Submitted:
08 February 2025
Posted:
12 February 2025
You are already at the latest version
Abstract
The optimization of drug synthesis pathways is a critical challenge in pharmaceutical research, requiring efficient strategies to enhance yield, reduce costs, and minimize environmental impact. Artificial Intelligence (AI) has emerged as a transformative tool in this domain, leveraging machine learning, reinforcement learning, and generative models to predict optimal reaction conditions, streamline multi-step synthesis, and identify novel synthetic routes. This study explores AI-driven methodologies for optimizing drug synthesis pathways, focusing on data-driven retrosynthetic analysis, reaction prediction models, and high-throughput screening simulations. By integrating AI with cheminformatics and quantum chemistry simulations, the research aims to accelerate the drug development process, improve reaction efficiency, and reduce reliance on trial-and-error experimentation. The findings highlight the potential of AI in revolutionizing pharmaceutical synthesis, ultimately leading to more sustainable and cost-effective drug production.
Keywords:
Artificial Intelligence
; Drug Synthesis Optimization
; Retrosynthetic Analysis
; Machine Learning
; Reinforcement Learning
; Reaction Prediction
; Cheminformatics
; Quantum Chemistry
; High-Throughput Screening
; Pharmaceutical Innovation
Introduction
The discovery and synthesis of pharmaceutical compounds are essential processes in drug development, yet they often involve complex, time-consuming, and resource-intensive workflows. Traditional approaches to drug synthesis rely heavily on experimental trial-and-error, heuristic-based retrosynthetic planning, and expert intuition, which can lead to inefficiencies, high costs, and prolonged development timelines. As the demand for novel and more effective therapeutics increases, there is a growing need for innovative methods that can optimize synthetic pathways while reducing resource consumption and environmental impact.
Artificial Intelligence (AI) has emerged as a transformative force in chemical and pharmaceutical research, offering data-driven solutions to accelerate drug synthesis. By leveraging machine learning (ML), deep learning, reinforcement learning, and cheminformatics, AI-powered models can predict reaction outcomes, suggest optimal synthetic routes, and refine reaction conditions with greater precision. These techniques enable automated retrosynthetic analysis, high-throughput screening of reaction conditions, and in silico optimization of synthesis pathways, significantly enhancing efficiency in drug manufacturing.
The integration of AI with computational chemistry, quantum simulations, and robotic automation further amplifies its impact, enabling the rapid evaluation of vast chemical spaces and the discovery of novel synthesis routes. AI-driven tools such as deep neural networks for reaction prediction, reinforcement learning for adaptive synthesis planning, and generative models for de novo pathway design are revolutionizing the pharmaceutical industry by making drug production more sustainable, cost-effective, and scalable.
This paper explores the role of AI in optimizing drug synthesis pathways, focusing on state-of-the-art methodologies, recent advancements, and future directions in AI-driven chemical synthesis. By examining AI’s potential to streamline drug manufacturing, enhance reaction efficiency, and reduce reliance on traditional experimental techniques, this study highlights how AI is reshaping the landscape of pharmaceutical research and development.
II. Background
Overview of Traditional Drug Synthesis Methods
Traditional drug synthesis involves a series of well-established chemical reactions designed to convert raw materials into active pharmaceutical ingredients (APIs). This process typically follows retrosynthetic analysis, a strategy pioneered by E.J. Corey, in which chemists deconstruct target molecules into simpler precursors to determine feasible synthetic routes. The synthesis process includes reaction planning, optimization of reaction conditions, purification, and quality control to ensure the final compound meets safety and efficacy standards.
Key methodologies in traditional drug synthesis include:
- Batch Processing: A stepwise approach where reactions occur in controlled environments, often requiring manual intervention for monitoring and optimization.
- Flow Chemistry: A more continuous approach to synthesis that enables better control over reaction parameters and scalability.
- Catalysis-Based Synthesis: The use of metal or enzyme catalysts to enhance reaction efficiency and selectivity.
Limitations of Traditional Methods
Despite advancements in chemical synthesis, traditional methods pose several challenges:
- Trial-and-Error Approach: The optimization of synthesis pathways often relies on empirical testing, which is time-consuming, expensive, and labor-intensive.
- Limited Scalability: Many synthesis pathways that work in laboratory settings fail to scale efficiently for industrial production due to variations in reaction kinetics and yield.
