Computer Aided Drug Design (CADD) is a multidisciplinary field that integrates computational and experimental methods to discover, design and optimize new therapeutic compounds. This approach significantly accelerates the drug discovery process by reducing the time and cost associated with traditional methods. CADD techniques employ various computational tools and algorithmstoanalyzebiologicaldata,predictmolecularinteractions and guide the development of new drugs. By interpreting and directing studies, computational techniques speed up the drug creation mechanism.^1–3 Computer-Aided Drug Design (CADD) divided into two categories such as Structure-Based Drug Design (SBDD) and Ligand-Based Drug Design (LBDD). The SBDD technique examines the 3D structural data of macromolecular sites, often proteins or RNA, to pinpoint essential interactions and locations crucial to each biological activity. Such knowledge can be used to produce antibiotic treatment that can compete with the site’s necessary interactions, disturbing the pharmacological pathways required for microorganism(s) to survive. The aim of the LBDD approach was to recognize known antibiotic ligand for a target and create a Structure Activity Relationship (SAR) between their physiochemical qualities and antibiotic effect.^4–6 The SAR provided data that could be utilized to increase existing molecules. This chapter will offer basic CADD methods for SBDD along with LBDD not limited to synthesizing on technique and targets often investigated in our research for the development of antibacterial drugs.^7,8

Computer-aided drug design plays a crucial role in streamlining the drug discovery process, offering insights into molecular interactions, predicting drug properties and accelerating the identification of promising drug candidates. The accompanying figures visually represent key stages and techniques in the CADD workflow.^9–12

Here’s a general outline of materials and methods commonly employed in CADD for drug discovery across various diseases.

High-performance computing clusters or cloud computing platforms for running resource-intensive simulations and calculations.

Molecular Docking Software: Autodock, Glide, GOLD.^13–20
Molecular Dynamics Software: GROMACS, AMBER, NAMD.
Quantum Mechanics Software: Gaussian, NWChem.
Structure-Based Design Tools: PyMOL, UCSF Chimera.²¹
Ligand-Based Design Tools: QSAR, CoMFA, CoMSIA.
Bioinformatics Tools: BLAST, Clustal Omega.
Machine Learning Tools: Scikit-learn, TensorFlow, PyTorch.

Protein Data Bank (PDB) for retrieving 3D structures of biological macromolecules.

Chemical databases for ligand structures and properties.^22–25

3D structures of target proteins obtained from experimental methods (X-ray crystallography, NMR) or predicted structures. Structure of known ligands or potential drug candidates.

Biological data related to the target, including binding affinities, activity profiles and known ligands.²⁶

If experimental structures are not available, homology modeling can be employed to predict the 3D structure of the target protein based on the known structure of a homologous protein.

Docking algorithms are used to predict the binding mode and affinity of ligands to the target protein. This helps identify potential drug candidates.

Figure 1:
Classification of CADD.²⁷

Figure 2:
Screening Workflow in Computer Aided Drug Design (CADD).²

Simulations are performed to study the dynamic behavior and stability of protein-ligand complexes over time. This provides insights into the binding interactions.

Quantum mechanics methods can be applied for accurate calculations of electronic properties and energy levels, especially for detailed studies of reaction mechanisms.

Structural information is used to guide the design of new compounds with improved binding properties. This may involve modifying existing ligands or designing de novo compounds.²⁸

Quantitative Structure-Activity Relationship (QSAR) models are developed to correlate the chemical structure of ligands with their biological activity. This helps predict the activity of new compounds.

Machine learning algorithms are trained on diverse datasets to predict various aspects of drug discovery, including compound activity, toxicity and ADMET properties.

High-throughput virtual screening is performed to filter large chemical databases and prioritize compounds with the highest likelihood of binding to the target.

Bioinformatics tools are employed for the analysis of sequence data, identification of potential drug targets and understanding the relationships between genes, proteins and diseases.²⁹

Figure 3:
Stages of CADD Process.²⁹

Predictions made through computational methods need to be validated experimentally. This involves synthesizing and testing selected compounds in vitro and in vivo.

In the rapidly evolving field of CADD, researchers often combine multiple methods to obtain more accurate and reliable results. The iterative process of computational prediction, experimental validation and refinement is crucial for the success of drug discovery efforts across various diseases.³⁰

Computer-Aided Drug Design (CADD) has made significant strides in drug discovery, but it also has its limitations. Here are some of the key limitations associated with CADD.

CADD often relies on simplified models of biological systems. The actual behavior of molecules in a living organism is incredibly complex and many factors may be oversimplified or not fully considered in computational models.³¹

Force fields used in molecular dynamics simulations may not be accurate enough to capture all interactions accurately. The approximation of intermolecular forces can lead to inaccuracies in predicting binding affinities and other important properties.

Our understanding of biological systems is still incomplete and many aspects of protein-ligand interactions are not fully understood. This lack of knowledge can limit the accuracy of predictions made by CADD tools.

Proteins and other biomolecules are flexible and dynamic, undergoing conformational changes. CADD methods may struggle to accurately predict these dynamic changes, leading to inaccuracies in predicting binding modes.³²

Validating the predictions made by CADD methods can be challenging. Experimental validation is essential, but it can be expensive and time-consuming. Additionally, predictions made in silico may not always translate accurately to real-world biological systems.

