**4. The impact of recently developed methods over pharmacodynmics approaches**

### **4.1 Genome-wide association studies (GWAS)**

Genome-wide association studies (GWAS) act as a tool to identify the position of the gene that affects the drug susceptibility and response to the body. The genomic variants cause the PD-PGx effects with respective to the variations in patient response in the pathways of drug-target interaction. Whole genome sequencing helps to determine the PGx markers exhibiting distinct features by employing conventional genetic screening [16].

The genetic architecture of drug response can be described in deep by incorporating the PD process with GWAS. This permits to find out the genetic influences regarding the drug response and adverse drug reactions through computational tools using the mathematical equation and statistical method [17].

In this regard, candidate-gene studies offer data on the cases that may vary from adverse drug reactions to the single gene's alleles. Some victorious examples speak about and suggest this approach for the drug toxicity investigation, interpretation of drug-response. The response for a drug might be differing from patient to patient because they may not respond properly for a particular drug or it may need an error-free dosage to acquire good results. Phase III of clinical trials adopts GWAS to determine the best drug with a precise dose [18].

### **4.2 Chronopharmacodynamics**

Chronopharmacology is a field which depends on the biological time and endogenous periodicities to discern the drugs' biochemical and physiological effects on body,

#### *Trending Advancements in Technologies Pertinent to Therapeutical Pharmacodynamics DOI: http://dx.doi.org/10.5772/intechopen.107341*

drug action mechanisms, drug-concentration relationship, and impact of circadian clock to the effect of the drug. Circadian clock has established its role in regulatory metabolism, detoxification, and few other physiological processes.

Chronopharmacodynamics validates the changes of circadian in the drugs' mode of action to provide optimized pharmacodynamic response. Chronotherapeutics deals with the scientific approach of synchronizing drug delivery with the body's circadian rhythm, to maximize the therapeutic index and enhance effectiveness. Studies have revealed that a drug may have different effects depending on the time of dosage [19].

Circadian pharmacodynamics/Chronesthesy could be a variation associated with time effects that allow some modifications in treatment to improve the chances of efficacy and safety and to lower the side effects of drug. However, side effects of drugs are influenced by physio-chemical properties and PK/PD of the drug. Since, circadian variation affect the PD, it becomes essential to consider the circadian rhythm prior to drug administration so as to prevent the timely variations that occur in the drug's MoA [20].

#### **4.3 Systems biology**

Systems biology sets a trend of using data mining and statistical tools which aid to analyze the networks to locate the topology of biological systems and to build dynamic ordinary differential equation (ODE) models to represent the mechanisms of biochemical reaction. Systems pharmacology has evolved as a discipline that possesses the features of both systems biology and pharmacology that could be employed in all the phases of drug research and development. The enhanced PD (ePD) models highlight the importance of testing the drug effects in a multilevel network and the treatment of an individual patient to promote accuracy and benefits of clinical decision-making [21]. It integratively employs various domains such as systems engineering, systems biology, and PK/PD which enhances the understanding of the complex biological systems by iterating the computation and mathematical model construction, experiments and quantitatively analyzing different interactions that occur between drug and biological systems [22].

The Network-based Systems Pharmacology is an impactful way to understand the adverse effects of drug. It benefits by increasing the drug efficacy, regulating signaling pathway with multiple channels, and provide higher success rate of clinical trials, and also lowering the costs of drug discovery and development [6].

#### **4.4 Pharmacometabolomics**

Pharmacometabolomics (PMx) related studies have turned their concentration towards the areas such as biomarker identification and metabolic patterns. PMx also known as Pharmacometabonomics, is a recently developed discipline that puts in together different aspects of an individual's metabolite profile via metabolic approaches and predicts the individual's variation with respect to drug response phenotypes. It paves a way to identify endogenic metabolites and their associated pathways, individual drug PK/PD characteristic predictions and biomarkers for observing disease progression.