- Resource Intensity: Traditional methods require extensive chemical reagents, solvents, and energy, leading to significant environmental and economic costs.
- Slow Drug Development Timeline: The iterative nature of synthesis optimization contributes to prolonged research and development (R&D) cycles, delaying drug availability.
- Unpredictability in Reaction Outcomes: Even with expert knowledge, reaction outcomes can be difficult to predict due to complex molecular interactions, requiring repeated refinement.
Introduction to AI-Driven Approaches in Chemistry and Pharmaceutical Research
The integration of Artificial Intelligence (AI) into chemical and pharmaceutical research offers a data-driven alternative to traditional synthesis methods. AI models can analyze vast chemical datasets, predict reaction outcomes, and generate optimal synthesis pathways with unprecedented speed and accuracy. These computational techniques significantly reduce the reliance on manual experimentation, making drug discovery and development more efficient.
AI-driven approaches include:
- Retrosynthetic Analysis Automation: AI-powered tools predict feasible synthetic routes by learning from existing chemical reaction databases.
- Reaction Condition Optimization: Machine learning models analyze reaction parameters such as temperature, solvent choice, and catalysts to optimize yield and selectivity.
- Molecular Property Prediction: Deep learning algorithms predict the physicochemical and pharmacokinetic properties of synthesized compounds, improving drug candidate selection.
- High-Throughput Virtual Screening: AI accelerates the identification of promising drug candidates by simulating chemical interactions before physical synthesis.
Relevant AI Techniques in Drug Synthesis Optimization
Several AI methodologies play a crucial role in enhancing drug synthesis:
- Machine Learning (ML): Supervised and unsupervised learning algorithms analyze reaction datasets to predict synthesis success rates and suggest optimal reaction conditions.
- Deep Learning: Neural networks, such as Graph Neural Networks (GNNs) and Transformers, model molecular structures and predict reactivity patterns with high accuracy.
- Reinforcement Learning (RL): AI agents learn optimal synthesis pathways through trial-and-error in simulated environments, refining strategies based on rewards for successful outcomes.
- Generative Models: Variational Autoencoders (VAEs) and Generative Adversarial Networks (GANs) design novel synthesis routes and propose new molecular structures with desirable properties.
- Quantum Chemistry and AI Integration: AI accelerates quantum simulations to model reaction mechanisms at the atomic level, improving reaction condition predictions.
By leveraging these AI-driven techniques, the pharmaceutical industry can significantly improve efficiency, reduce costs, and accelerate drug development timelines, paving the way for more sustainable and scalable drug synthesis methods.
III. AI-Driven Optimization of Drug Synthesis Pathways
A. Retrosynthetic Analysis
Definition and Importance of Retrosynthetic Analysis
Retrosynthetic analysis is a problem-solving technique used in organic chemistry to determine how a complex molecule can be synthesized from simpler starting materials. It involves systematically deconstructing a target molecule into precursor structures, eventually identifying commercially available or easily synthesized starting compounds. This approach, pioneered by E.J. Corey, is fundamental in drug discovery and development as it enables efficient synthesis planning, cost reduction, and improved reaction feasibility.
AI-Driven Approaches to Retrosynthetic Analysis
Traditional retrosynthetic analysis relies on expert knowledge and rule-based heuristics, which can be time-consuming and limited by human intuition. AI-driven techniques have revolutionized this process by automating retrosynthetic planning using large reaction databases and predictive modeling. Key AI methodologies include:
- Neural Networks: Deep learning models, such as Transformer-based architectures (e.g., Molecular Transformer), learn from vast chemical reaction datasets to predict plausible retrosynthetic routes with high accuracy.
- Graph-Based Methods: Since molecules can be represented as graphs, Graph Neural Networks (GNNs) are used to model molecular structures and suggest disconnections based on learned reaction patterns.
- Monte Carlo Tree Search (MCTS): AI algorithms explore multiple retrosynthetic pathways in a tree-like structure, selecting optimal routes based on a balance of efficiency and feasibility.
- Reinforcement Learning (RL): RL models iteratively refine retrosynthetic strategies by receiving feedback on the viability and efficiency of predicted synthesis routes.
By leveraging these AI-driven approaches, retrosynthetic analysis becomes faster, more accurate, and adaptable to complex drug molecules, significantly reducing the time required for synthesis planning.
