*2.3.2 Target/molecular docking-based drug repurposing*

In-silico screening of various compounds by generating drug library could find a lead molecule resulting targeted therapies. Drug compounds of specific interest from drug libraries can be selected by molecular docking or ligand-based screening which incorporates high-throughput screening (HTS) as large number of compounds are screened in this method [19]. The other methods include standard precision (SP), extra precision (XP).

During this screening there is no incorporation of information related to any biological or pharmacological as the screening is blinded. Target-based approach links the targets such as any receptor with the pathophysiology behind the disease and therefore the process of drug-discovery revamps. Withstanding to this targetbased approach cannot predict the novel/unknown mechanism with the currently available targets of the disease [1].

**69**

*Drug Repurposing and Orphan Disease Therapeutics DOI: http://dx.doi.org/10.5772/intechopen.91941*

structed for drugs repositioning [19].

drugs [19].

*2.3.5 Machine learning*

*2.3.6 Network models*

*2.3.4 Target mechanism-based drug repurposing*

network-based propagation approaches [19].

*2.3.6.1 Network-based cluster approaches*

*2.3.3 Pathway mapping/pathway-based drug repurposing*

Information such as protein-protein interaction, cell signaling and metabolic pathways can be useful for predicting the intersection between disease and drugs. The best possible example could be data available from the central database of patients can define the pathways involved in the specific disease are possibly recon-

Predicting the novel mechanism of action by this approach and resulting in drug repurposing for existing drugs. This is possible by gathering information from signaling pathway information, interaction networks of various proteins and also by data obtained from omics. Additionally, this will contribute to precision medicine. As, increasingly the individuals with their respective diseases have different pathophysiology among them. Precision medicine is upcoming solution for all these unique disease specific pathophysiological individuals. Advantages for drug repurposing by these approaches can discover not only different pathophysiological mechanisms but also providing treatment to them with respective

Machine learning (ML) techniques such as deep learning (DL), gradient boosted machine with trees (GBM), random forest (RF), support vector machine (SVM), logistic regression with elastic net regularization (EN), deep neural networks (DNN) have been usually applied for repositioning of various drugs [20, 21].

The various interaction patterns such as protein-protein, drug-disease, diseasegene, drug-target, drug-drug, disease-disease and transcriptomes and cell signaling networks are usually procured from different databases, which are interpreted computationally. The network models represent each and everything like drug, disease, gene and related products and their interaction patterns, for example, nodes (drug, disease and genes) and edges (interaction patterns of nodes). Network models are divided in to two types, these includes: network-based cluster approaches and

These kinds of approaches have been proposed in order to determine the mechanism/relationships between the drug and the disease or the drug and the target/ receptor. The biological interaction pattern in our human body has a characteristic network. The entities such as disease, drug and the protein share similar kind of interaction pattern in the network-based cluster approaches. This approach has been incorporated to develop various kinds of modules using the network topology-based cluster algorithms such as cliques/clusters/groups, subnetworks. They portray relationships or pattern of interaction between drug-drug, disease, target/receptors. The overlapping clusters cannot be detected by CLIQUE, OPTICS, DBSCAN and STING [20]. These modules are most commonly used network-based cluster approaches. Examples of repurposed drugs using network approach are atomoxetine indicated for Parkinson's disease and repurposed for attention deficit

*Drug Repurposing - Hypothesis, Molecular Aspects and Therapeutic Applications*

that are being performed routinely in drug development include cell apoptosis, infection, cell motility, cell cycle arrest and many more [17]. Cousin et al. [18] have used zebrafish model to evaluate the efficacy of compounds against tobacco dependence. In this study the authors have found that the agents like apomorphine may be useful in modifying nicotine or ethanol induced dependence behavior [18]. Over the decades in drug discovery, these approaches are getting shifted more toward molecular targets based on phenotypic methods, which rely mainly on molecular targets of drugs and diseases and related mechanisms for hypothesis

Clinical trials providing positive outcomes are quite rare as most of the drugs fail during Phase II/III trials. But many drugs which already have been marketed during post-marketing surveillance provides different outcome. Some may display different adverse events and some may treat specific kinds of disease with no labeled indication. Many drugs have been repurposed with the help of these trials. Some such examples are apomorphine was indicated for Parkinson's disease and it was repurposed for erectile dysfunction, drospirenone—oral contraceptive and repurposed for hypertension, dapoxetine—analgesia and depression and repurposed for hypertension [3]. These are very few examples of clinically repurposed drugs, there are many such drugs which have been repurposed with many new

This approach is based on already available data such as ligands and receptors. Specific models can be developed in order to locate the targets that have not been discovered/explored yet. These models are developed to discover the novel bio-markers, pathophysiology and receptors for various diseases. This kind of approach could also be beneficial in predicting different adverse reaction related to drugs, their structure activity relationships (SAR), ligands targeting the different pathways, etc., So, this could be mechanism based, pathway-based, receptor-based

In-silico screening of various compounds by generating drug library could find a lead molecule resulting targeted therapies. Drug compounds of specific interest from drug libraries can be selected by molecular docking or ligand-based screening which incorporates high-throughput screening (HTS) as large number of compounds are screened in this method [19]. The other methods include standard

During this screening there is no incorporation of information related to any biological or pharmacological as the screening is blinded. Target-based approach links the targets such as any receptor with the pathophysiology behind the disease and therefore the process of drug-discovery revamps. Withstanding to this targetbased approach cannot predict the novel/unknown mechanism with the currently

**68**

generation.

indications [1].

**2.3 Computational perspective**

*2.3.1 Knowledge-based repurposing*

repositioning of drugs [1].

precision (SP), extra precision (XP).

available targets of the disease [1].

*2.3.2 Target/molecular docking-based drug repurposing*

**2.2 Clinical perspective**

*2.2.1 Clinical analysis (human experiments)*