Metabotype (an individual metabolic profile) is one of a new strategy that consists of baseline metabotype and treatment metabotype. *Baseline metabotype* can be obtained from pre-dose sample which deals with the heterogenic data and subtypes of disease. *Treatment metabotype* are obtained from the samples which are collected while dosing in order to determine the effect/side effects of drug and changes in response

exhibited in molecular pathways. These two metabotypes help to find the variation occurred due to the effect of drug and that is exposed by inter and intra-patients. The investigation of therapeutic responses depends on these two metabotypes and the metabolic signatures that help to build models of patient's response to drug.

Another strategy is the PMx based biomarker identification which involves the steps such as data collection, process, statistical analysis, and biochemical interpretation. The biomarkers identified as an end-product of the PMx procedure may act effectively for responders and safely for non-responders in addition to some other unfavorable action and they are used to fix the range of drug dosage based on the metabolites discovered from treatment samples. PMx also identifies biomarkers involved in the downstream effects of pathophysiology and PD/PK events [11].

### **4.5 Artificial intelligence (AI)**

AI speeds up the validation and optimization of the target and the drug structure respectively. AI approaches have been employed to check the safety and efficacy of the drug molecules by developing and analyzing big data.

Recently, QSAR approaches have also been transformed as AI-dependent QSAR approaches, namely, linear discriminant analysis (LDA), support vector machines (SVM), random forest (RF), decision trees (DT) and so on. The nearest-neighbor classifier, RF, extreme learning machines, SVMS, and deep neural networks (DNNs) are able to predict the *in vivo* activity and toxicity, physicochemical properties and bioactivity of the developed drugs. With the usage of training data set (n number of compounds), predictive models have been developed and tested to predict the properties of the data (n number of compounds lesser than training data set) based on some parameters such as molecular surface area, molecular mass, total hydrogen count, refractivity, volume, logP, total polar surface area, sum of E-states indices, solubility index (log S), and rotatable bonds. By choosing anyone of the drug feature and its respective target, AI can predict the binding affinity of the drug to its target. Tools designed based on ML and DL approaches such as KronRLS, SimBoost, DeepDTA, and PADME are used to determine drug-target binding affinity. DL-based methods like DeepDTA, PADME, WideDTA and Deep Affinity produce better results when compared to the ML-based methods via the application of network-based methods [23].

Deep learning (DL) models surpassing the PK/PD methods are being followed in the clinical trials to predict the time course of the response for the respective dose of the administered therapeutic. This requires quite a lot man power to model the dynamic systems for the bulk amount of data to produce the best product (drug).

The integration of major concepts with deep learning workflow may provide an effective outcome. The models which are built with the combination of machine and human support gain a center role in the diagnostic procedures to monitor the real time treatment data. The deep learning and PK/PD approach together help to meet the necessities of automatic modeling to define the dosage by reducing the amount of time and human energy has been spent.

AI and ML system is used to detect and predict the data of individuals' symptom by means of feedback loop using digital biomarkers thereby leading to precision medicine. Certain companies involving the AI robots combined with psychological models in therapeutic areas. Some other companies design pills with an ingestible mini sized sensor coated with copper and magnesium on alternate side. Once the pill is administered to a person (patient), it will pair the two sides and start to generate

*Trending Advancements in Technologies Pertinent to Therapeutical Pharmacodynamics DOI: http://dx.doi.org/10.5772/intechopen.107341*

signal for the wearable sensor patch. The person who worn the sensor patch may get a digital record via a mobile app with the permission of his/her doctor (healthcare provider). The sensor patch will monitor the persons' activity, heart rate, sleep quality, temperature and even monitors disease conditions like diabetes, hypertension, etc.

Although AI is a boon for the pharmacological area, it has some notable limitations in handling data, such as its range, growth, variability, and unpredictability. The future of AI in clinical pharmacology has been expected to predict the concentration and effect of the drug with the help of DL models which could be implemented in digital devices that aid to monitor the treatment data automatically [24].
