(1) Computational Medicinal Chemistry Laboratory, Department of Biochemistry, Abdul Wali Khan University Mardan, Mardan-23200, Pakistan
* Corresponding author Email:
Drug discovery and development is an intense, lengthy and an inter-disciplinary venture. Recently, a trend towards the use of in-silico chemistry and molecular modelling for computer-aided drug design has gained significant momentum. In-silico drug design skills are used in nanotechnology, molecular biology, biochemistry etc. The main benefit of the in-silico drug design is cost effective in research and development of drugs. There are wide ranges of software that are used in in-silico drug design, Grid computing, window based general PBPK/PD modelling software, PKUDDS for structure based drug design, APIS, JAVA, Perl and Python, in-silico drug design as well as software including software libraries. There are different techniques used in in-silico drug design visualization, homology, molecular dynamic, energy minimization molecular docking and QSAR etc. In-silico drug design can take part considerably in all stages of drug development from the preclinical discovery stage to late stage clinical development. Its exploitation in drug development helps in the selection of only a potent lead molecule and may thus thwart the late stage clinical failures; thereby a major diminution in cost can be achieved. This article gives an insight into all the aspects of in-silico drug design; its potential, drivers, current development and the future prospects.
In-silico methods have been of great importance in target identification and in prediction of novel drugs.
Drug discovery and development is a very complicated, time consuming process and there are many factors responsible for the failure of different drugs such as lack of effectiveness, side effects, poor pharmacokinetics, and marketable reasons. The expenditure of this process has amplified ominously during the past thirty-four years. The Pharmaceutical Manufacturer’s Association received the industry average reports which have shown that the expenditure of drug development has enlarged from $4 million in 1962 to over $350 million in 1996. The improvement time of a drug from the first synthesis to its foreword in the market, has almost multiplied between 1960 and 1980. It has kept on comparatively unaffected since 1980 with a present time period of 9-13 years.[1,4] According to pharmaceutical research and manufacturers the estimated cost of the complete drug discovery process is about US$880 million and takes up to 14 years from initial research stage to the successful marketing of an new drug in 2001.
The recent figures of DiMasi at the Tufts Centre for Study of Drug Development (CSDD) is about US$802 million spread over 12 years, which was reported in 2003, while the Boston Consulting Group estimates the cost as $880 million over 15 years. At present the cost involved in the drug discovery process ranges from $800 million to $1.8 billion, The establishment of the Computer-Aided Drug Design (CADD) Centre was to endorse joint research among the scientists of various fields like biology, biophysics, structural biology and computational scientists. The main objective of CADDC entries is to start these partnerships that lead to the implementation of research projects to discover new compounds with the potential to be transformed into new therapeutic agents. The in-silico drug design is a vast field in which the different sides of basic research and practice are combined and inspire each other, modern techniques such as QSAR/QSPR, structure-based design, combinatorial library design, cheminformatics, bioinformatics and the increasing number of biological and chemical databases are used in the field. Furthermore, large numbers of the available tools provide a much developed basis for the design of ligands and inhibitors with preferred specificity. The aim of this review was to discuss the process of In-silico drug design.
Methods used in in-silico drug design
There are many important methods in in-silico drug design research that are discussed below.
Homology modelling, is also recognized as comparative modelling of protein and it is a method that allows to generate an unknown atomic-resolution model of the "target" protein from its amino acid sequence and an experimental three dimensional (3D) structure of a related homologous protein (the "template"). Homology modelling involves the recognition of one or more identified protein structures probably to show resemblance with the structure of the query sequence, and on the making of an alignment that maps residues in the query sequence to residues in the template sequence. It has been reported that the protein structures are more conserved than protein sequences amongst homologues, but sequences have less than 20% sequence identity and can have very different structures. The proteins which are related with evolution have similar sequences and naturally occurring homologous proteins have similar protein structure. It has been revealed through the research that the evolutionarily protein three dimensional structure is more conserved than expected because of the sequence conservation to generate a structural model of the target using the sequence alignment and template structure.10 Since the protein structures are more conserved than DNA sequences, detectable levels of sequence similarity usually involve substantial structural similarity.11 Bioinformatics software tools are used to generate the 3D structure of the target on the basis of the known 3D structures of templates.12 The Modeller is a popular tool in homology modelling, and SWISS-model repository is a database of protein structures created with homology modelling.13
Molecular docking (Interaction networks)
In the field of molecular modelling docking it is a technique which envisages the favoured orientation of one molecule to a second, when bound to each other to form a stable complex. Molecular docking denotes ligand binding to its receptor or target protein. Molecular docking is used to recognize and optimize drug candidates by examining and modelling molecular interactions between ligand and target macromolecules. Molecular docking are used to generate multiple ligand conformations and orientations and the most appropriate ones are selected.[15,16] There are several molecular docking tools available that includes ArgusDock, DOCK, FRED, eHITS, AutoDock and FTDock. Molecular modelling involves scoring methods that are used to rank the affinity of ligands to bind to the active site of a receptor. In virtual high-throughput screening, compounds are docked into the active site and then scored to determine which one is more likely to bind tightly to the target macromolecule.