Some CADD methods, especially those involving molecular dynamics simulations and quantum mechanics calculations, can be computationally intensive. This limits the speed at which large-scale virtual screenings can be performed.

The accuracy of CADD heavily depends on the quality and quantity of available data.³³ Insufficient or biased data can lead to inaccurate predictions, especially in machine learning-based approaches.

Achieving high specificity and selectivity for a particular target can be challenging. CADD methods may sometimes identify compounds that bind to the target but lack the desired selectivity or have off-target effects.

Biological systems, especially pathogens, can adapt and develop resistance to drugs over time. CADD methods may not fully account for the dynamic nature of these systems, leading to challenges in predicting long-term efficacy.

The use of CADD raises ethical concerns, especially when predicting potential drug candidates without extensive experimental validation. Issues related to intellectual property, safety and regulatory approval can also be challenging. Despite these limitations, CADD remains a valuable tool in drug discovery, often used in conjunction with experimental methods to accelerate the identification and optimization of potential drug candidates. Ongoing research aims to address these challenges and improve the reliability of computational methods in drug design.^34,35

CADD expedites the drug discovery process by facilitating virtual screening, lead optimization and the prediction of ligand-receptor interactions. This acceleration is crucial in addressing urgent medical needs and optimizing resources. Computational techniques enable rational drug design by providing insights into the three-dimensional structures of biological macromolecules and predicting how potential drug candidates interact with their targets. This rational approach enhances the likelihood of success in the drug development pipeline. CADD reduces the time and costs associated with experimental screening by allowing researchers to focus on a narrower set of compounds with a higher probability of success. This efficiency is particularly important in the context of large chemical spaces. The use of Quantitative Structure-Activity Relationship (QSAR) models and machine learning allows for the prediction of compound properties, including biological activity, toxicity and Absorption, Distribution, Metabolism, and Excretion (ADME) properties. This predictive capability guides decision-making in the early stages of drug development. Computational tools assist in the identification and validation of drug targets through bioinformatics analyses, helping researchers prioritize targets based on their relevance to specific diseases. Molecular docking and dynamics simulations enhance our understanding of the dynamic nature of ligand-receptor interactions. This knowledge is critical for designing molecules with optimal binding affinity and specificity. CADD techniques can be tailored to address the unique challenges of specific diseases. Whether dealing with infectious diseases, cancer, neurodegenerative disorders, or other conditions, computational methods provide a versatile platform for designing disease-specific interventions. While computational predictions are valuable, the synergy between in silico and in vitro/in vivo experimental validation is essential. This integrated approach ensures the reliability and applicability of the identified drug candidates. The field of CADD is dynamic, with ongoing innovations in algorithms, software tools and methodologies. Researchers continually refine and expand their computational approaches to overcome challenges and improve the accuracy of predictions. As CADD becomes increasingly integral to drug discovery, ethical considerations regarding the balance between in silico predictions and experimental validation, as well as issues related to data privacy and intellectual property, merit careful attention. In essence, computer-aided drug design has transformed the drug discovery landscape, offering a powerful suite of tools that complement and enhance traditional experimental approaches. The continued advancement of computational techniques holds great promise for the development of innovative and targeted therapies to address the complex challenges posed by various diseases.

In this in silico assisted drug design process offering a powerful arsenal of tools collectively known as in silico methods. From molecular docking and molecular dynamics simulations to Quantitative Structure-Activity Relationship (QSAR) models and artificial intelligence-based approaches, these methods play a pivotal role in expediting drug discovery and optimization. The utilization of molecular docking facilitates the prediction of ligand-receptor interactions, aiding in the identification of potential drug candidates. Predictions made through computational methods need to be validated experimentally. This involves synthesizing and testing selected compounds in vitro and in vivo. CADD expedites the drug discovery process by facilitating virtual screening, lead optimization and the prediction of ligand-receptor interactions. This acceleration is crucial in addressing urgent medical needs and optimizing resources. Computational techniques enable rational drug design by providing insights into the three-dimensional structures of biological macromolecules and predicting how potential drug candidates interact with their targets.

In conclusion, Computer-Aided Drug Design (CADD) techniques play a pivotal role in modern drug discovery efforts across various diseases. The integration of computational methods has significantly accelerated the identification and optimization of potential therapeutic agents.

We extend our sincere gratitude to the Dean & HOD Pharmacy, MJPRU, Bareilly whose contributions have enriched the completion of this review paper. Their support, guidance and expertise have been invaluable in navigating the complexities of the subject matter.

Special appreciation goes to the Directorate of Research, MJP Rohilkhand University, Bareilly, UP, India for providing support that facilitated the research and writing of this review. Their commitment to advancing scientific inquiry has been a crucial factor in the success of this endeavor.

We would like to acknowledge the valuable input received from colleagues and peers who generously shared their expertise and insights during work. Their contributions have broadened our perspective and enriched the context of this review. This review paper stands as a collaborative effort and we are grateful to each individual and entity that has played a part, no matter how small, in bringing this work to fruition.