Computer simulation in pharmaceutical research and drug development for pharmacokinetic and therapeutic optimization

Dear Editors

In the last decade or so, the use of artificial intelligence (AI) has increased exponentially. Huge digital data is being received by pharmaceutical fraternity and medical sector. However, this data needs to be handled with extremely skilled personnel and professionals as it comes with the challenges like acquiring, scrutinizing, and applying the knowledge in hand to solve complex clinical problems. AI is an innovative technology that imitates the human intelligence such as learning, reasoning, planning and problem-solving skills and therefore can be capable to search, process and analyse the large pool of data and to respond accordingly. Therefore, in the area of drug development, AI can be helpful in improving the productivity of clinical trials and reducing animal experimentation. According to the McKinsey Global Institute, the rapid advances in AI-guided automation will be likely to completely change the work culture. The present research time is shifting towards the modern development of the latest technology and cutting-edge computer simulation.¹ Computer simulation is an important part of newest pharmacokinetics and pharmacodynamics studies for rapid development of dosage forms, prediction of the drug behaviour inside the biological system, In-vitro In-vivo correlation and expected outcome of the treatment. Computer simulation is economical in terms of time, money and manpower as compared to real laboratory experimentations and reproduces reliable results. But despite many advantages, computer simulation is totally relying on the already available literature data from pharmacokinetics and pharmacodynamics studies and has the scope of further improvement with the AI integration. AI-enabled computer simulation presents more reliable, realistic and robust simulation results. AI can assist computer simulation in target identification, lead compound screening, dosage form development, animal models and clinical trials such as data search, data categorization and predictions of the expected outcome from different data sets.^{2, 3, 4} More use of these technologies will prove to be significant as it may play an important role in selection of major studies and clinical trials to be performed, clinical trials to be performed, pharmacokinetics studies, clinical pharmacology, statistics, programmes etc.

However, over the last decade, we have seen huge development and testing of computational (In-Silico) methods in pharmacology. These methods include databases, quantitative relationships between structure and activity, pharmacophores and other models for molecular modelling, machine learning, data mining, network analysis and computer-use data analysing tools. Molecular simulation is a rigorous theoretical tool to provide reliable answers to questions related to the structure-function relationship of proteins when used efficiently. Protein dynamics collected data can be translated into useful statistical data that can be used to detect thermodynamics and kinetics that are essential for elucidating biological process modulation mechanisms, such as protein-ligand binding and the association of protein-proteins.

AI-assisted computer simulations could give molecular-level information which are not reachable with real-time animal model experiments and help to explicate the mechanism of the passive permeation process at a molecular level.^5,6

The safety and efficacy of drug and drug products has always been a major concern for everyone and hence extensive care is been taken during the developmental phase including preclinical and clinical studies. Computer simulation can be used in the prediction of drug behaviour inside the body and its interaction with the biological system before it is actually given to a human subject to ensure the safety of the drugs. During the clinical trials of the drugs, a large pool of data is generated over the period of years from patient follow-up and statistical comparison of data among different study groups to reach to a conclusion of the trials. But manual processing of the data may have data loss at multiple steps such as during data collection, data storage, data retrieval and statistical calculations which further affect the final outcome of the trial. AI-based computer simulation can fill the hole and gives better outcome in terms of drug safety as it can handle diverse aspects of drug research at a time which is almost impossible for humans.^7,8 However, use of technology is always not error free as it has its own limitations. Here, the accuracy and precision of these stimulation software's are the major challenges and there is the need of the day that they must be analysed and updated according to the need.

Four levels of these software have already been discovered i.e., Computer Simulation of the Whole Organism, Computer Simulation of Isolated Tissues and Organs, Computer Simulations of the Cell and Computer Simulations of Proteins and Genes and their utilization is being popularised day by day. The level three and four i.e., the cellular level and the molecular level have to refine and their prediction power with the help of statistics and relations must be developed for more précised results.^9,10 A lot of examples are available in literature for the application of computer simulation software in different areas of drug discovery, such as prediction of safety and toxicity, ADME, feasibility of drug molecule synthesis and drug-receptor affinity and efficacy and detection of disease at earlier stage.⁸ The use of these software could prove the golden hammer in the field of research and development of new pharmaceuticals, especially drugs. Despite many advantages, AI and computer simulation has some drawbacks such as cost of software development, AI-enabled machine learning being tedious, requiring high-end training on AI and high dependency on machines not always being beneficial.

Patients/ Guardians/ Participants consent

Patients informed consent was obtained.

Ethical clearance

Not Applicable.

Source of support

Nil.

Disclosure of competing interest

The authors have none to declare.