Virtual high-throughput screening
Virtual screening is a computational technique where large libraries of compounds are evaluated for their potential to bind specific sites on target molecules such as proteins, and well-matched compounds tested. The research in the drug discovery process involves virtual screening (VS) which is a computational method used for the rapid exploration of large libraries of chemical structures in order to identify those structures that are most likely to bind to a drug target, usually a protein receptor or enzyme. Virtual screening plays a vital role in the drug discovery process. The term "virtual screening" is relatively new as compared to the more general and old concept of database searching. Walters, et al. define virtual screening as "automatically evaluating very large libraries of compounds" using a computer program. It is clear from above definition that VS has been a numbers game at large scale and it is focusing to find out answers of questions like how can we screen down the huge chemical space of over 10 possible compounds to a practicable number that can be synthesized, purchased, and tested. Although filtering the whole chemical universe might be an interesting question, more practical VS scenarios focus on designing and optimizing targeted combinatorial libraries and enriching libraries of available compounds from in-house compound repositories or vendor offerings. It is less expensive than High-Throughput Screening, Faster than conventional screening, scanning a large number of potential drugs like molecules in very little time. HTS itself is a trial and error approach but can be better complemented by virtual screening.
Quantitative structure activity relationship (QSAR)
Quantitative structure-activity relationships (QSAR) methods are used to show a relationship of structural and/or property descriptors of compounds with their biological activities. These descriptors explaining the properties like steric, topologic, electronic, and hydrophobic of numerous molecules, have been determined through empirical methods, and only more recently by computational methods.
Hologram quantitative structure activity relationship (HQSAR)
In Hologram QSAR, a distinctive QSAR procedure, there is no need for precise 3D information about the ligands. In this method, the molecule breaks to a molecular fingerprint encoding the frequency of occurrence of various kinds of molecular fragments. Simply, the minimum and maximum length of the fragments depends on the size of the fragment to be included in the hologram fingerprint. Molecular holograms are caused by a generation of linear and branched fragments, ranging in size from 4 to 7atoms.22
Comparative molecular field analysis (CoMFA)
Comparative molecular field analysis (CoMFA) is a constructive novel technique to explain structure activity relationship. It is a well-known 3D QSAR method and work on CoMFA began in the 70’s. It delivers values of ClogP which means the solvent repellent constraints the ligands and also explains the steric and electrostatic values of the ligands.[23,24]
Comparative molecular similarity indices analysis (CoMSIA)
Comparative Molecular Similarity Indices Analysis (CoMSIA) is recognized as one of the new 3DQSAR approaches. It is generally used in the drug discovery process to locate the common characteristics, essential for the proper biological receptor binding. This method deals with the steric and electrostatic characteristics, hydrogen bond acceptors, hydrogenbond donor and hydrophobic fields.25
3D pharmacophore mapping
The 3D pharmacophore search is an imperative, vigorous and simple method to quickly recognize lead compounds alongside a preferred target. Conventionally, a pharmacophore is defined as the specific 3D arrangement of functional groups within a molecular framework that are indispensable to attach to an active site of an enzyme or bind to a macromolecule. It is essentially the first step to describe a pharmocophore in order to understand the interaction of a ligand with a receptor. Once a pharmacophore is recognized, the medicinal chemist utilizes the 3D database search tools to retrieve novel compounds that are suitable for the pharmacophore model. The modern drug design process has been used to make it one of the most successful computational tools because the search algorithms have made advancements over the years to efficiently identify and optimize lead focus combinatorial libraries and help in virtual high-throughput screening. Many improvements have been made in the computational view and application of pharmacophores in drug discovery, database searching and compound libraries. For example, in order to effectively make a partition in a library into a trial set of pharmcophore the hierarchical set of filtering calculations has emerged which can be used. This sequential filtering