AI Is Game Changer For Toughest Areas Of Drug Discovery

The world of drug research and development (R&D) is undergoing a seismic shift, thanks to the power of artificial intelligence (AI) according to industry experts. Machine learning and predictive algorithms focused on large molecules such as antibodies are creating a new era where drug discovery is getting faster and more efficient. Traditionally one of the most complicated areas of research, these advances are happening now with major investment coming from pharma and venture capital.

In the past few years AI has begun to make a mark in pharma by focusing on small molecules. These tiny powerhouses are immensely important since small molecules constitute as much as 90% of global pharmaceutical sales in 2021. AI has been used to predict how they interact with their targets, optimize their effectiveness, and even forecast their safety.

Now AI is being applied to even tougher challenges, as companies have started to use the technology for larger molecules like antibodies, proteins, gene therapies, and RNA-based treatments, which represented 40% of new drug approvals in 2022, and are expected to be the future pipeline of the biopharma industry. Consider oncology, where around half of the pharma revenue is expected to come from large molecules by 2030, and will reach an overall market size $3.20 trillion according to industry analysts. Of that, more than 80% could be attributed to antibodies. So, how is AI contributing to this growth?

AI is being applied in three key areas of large molecule drug discovery. A new article published in the prestigious journal Nature details the company landscape of how AI is changing large molecule research. The expensive, time consuming challenges in large molecule science have been areas such as accurately understanding structures, predicting function and designing safe new therapies.

The Role of AI in Large Molecule Drug Discovery

First off, predicting the structure of proteins is a big deal. Think of it like knowing the blueprint of a building before you start constructing it. AI is helping make breakthroughs in highly accurate protein structure prediction, with tools like AlphaFold2 (AF2) developed by DeepMind. AF2 is an AI system that predicts the three-dimensional (3D) structures of proteins from amino acid sequences. AF2 resutlts were published in Nature in July 2021. Others like RoseTTAFold are improving the ease of use, scalability, and performance. According to the Institute for Protein Design at the University of Washington, RoseTTAFold is a software tool that uses deep learning to quickly and accurately predict protein structures based on limited information. Such a tool can eliminate years of laboratory work to determine the structure of just one protein which can now be computed in as little as ten minutes on a single gaming computer, based on results published in the prestigious journal Science.

Next, AI is becoming able to forecast how large molecules function. Imagine AI models predicting how well a key fits into a lock. These predictions are essential for therapeutic candidates, especially when it comes to binding, interaction, and pharmacokinetics. Fancy terms aside, AI is getting better at understanding how proteins bind to their targets and the movement of antibodies within the body, using techniques like deep learning and neural networks to do the heavy lifting. Deep learning models have been used to increase the power of more-accurate protein structures and small-molecule binding with the goal of designing a new entities that will elicit appropriate therapeutic responses, as published in Science.

AI is also helping generate new large-molecule therapies. With the current explosion of data, AI algorithms can design proteins, antibodies, and more. Imagine designing a protein that’s perfect for tackling cancer or developing an mRNA structure that fights off infections. Algorithms are being used to design more stable and safer mRNA vaccines for diseases such as COVID as published recently in Nature.

The Future Is AI-Powered

The biotech landscape is buzzing with over 80 companies diving into AI-driven large-molecule drug discovery. The majority were founded in the last five years, a sign of the industry’s rapid growth, fueled by technological leaps and new funding.

Investment has poured in heavily, hitting $3.9 billion in 2021 with nearly $3 billion coming from venture capital. Companies like AbCellera and Absci, dedicated to antibody and biologics discovery, had successful public offerings. Established giants in biopharma are also participating. Genentech, for instance, acquired Prescient Design, an AI-powered player in antibody discovery. Lilly has partnered with AbCellera on COVID research.

AI’s potential in large-molecule drug discovery is huge. But there are challenges to tackle. Integrating AI models into research processes, creating the right technical environment, and merging AI across the entire R&D process are crucial steps. This isn’t just about creating a new tech tool; it’s about revolutionizing an industry to save lives and make treatments more accessible. With proper use, AI can help accelerate a new era of medical research.