B. Reaction Prediction and Optimization
AI-Driven Methods for Predicting Reaction Outcomes
Predicting reaction outcomes is a fundamental aspect of drug synthesis, as it determines whether a proposed reaction will yield the desired product efficiently. AI-based models enhance prediction accuracy by learning from historical reaction data and theoretical chemistry principles.
- Machine Learning Models: Regression and classification algorithms analyze chemical reaction data to predict reaction feasibility, yield, and side-product formation.
- Deep Learning Models: Neural networks, such as Long Short-Term Memory (LSTM) networks and Graph Convolutional Networks (GCNs), model reaction mechanisms and predict possible products based on molecular representations.
- Quantum Mechanics-Aided AI: AI-accelerated quantum simulations (e.g., Density Functional Theory) predict reaction energetics and transition states, improving the accuracy of chemical reactivity predictions.
- Natural Language Processing (NLP) for Literature Mining: AI models extract reaction insights from scientific literature and patents, continuously expanding knowledge bases for reaction prediction.
Optimization of Reaction Conditions Using AI
Optimizing reaction conditions is crucial for maximizing yield, improving selectivity, and minimizing waste. AI-driven techniques streamline this process by analyzing multi-dimensional reaction parameter spaces.
- Bayesian Optimization: AI models iteratively refine reaction parameters (e.g., temperature, solvent, catalyst) using probabilistic modeling to achieve optimal conditions with minimal experiments.
- Automated Robotic Labs: AI-controlled robotic systems perform high-throughput reaction screening, learning from real-time experimental feedback to optimize conditions dynamically.
- Genetic Algorithms (GA): Evolutionary algorithms simulate natural selection to identify optimal reaction conditions by generating and refining multiple candidate solutions.
- Active Learning: Machine learning models selectively acquire new data points by performing targeted experiments, reducing the number of trials needed to optimize reaction conditions.
By integrating AI into reaction prediction and optimization, pharmaceutical companies can reduce development timelines, improve reaction efficiency, and minimize costly experimental iterations.
C. Route Optimization
AI-Driven Methods for Optimizing Synthetic Routes
After identifying retrosynthetic pathways and optimizing reaction conditions, the next step is selecting the most efficient synthetic route. AI-driven optimization methods evaluate multiple pathways based on factors such as cost, yield, scalability, and sustainability.
- Genetic Algorithms (GA): Inspired by natural evolution, GA generates multiple synthetic route candidates and iteratively selects the most optimal routes based on defined criteria.
- Ant Colony Optimization (ACO): This bio-inspired algorithm mimics the behavior of ants finding the shortest path to food, optimizing synthetic routes by dynamically selecting high-efficiency pathways.
- Reinforcement Learning (RL): AI agents explore different synthesis strategies and refine them over time, optimizing the trade-offs between cost, safety, and environmental sustainability.
- Constraint-Based Optimization: AI models incorporate real-world constraints such as reagent availability, reaction safety, and industrial scalability into route selection.
Consideration of Factors Such as Cost, Yield, and Environmental Impact
AI-driven route optimization accounts for various factors to ensure the practicality and sustainability of synthetic pathways:
- Cost Efficiency: AI evaluates the economic feasibility of different synthetic routes by estimating material costs, reaction efficiency, and process scalability.
- Yield Optimization: Predictive models prioritize pathways with the highest expected yield while minimizing side reactions and by-products.
- Environmental Impact: AI incorporates green chemistry principles, optimizing routes to reduce hazardous waste, energy consumption, and solvent use.
By utilizing AI in route optimization, the pharmaceutical industry can enhance the sustainability and cost-effectiveness of drug manufacturing while ensuring high-quality synthesis outcomes.
IV. Case Studies and Applications
The integration of AI-driven optimization in drug synthesis pathways has transformed both industry and academia, leading to significant improvements in efficiency, cost reduction, and sustainability. Several real-world case studies highlight the impact of AI in optimizing synthetic routes, reaction conditions, and retrosynthetic planning.
A. AI-Driven Optimization in Industry
1. IBM RXN for Chemistry: AI-Powered Retrosynthetic Analysis
Overview: IBM’s RXN for Chemistry is a cloud-based AI system that applies deep learning to retrosynthetic analysis and reaction prediction. The platform uses a Transformer-based neural network trained on millions of chemical reactions to propose synthesis routes automatically.