permits vast libraries to be proficiently handled, and also examine the compounds revealed as hits in countless detail. Furthermore, novel and protracted techniques of QSAR analysis have advanced to transform pharmacophore detail into QSAR models that, in turn can be used as virtual high-throughput screens for activity profiling of a library. Additionally, an effective application of the finger printing method was formerly utilized to make 10,549 three-point pharmacophores by enumerating numerous distance ranges and pharmacophoric features. Consequently, the fingerprint involved partial least squares as a descriptor to a QSAR model. An additional broad idea of descriptor pharmacophore was presented, which uses a variable selection QSAR as a division of molecular descriptors that afford the maximum statistically important structure-activity relationship. These approaches comprise of partial least squares and K-nearest neighbours. Thus, chemical similarity searches involving descriptor pharmacophores produces resourceful pulling out of chemical databases or virtual libraries to find out compounds with anticipated biological activities.
Microarray analysis is a new technique, known as DNA technology which plays a very significant role in the advancement of biotechnology further. These are basically properly arranged sets of known sequence DNA molecules. Mostly rectangular, which can be consisted of hundreds of thousands sets. Each single feature drives on the array at the accurately demarcated position on the substrate. The identity of the DNA molecule associated to each feature definitely does not change. Scientists use this information to know the results of their experiments. The microarray study helps scientists to perceive numerous genes in a small sample immediately and also to carry out the analysis of the expression of these genes. That safety is given to facilitate biotechnology and pharmaceutical companies to identify target molecules. Microarray analysis can assist medical companies to participate in the selection of the most suitable candidates in clinical trials of new drugs. This development has a potential as a future technology to help medical experts in the selection of the most effective drugs, or to help those with less side effects for individual patients. It has wide applications in many fields, such as transgenic animal studies, cancer tissue microarrays and other diseases, normal tissues and cells during development. This approach can be used to develop new and potent drugs.
Conformational analysis deals with deformable molecules and their minimum energy configurations through various calculation methods and interaction networks involves comparing a molecular receptor site of another molecule and calculating the most energetically satisfactory3-D conformation.31
Monte Carlo simulation
The principles of statistical mechanics are involved in Monte Carlo simulation which produces adequate different conformations of a system by computer simulation to permit the preferred thermodynamic, structural, and numerical properties to be calculated as a weighted average of these properties over these conformations. A valuable presentation has joined Monte Carlo sampling with flexible temperatures (simulated annealing) to enhance the fixing of ligands into active sites.
Molecular dynamic (MD) simulation
Molecular dynamics is an effective procedure and depends on the molecular motion simulation by solving Newton's equations of motion for each atom and increasing the speed and position of each atom by a small increase of the time duration. MD simulations characterize alternative methods to sample configuration space, based on the above mentioned rule. That is shared with temperatures using "reasonable" (a few hundreds or thousands of degrees), this means that only the local area around the sampled point, and only relatively small barriers (a few tens of kJ / mol) are overcome. Generation may be different (local), minimum may be accomplished by selecting configuration appropriate times during the simulation and thus minimize these structures. MD methods utilize the inherent dynamics of the system to search deformation modes of low energy and can be used for sampling of the conformational space of a large confined system.
During the process of selection of novel drug candidates many essential steps are taken to eliminate such compounds that have side effects and also show interaction with other drugs. In-silico drug designing softwares play an important role to design innovative proteins or drugs in biotechnology or the pharmaceutical field. The drug designing softwares and programs are used to examine molecular modelling of gene, gene expression, gene sequence analysis and 3D structure of proteins. In-silico methods have been of great importance in target identification and in prediction of novel drugs.