For five decades, protein folding has been one of the toughest problems in life science. The breakthrough came in 2020 when Google DeepMind successfully tackled this long-standing challenge with AlphaFold. This achievement not only marked a turning point but also unlocked a multitude of possibilities for leveraging AI in drug development and broader healthcare applications.

Following AlphaFold’s success, other players entered the scene. Meta introduced ESMFold, while the Chinese biotech company Helixon pioneered OmegaFold. Generate Biomedicines contributed Chroma, and Baker Lab brought forth RoseTTAFold and RoseTTAFoldDiffusion, expanding the array of innovative solutions in this domain.

And one of the strongest players in this field of using AI in healthcare is Cambridge-based AstraZeneca. AIM got in touch with Siva Padmanabhan, Managing Director, AstraZeneca, India to understand how AI plays a crucial role in redefining medical science, providing a platform for the discovery, testing, and acceleration of potential medicines along with protein folding. 

However, the incorporation of AI into drug development brings about advantages and challenges. Amid the vast amount of data accessible today, the key lies in effectively analysing, interpreting, and applying this information. 

“In our research and development efforts, AI plays a vital role in decoding extensive datasets to enhance our understanding of specific diseases, identify new medicinal targets, guide molecule synthesis, and enhance predictions of clinical success,” Padmanabhan told AIM, stating that the application of AI goes beyond the laboratory and extends into various clinical approaches.

Consider the clinical trial process, for example, where AI and ML tools are employed to extract valuable insights from trial data. Proficiency in utilising trial data for safety and efficacy analysis has been demonstrated, and initiatives are in progress to maximise the potential of previously collected data. AI further contributes to event adjudication in clinical trials, streamlining processes across different stages with the overarching goal of reducing overall time investments. 

AI in Protein Folding 

The role of AI in protein folding research encompasses various aspects with potential benefits and limitations. Disease-causing proteins and antibodies’ atomic structures offer crucial insights into the functioning of therapeutics, aiding in the design of potent and safer drugs. Traditional methods such as X-ray crystallography, NMR, and cryo-EM play a pivotal role in structure determination for drug discovery. However, as per Padmanabhan, the emergence of powerful AI models has introduced an alternative approach by accurately predicting protein structures that are challenging, expensive, and time-consuming to determine experimentally.

In the space of predicting protein structures, optimising folding simulations, identifying novel proteins and their functions, and designing new protein structures, LLMs have shown promise. These models, trained on a vast repository of known protein structures, can propose protein sequences that enhance functionality and other desired properties. In the context of antibody drug discovery, these AI models can suggest sequences with tight binding to target proteins and improved developability properties. Leveraging publicly available antibody datasets, along with information about the proteins they bind to, serves as a valuable resource for building and fine-tuning models tailored to specific targets of interest.

AI-driven structure prediction facilitates the structure-guided discovery of small molecules, peptides, and antibody therapeutics. “Despite this advancement, existing AI models have limitations, primarily in predicting only the overall protein fold as their capability to predict changes caused by single amino acid mutations is restricted, and these tools provide a static snapshot of the protein while lacking insights into its dynamic nature,” added Padmanabhan.

However, according to Padmanabhan, interdisciplinary collaborations play a crucial role in advancing AI applications in protein folding research. To build effective protein folding models, a diverse team is needed, including data engineers, data scientists, structural biologists, and machine learning experts. Additionally, “the adoption of federated learning, where models are trained on data from various pharmaceutical companies and research centers without exposing the data to other entities, holds significant potential in transforming this field,” he commented.

What Next

AstraZeneca is actively engaged in exploring innovative approaches to leverage data and technology for optimising the efficiency of discovering and delivering potential new medicines.

“Throughout our R&D processes, AI is integrated to empower our scientists in pushing the boundaries of scientific exploration with the aim of delivering impactful and life-changing medicines,” concluded Padmanabhan.

Through the simulation of intricate molecular interactions, AI significantly contributes to the efficient discovery of novel drug candidates, potentially expediting the drug discovery process. This cutting-edge technology is revolutionising drug development, offering researchers the tools to combat diseases more effectively. Ultimately, this approach enhances the success rates of drug candidates, contributing to the overall improvement of patient outcomes on a global scale.

Step into the future of healthcare at Cypher 2023, where the fusion of AI and healthcare meets from October 11th to 13th at the Hilton Garden Inn Embassy, Bengaluru. In a space like healthcare where every breakthrough directly touches lives, join us to witness the metamorphosis of the healthcare industry through the lens of AI.

Read more: From Humble Beginnings to Scientific Stardom: Meet the Protein Prodigy from Bengal