Note for: Computer-Aided Drug Design of Bioactive Natural Products
Note for: Computer-Aided Drug Design of Bioactive Natural Products
The associated data mining tools are useful for creating databases, and molecular docking is capable of identifying potential targets by docking drugs to large libraries of proteins.
The process of hit identification can be performed using high throughput screening (HTS) and virtual screening.
Virtual screening is an effective means of searching for potential compounds by using computational approaches. One widely used computational method in this process is molecular docking.
Active compounds with good binding affinity to the target, represented
by a docking score.
Privileged structures are defined as molecular substructures that are capable of binding to a diverse array of receptors, and the modification of these substructures can provide an alternative approach to the discovery of novel receptor agonists and antagonists.
Diverse types of privileged structures have been identified from natural products, e.g., indole, quinolone, isoquinoline, purine, quinoxaline, quinazolinone, tetrahydroquinoline, tetrahydroisoquinoline, benzoxazole, benzofuran, 3,3-dimethylbenzopyran, chromone, coumarin, carbohydrate, steroid and prostanoic acid
Lipinski’s rule suggests that drug-like compounds are
1. molecules with molecular weights (MW) < 500 Da,
2. calculated octanol/water partition coefficients (clogP) < 5,
3. a number of hydrogen-bond donors < 5 and
4. a number of hydrogen-bond acceptors < 10.
Databases
broadly classified into two groups: bioactivity databases and target databases.
Bioactivity databases are valuable tools for identifying hit chemical compounds.
Bioactivity database
http://www.chemnavigator.com/
http://zinc.docking.org
https://www.ebi.ac.uk/chembl/
Pubchem
Target databases are important for identifying druggable proteins that are involved in pathogenesis.
Target database
http://tdrtargets.org
http://www.dddc.ac.cn/pdtd/
http://bidd.nus.edu.sg/group2017/index.html
Chemical space is the total possible number of descriptors from chemical compounds.
Osada and Hertweck claimed that the chemical space of natural products is populated naturally by gene clustering, where gene natural product synthesizer enzymes are altered to increase their chemical space.
In contrast, the results showed that most of the bioactive natural products exhibited drug-likeness despite having increased numbers of hydrogen bonds donors and acceptors. This result suggested that natural products have desirable properties in drug discovery and development because compounds that obey the rule-of-five are orally active and very specific in binding to their targets.
Semi-synthesis is performed by the chemical modification of natural products to improve potency, selectivity and other properties.
Fragment exchange is a complementary approach that replaces chemical fragments of natural products with synthetically derived fragments.
Diversity-oriented synthesis (DOS) is an effective tool to achieve a library of structurally diverse compounds with desirable biological properties.
Function-oriented synthesis (FOS) is an effective strategy for producing therapeutic lead compounds in a step economical fashion such that small molecules are generated with less structural complexity and with preferable properties.
It should be noted that natural products are most likely bind to multiple targets, and they are not designed for human therapeutic use.
The benefits of FOS have been noted to address these problems by reducing undesired side effects, enhancing desired biological activities and improving pharmacokinetic properties.
Biology-oriented synthesis (BIOS) is based on the structural analysis of small molecules and target proteins, where biological relevance is a prime criterion for the selection of starting scaffolds for the synthesis of biologically active compound collections.
Quantitative structure-activity/property relationships (QSAR/QSPR) describe mathematical and statistical relationships between molecular descriptors of compounds (X) and their biological activities/properties (Y).
The development of a QSAR/QSPR model is essentially comprised of five major steps: i) calculating the molecular descriptors; ii) selecting relevant and informative molecular descriptors; iii) dividing the data into training/internal and testing/external sets; iv) establishing the QSAR/QSPR model using the training set; and v) validating the QSAR/QSPR model.
a QSAR/QSPR model must perform well on both training and testing sets to be an effective and efficient model.
Currently, a few well-known QSAR/QSPR models based on machine learning techniques include multiple linear regressions (MLR), partial least square (PLS), k-nearest neighbor (k-NN), artificial neural network (ANN), support vector machine (SVM), decision tree (DT), and random forests (RF).
Machine learning tasks are typically classified into two broad categories consisting of classification and regression tasks.
The MLR, PLS, ANN, SVM, and RF methods can be utilized in both classification and regression tasks, whereas k-NN and DT are used only in the classification task.
The concept proposed by Ehrlich stated that a pharmacophore is not the same as a functional groups
of molecules; rather, it is a molecular scaffold that carries essential features responsible for the compounds’ bioactivity.
Pharmacophores can be grouped into two classes on the basis of the method that is used to obtain them;
1. structure-based pharmacophores - probing the possible interaction points between the ligand and the target
2. ligand-based pharmacophores - purely on the structure and binding data of the ligand to the target without consideration of the three-dimensional structure of the target proteins for which many
active molecules are superimposed to extract the common features that are crucial for bioactivity
The structures of ligands are superimposed to identify common features that are responsible for their biological activities without requiring 3D structures of target proteins.
In any docking scheme, two conflicting requirements should be balanced: the desire for an accurate procedure and the desire to keep the computational demands at a reasonable level.
Molecular docking, molecular dynamics (MD) simulations, which represent one of the most versatile computational techniques for studying the dynamics of biomolecules, are more computationally expensive and sophisticated.
The use of combined docking and MD methods has been broadly applied for the identification of new therapeutic agents from compounds of natural origin and the optimization of new lead candidates derived from natural compounds.
In other words, it can be stated that scaffolds of natural products are evolutionary-chosen. scaffolds from natural products serve as structural starting points to explore the biologically relevant chemical space,
and the modification to these privileged structures is required for preferable therapeutic properties
I read your blog CADD (computer-aided drug design) is a technique where we can uses software for predicting the structure value of properties like we can say known, unknown, stable, and unstable molecular these are the things which include in it. The provided information is very useful for Computer-aided drug designing. Keep continuing to post further.
ReplyDeleteYour blog is truly innovative! Let me tell you about Computer Aided Drug Designing or CADD helps in speeding things up in a remarkable manner. Accuracy, Cost, Time, Information about the disease, less manpower required are few various advantages of CADD in drug designing.
ReplyDelete