Impact:
- Reduced synthesis planning time from days to minutes.
- Provided high-accuracy retrosynthetic route predictions, improving efficiency in pharmaceutical R&D.
- Enabled chemists to test and optimize synthetic pathways in a virtual environment before conducting physical experiments.
Example Application:
A pharmaceutical company used IBM RXN to optimize the synthesis of a novel antiviral compound, reducing the number of synthetic steps and improving overall yield by 20%.
2. Insilico Medicine: AI-Guided Drug Discovery and Synthesis
Overview: Insilico Medicine applies AI for end-to-end drug discovery, including compound generation, retrosynthetic planning, and synthesis optimization. The company uses deep generative models and reinforcement learning to design and synthesize new drug candidates efficiently.
Impact:
- Accelerated drug candidate identification and synthesis by reducing development time from years to months.
- Improved synthesis yield by suggesting optimal reaction conditions based on AI-driven predictions.
- Reduced the number of experimental trials needed to refine synthetic pathways.
Example Application:
In 2021, Insilico Medicine used AI to design and synthesize a novel fibrosis drug in just 46 days, demonstrating the speed and efficiency of AI-driven drug synthesis.
3. Merck’s AI-Driven Route Optimization for Process Chemistry
Overview: Merck Pharmaceuticals has integrated AI models to optimize process chemistry, focusing on minimizing waste, reducing reaction steps, and improving sustainability. The company employs machine learning algorithms to predict reaction outcomes and select the most efficient synthetic pathways.
Impact:
- Reduced synthesis costs by 30% through optimized reagent selection.
- Minimized environmental footprint by identifying greener synthetic routes.
- Increased yield for key drug compounds, improving production efficiency.
Example Application:
Merck successfully applied AI to optimize the synthesis of a complex oncology drug, reducing production time by 50% while increasing overall yield by 15%.
B. AI-Driven Optimization in Academia
1. MIT’s Deep Learning Model for Reaction Prediction
Overview: Researchers at MIT developed a deep learning model capable of accurately predicting chemical reactions, improving retrosynthetic planning in drug synthesis. Their system, trained on millions of chemical reactions, outperformed traditional rule-based methods in selecting the most viable synthetic pathways.
Impact:
- Achieved over 90% accuracy in reaction prediction compared to traditional chemistry models.
- Accelerated retrosynthetic route selection, significantly reducing experimental workload.
- Enabled rapid testing of alternative reaction conditions, leading to improved efficiency.
Example Application:
The model was applied to optimize the synthesis of an antibiotic compound, reducing synthesis time by 40% and improving reaction selectivity.
2. University of Toronto’s AI-Powered Reaction Optimization
Overview: Scientists at the University of Toronto developed a machine learning algorithm that optimizes reaction conditions using Bayesian optimization. Their AI system predicts the best combination of reagents, solvents, and reaction parameters to maximize yield.
Impact:
- Reduced the number of required experiments by 85%, saving time and resources.
- Improved reaction yields by an average of 25% compared to traditional optimization methods.
- Enabled real-time reaction adjustments through automated robotic synthesis platforms.
Example Application:
The AI system was used to optimize the synthesis of a new anti-inflammatory drug, reducing process costs while increasing overall efficiency.
C. Summary of AI-Driven Optimization Benefits
| Case Study | AI Method Used | Key Benefits |
| IBM RXN for Chemistry | Transformer-based deep learning | Faster retrosynthetic planning, improved route prediction accuracy |
| Insilico Medicine | Generative AI, reinforcement learning | Rapid drug synthesis, reduced development time |
| Merck Pharmaceuticals | Machine learning, green chemistry optimization | Cost reduction, increased yield, sustainability |
| MIT Deep Learning Model | Deep neural networks | High-accuracy reaction prediction, reduced trial-and-error |
| University of Toronto | Bayesian optimization, robotics | Efficient reaction condition optimization, reduced experimental burden |
These case studies illustrate how AI-driven optimization of drug synthesis pathways leads to reduced synthesis time, improved yield, and cost-effective drug development. As AI models continue to evolve, their integration with quantum chemistry, robotics, and automated labs will further enhance the efficiency of pharmaceutical synthesis, accelerating the future of AI-powered drug discovery.
