docking provides a new route to discover the suitable target of a drug compound and reveal the drug action mechanism [78].

11. Conclusion

The findings of this chapter demonstrate that docking programs are highly focused on the development of new pharmaceutical compounds using molecular modeling. In this decade, new docking software designs are emphasized. These trends are focused on improving docking accuracy by using more accurate molecular energy calculations without any fitting parameters, such as quantum-chemical methods, implicit solvent models, and new global optimization algorithms that can treat ligand flexibility and protein atom mobility at the same time. Current docking applications are not reliable enough to estimate binding affinity due to the insufficient molecular structure and the inadequacies of the scoring algorithm. However, by including a huge amount of biological data into the scoring function, the present molecular docking technique can be improved. Finally, it is demonstrated that all of the conditions for improving docking accuracy may be met in practice. Furthermore, some expanded sampling strategies are no longer an exclusive methodological exercise but have become accessible to a wide range of research organizations, with real-world applications in drug discovery. Molecular docking, technological advancements, and novel MD computational approaches have all made it possible to simulate increasingly large conformational shifts. By providing a mechanical understanding of binding pathways, the ability to recreate present folding and binding processes can be used to address the long-standing argument regarding “induced-fit” and “conformational selection” binding theories.

Acknowledgments

The authors gratefully acknowledge the R&D wing of Integral University, Lucknow, for their support and guidance.

Conflict of interest

The authors declare no conflict of interest.