Artificial intelligence in drug discovery
AI has been an important driver for many new business innovations such as web search, content referrals, product recommendations. Dr. Amit Gangwal, Associate Professor, SVKM Institute of Pharmacy, Dhule, Maharashtra, gives an overview
Artificial intelligence (AI), especially the field of deep learning (DL), has achieved tremendous success in a wide range of uses such as computer games, natural language processing, voice recognition , computer vision, driverless vehicles, automated interfaces for visual perception, decision making, translation between languages and others. None of the fields are spared by the AI. AI and robotics are no longer science fiction, they are transforming healthcare, although a bit late compared to other areas such as automotive, gaming, telecommunications, banking markets and financial, e-commerce, manufacturing, education, supply chains, marketing and others.
Currently, there is no AI challenge. For AI (actionable insights), only AI (artificial intelligence) should be used. AI is becoming more and more sophisticated to do what human experts do, but with more skill, faster and at a very competitive cost. Within the technology industry, AI has been an important driver for many new business innovations such as web search, content referrals, product recommendations, etc.
According to a report published by Accenture, “Explainable AI will not replace human workers; on the contrary, it will complement and support people, so that they can make better, faster and more accurate decisions”. AI technology can improve business productivity by up to 40%.
According to another assessment (by research firm PWC), “by 2030, global GDP could increase by 14% thanks to AI-based activities. That equates to $15.7 trillion.”
It is therefore no exaggeration to say that AI is changing our daily lives. Lately, there is a growing interest in the exploration of AI, machine learning (ML and its subtype, DL) to discover drug molecules for various diseases and to predict reactions and retrosynthetic analysis in plus other domain-specific applications. AI has a say in almost every department in healthcare industries and institutions; be it medical imaging (AI-based interpretation of various medical scans), drug discovery (including clinical trials), drug repositioning, quality assurance, marketing, sales , production, pharmaceutical analysis and others.
Current AI techniques include ML methods and DL models. The much-talked-about term, ML, was coined by Arthur Samuel of IBM in 1959. According to him, “machine learning is the field of study that gives computers the ability to learn without being explicitly programmed.”
Professor Samuel was an AI pioneer and IBM employee. Finally, DL is a subtype of ML that uses layers of artificial neurons, called neural networks, and has established improved performance over standard computer vision algorithms.
The potential of AI and robotics in healthcare is vast. Just like in our everyday life, AI and robotics are gradually becoming part of our healthcare system, just like e-commerce sites or streaming platforms analyzing our browsing and purchasing history before providing us with highly personalized data using various ML and DL models.
More and more data being generated like never before, in clinics, pathologies enables and encourages more applications of AI, ML and DL. Similarly, in the pharmaceutical industries, an enormous amount of data is generated at every stage of drug discovery and development, ranging from lead identification to post-market surveillance. High-speed internet connectivity, ultra-fast parallel processing computing unit (i.e. graphics processing units-GPUs), collaborations with cross-functional teams (like AI, tech, pharma, and medical) and Decentralized data access (as opposed to data silos) through federated networks machine learning (still widespread application is limited) are among the key enablers of the widespread and faster acceptability of AI across all domains. These applications have changed and will continue to change the way physicians and data scientists approach solving clinical problems.
A significant number of AI companies are developing and deploying their in-house developed patented tools, such as AI platforms and algorithms. These proprietary products come with such powerful features (to assist, guide, empower and complement experts in various fields of health or medicine) as drug discovery, understanding clinical trial studies, diagnosis by medical imaging. Interestingly, few companies have shown an openness to collaborating with pharmaceutical majors (like Novartis, Pfizer, GSK, Roche, AstraZeneca and others) or healthcare majors and few use their proprietary products to generate insights. for their own team, engaged in drug discovery, medical imaging and clinical trials.
Biopharmaceutical companies continue to invest heavily in AI and ML to improve their R&D and commercialization decision-making, and to deliver better outcomes for patients, physicians and payers. The pharmaceutical industries are no exception, and although belatedly or through some coming together, various organizations are adopting AI tools at different stages of drug discovery, such as lead identification, target study, clinical trials. Two main pathways are used by drug discovery scientists for AI-driven or mediated drug discovery: de novo drug design and drug reuse.
Two great organizations that deserve special mention here are Exscientia and Insilico Medicine. These are making headlines around the world leveraging AI in the drug discovery process (in target selection, ligand selection and Insilico Medicine in a clinical trial as well) through their advanced technologies. patented AI developed in-house. It is inevitable to study AI in drug discovery and not discuss these two big ones because these two are beacons not only for AI companies engaged in drug discovery but also for giants pharmaceuticals. Their frequency of announcing breakthroughs is incomparable. In two separate developments in January 2022, Insilico Medicine and Exscientia announced strategic alliances with Fosun Pharma and Sanofi respectively.
Yes, AI-based drug discovery has a promising future, but like other innovations or disruptive innovations, it also has limitations or misfires, as the technology has advanced a lot in other fields, but in the case of drug discovery, it is fairly new and must taste success through clinical trial pathways and even after that as well.