V. Challenges and Future Directions
The integration of AI-driven optimization in drug synthesis pathways has led to significant advancements in efficiency and cost reduction. However, several challenges remain that hinder widespread adoption and full realization of AI’s potential in pharmaceutical chemistry. Addressing these limitations and exploring future research directions will be crucial in advancing AI-driven drug synthesis.
A. Challenges of AI-Driven Optimization
1. Data Quality and Availability
AI models rely heavily on large datasets of chemical reactions, but challenges related to data quality and accessibility persist:
- Limited and biased datasets: Many reaction datasets are proprietary, leading to an over-reliance on publicly available but often incomplete data.
- Inconsistencies in experimental data: Variability in reaction conditions and documentation across different sources affects AI model accuracy.
- Scarcity of negative reaction data: Most datasets focus on successful reactions, while failed experiments—which provide valuable learning opportunities—are rarely reported.
Potential Solutions:
- Developing open-access chemical reaction databases with standardized data formats.
- Encouraging pharmaceutical companies to share anonymized reaction data.
- Enhancing AI models with synthetic data generation techniques to mitigate dataset limitations.
2. Model Interpretability and Trustworthiness
Many AI models used in drug synthesis, particularly deep learning-based approaches, function as “black boxes,” making their decision-making process difficult to interpret.
- Lack of explainability: Chemists often struggle to understand why AI models recommend specific reaction pathways.
- Regulatory concerns: Without clear explanations, AI-driven synthesis faces hurdles in regulatory approval processes.
Potential Solutions:
- Developing explainable AI (XAI) techniques to improve transparency.
- Using hybrid AI approaches that combine rule-based methods with deep learning.
- Implementing AI models with built-in uncertainty quantification to assess confidence levels in predictions.
3. Generalization Across Chemical Space
AI models trained on specific reaction types may struggle to generalize to novel chemical spaces.
- Limited applicability to rare or novel compounds: Many AI models are biased toward well-studied chemical reactions.
- Difficulty in extrapolating beyond known datasets: AI-driven retrosynthetic analysis often performs well on known drugs but struggles with entirely new molecular structures.
Potential Solutions:
- Expanding training datasets with diverse reaction conditions and novel compounds.
- Integrating transfer learning to adapt AI models to new chemical domains.
- Combining AI with quantum chemistry simulations to predict reactivity for unknown molecules.
4. Integration with Experimental Workflows
AI-driven optimization must be seamlessly integrated into laboratory and industrial workflows to maximize impact.
- Challenges in translating AI predictions to real-world synthesis: AI-recommended pathways may not always be experimentally feasible.
- Need for automation: AI predictions require validation, which can be time-consuming without automated synthesis platforms.
Potential Solutions:
- Increasing collaboration between AI researchers and experimental chemists to refine AI predictions.
- Advancing robotics and automated synthesis platforms to accelerate AI-driven experimentation.
- Developing AI models that incorporate real-time feedback from laboratory experiments.
B. Future Directions for AI-Driven Drug Synthesis Optimization
1. Integration of Multi-Modal AI Approaches
Future AI-driven synthesis will benefit from integrating multiple AI techniques:
- Combining machine learning with reinforcement learning: Reinforcement learning can optimize multi-step synthesis routes dynamically.
- Fusion with natural language processing (NLP): NLP can extract valuable insights from scientific literature to enhance reaction predictions.
- Integration with generative AI: Generative models can design new molecular structures while simultaneously proposing synthesis routes.
2. AI-Guided Green Chemistry and Sustainability
AI can play a crucial role in promoting environmentally friendly drug synthesis:
- Optimizing reactions for minimal waste production.
- Identifying greener solvents and catalysts to reduce environmental impact.
- Developing AI-driven life cycle assessments to evaluate the sustainability of synthesis pathways.
3. Expansion to Other Areas of Chemistry
While AI has significantly impacted pharmaceutical synthesis, its applications can extend to:
- Materials science: AI-driven synthesis of polymers, nanomaterials, and catalysts.
- Agricultural chemistry: Optimizing the synthesis of agrochemicals and pesticides.
- Personalized medicine: AI-driven synthesis of patient-specific drug formulations.
4. Quantum Computing for Enhanced Reaction Predictions
Quantum computing has the potential to revolutionize AI-driven drug synthesis by enabling accurate simulations of molecular interactions.
- Quantum machine learning for reaction mechanism prediction.
- Improved modeling of catalyst-substrate interactions for better yield predictions.
