Generative antibody design: When AI meets DNA

• Author:
• Author name

  Quantumrun Foresight
• September 7, 2023

Insight summary

Antibody design using generative artificial intelligence (AI) to create novel antibodies that outperform traditional ones can expedite and reduce the cost of therapeutic antibody development. This breakthrough can make personalized treatments feasible and potentially enhance medical outcomes while boosting economic productivity through reduced disease burden. However, such advancements have associated challenges, including job displacement, data privacy concerns, and ethical debates on access to personalized treatments.

Generative antibody design context

Antibodies are protective proteins created by our immune system that eliminate harmful substances by binding to them. Antibodies are frequently used in therapeutic applications due to their unique characteristics, including reduced immunogenic responses and enhanced specificity to target antigens. The initial phase in developing an antibody drug involves the identification of a principal molecule. 

This molecule is typically found by screening extensive libraries of diverse antibody variants against a specific target antigen, which can be time-consuming. The molecule's subsequent development is also a lengthy process. Therefore, it is crucial to devise speedier methods for antibody drug development.

Absci Corp, a company based in New York and Washington, made a breakthrough in 2023 when they employed a generative AI model to design novel antibodies that bind more tightly to a specific receptor, HER2, than traditional therapeutic antibodies. Interestingly, this project began with the removal of all existing antibody data, preventing the AI from merely duplicating known effective antibodies. 

The antibodies designed by Absci's AI system were distinctive, with no known counterparts, emphasizing their novelty. These AI-designed antibodies also scored high on "naturalness," suggesting ease of development and the potential to induce robust immune responses. This pioneering use of AI to design antibodies that function as well or better than our body's creations may drastically cut the time and expense of therapeutic antibody development.

Disruptive impact

Generative antibody design holds considerable promise for the future of medicine, particularly for personalized treatments. Since each person's immune response can vary significantly, creating bespoke treatments tailored to an individual's specific immune characteristics becomes possible with this technology. For example, researchers could design particular antibodies that bind to the unique cancer cells in a patient, providing a highly individualized treatment plan. 

Traditional drug development is an expensive, time-consuming process with a high failure rate. Generative AI can accelerate the process by identifying potential antibody candidates quickly, dramatically cutting costs and potentially increasing the success rate. Additionally, AI-designed antibodies can be modified and adapted faster in response to any resistance the target pathogens develop. This agility is vital in rapidly evolving diseases, as witnessed during the COVID-19 pandemic.

For governments, embracing generative AI in antibody design can impact public health. Not only can it expedite the response to health crises, but it can also make healthcare more accessible. Traditionally, many novel drugs are prohibitively expensive due to the high development costs and the need for pharmaceutical companies to recoup their investment. However, if AI can reduce these costs and accelerate the drug development timeline, the savings could be passed on to patients, making novel treatments more affordable. Moreover, responding rapidly to emerging health threats can significantly reduce their societal impact, enhancing national security.

Implications of generative antibody design

Wider implications of generative antibody design may include: 

• Individuals gaining access to personalized medical treatments resulting in improved healthcare outcomes and life expectancy.
• Health insurance providers lowering premium rates due to more cost-effective treatments and better health outcomes.
• Reduction in societal burden of disease leading to increased productivity and economic growth.
• Generation of new jobs and professions focused on the intersection of AI, biology, and medicine, contributing to a diversified job market.
• Governments being better equipped to respond to biological threats or pandemics leading to enhanced national security and societal resilience.
• Pharmaceutical companies shifting towards more sustainable and efficient research practices due to the decrease in animal testing and resource consumption.
• Universities and educational institutions adapting curricula to include AI and antibody design, fostering a new generation of interdisciplinary scientists.
• Risks associated with privacy and data security as more health and genetic data are needed for personalized antibody design.
• Political and ethical implications surrounding access to personalized treatments leading to debates about healthcare equity and fairness.

Questions to consider

• If you work in healthcare, how else might generative antibody design improve patient outcomes?
• How might governments and researchers work together to scale up the benefits of this technology?