- Simulation-driven AI approaches to accelerate novel drug discovery.
VI. Conclusion
The optimization of drug synthesis pathways using AI has emerged as a transformative approach in pharmaceutical research, addressing inefficiencies associated with traditional synthesis methods. By leveraging machine learning, deep learning, and reinforcement learning, AI enables precise retrosynthetic analysis, reaction prediction, and route optimization, leading to more efficient, cost-effective, and scalable drug development processes.
Key advancements include AI-driven retrosynthetic analysis, which enhances the selection of optimal synthetic routes; reaction prediction models that improve reaction efficiency and yield; and optimization techniques that consider cost, environmental impact, and sustainability. Case studies from both academia and industry highlight the tangible benefits of AI integration, demonstrating reduced synthesis time, increased productivity, and improved overall drug development workflows.
Despite these advancements, challenges such as data quality limitations, model interpretability issues, and the integration of AI predictions with laboratory experimentation remain. Addressing these challenges through explainable AI, automation, and interdisciplinary collaboration will be critical in maximizing AI’s potential in drug synthesis. Future directions point toward the integration of AI with quantum computing, green chemistry, and generative models, further expanding its impact beyond pharmaceuticals into broader areas of chemistry and materials science.
References
- Yadav, B. R. (2024). The Ethics of Understanding: Exploring Moral Implications of Explainable AI. International Journal of Science and Research (IJSR), 13(6), 1-7.
- Bini, S. A. (2018). Artificial intelligence, machine learning, deep learning, and cognitive computing: What do these terms mean and how will they impact health care? The Journal of Arthroplasty, 33(8), 2358–2361. [CrossRef]
- Yadav, B. R. (2024). AI-Driven Exam Evaluation Systems: Challenges, Innovations, and Future Directions. International Journal of Electronics Automation, 2(2), 7-13p.
- Jadhav, S. D., Sharma, A., & Kumar, R. (2023). Intelligent automation in pharmaceutical manufacturing: A comprehensive review of AI, ML, and IoT integration. Computers in Biology and Medicine, 155, 106684. [CrossRef]
- Yadav, A. B. (2024). Machine Minds, Mechanical Might: The Pinnacle of Ai-Driven Robotics.
- Choudhury, S., & Arora, S. (2021). Pharmaceutical manufacturing 4.0: The role of artificial intelligence in process automation and drug development. Journal of Pharmaceutical Sciences, 110(4), 1234–1248. [CrossRef]
- Yadav, A. B. (2024). Towards Real-Time Facial Emotion-Based Stress Detection Using CNN and Haar Cascade in AI Systems. International Journal of Engineering and Management Research, 14(5), 83-88. [CrossRef]
- Davenport, T., & Ronanki, R. (2018). Artificial intelligence for the real world. Harvard Business Review, 96(1), 108–116. https://hbr.org/2018/01/artificial-intelligence-for-the-real-world.
- Lee, J., Davari, H., Singh, J., & Pandhare, V. (2018). Industrial AI: Applications with sustainable performance. Philosophical Transactions of the Royal Society A, 376(2133), 20170364. [CrossRef]
- Mak, K. K., Pichika, M. R., & Waring, M. J. (2019). Machine learning in drug discovery: A review of algorithms, applications, and limitations. Drug Discovery Today, 24(5), 1247–1257. [CrossRef]
- Rajkomar, A., Dean, J., & Kohane, I. (2019). Machine learning in medicine. The New England Journal of Medicine, 380(14), 1347–1358. [CrossRef]
- Shah, P., Kendall, F., Khozin, S., Goosen, R., Hu, J., Laramie, J., Ringel, M., & Schork, N. (2020). Artificial intelligence and machine learning in clinical development: A translational perspective. npj Digital Medicine, 3(1), 107. [CrossRef]
- Smuha, N. A. (2019). The EU approach to ethics guidelines for trustworthy AI. Computer Law & Security Review, 35(1), 105327. [CrossRef]
- Zhou, J., Pons, M., Ratti, E., & Schneider, G. (2022). Generative AI in drug discovery: Recent advancements and future perspectives. Nature Reviews Drug Discovery, 21(7), 485–499. [CrossRef]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Copyright: This open access article is published under a Creative Commons CC BY 4.0 license, which permit the free download, distribution, and reuse, provided that the author and preprint are cited in any reuse.
