Select your location

3 Trends I’m Watching in Antibody Generation

David Laine, Director Antibody Generation at Teva Pharmaceuticals

David Laine, Director Antibody Generation at Teva Pharmaceuticals, explores how artificial intelligence is transforming antibody generation, improving drug design, safety, and accelerating the development of future antibody therapies.

Trend one: AI simplifies complex biology process to make better predictions

Proteins are essential to human health, including for strength and helping our bodies grow and repair. One important type of protein is an antibody. An antibody is a specialized protein designed to recognize and bind to specific targets in the body, such as bacteria and viruses.

The antibody is like a key that binds − or locks − into its target.

At Teva, we are generating antibodies to lock onto dysregulated proteins that contribute to a disease.

Teva is excited about the potential to use artificial intelligence (AI) to analyze or predict the structure of the target protein to guide antibody generation 01 .

AI can help predict the shape of the protein (the lock) based on its genetic code then visualize how it folds and moves like seeing a 3D model of the lock. Then we can find the best spots on the protein where an antibody could attach, like identifying the keyhole.

Antibodies that bind more effectively to disease-causing proteins, potentially blocking their harmful activity and restoring healthier biological function.

Trend two: Enhancing antibody discovery with AI

The next trend is related to the first trend. It also uses AI to refine antibodies, but in a slightly different way. The analogy I use is finding a needle in a haystack.

We have billions of different antibodies that recognize a vast array of molecules, and we want to find the one that binds to the dysregulated protein.

AI can look at millions of antibodies bound to their target and learn patterns. Then, it can predict which antibodies are most likely to work against a new target, even before we test them in the lab.

Say we isolate millions of antibodies, only 0.1% bind the target of interest. We can work with AI, that acts like a super-smart assistant, to find the 1 antibody that binds the target at the right location and with the right strength to be effective.

In the future, we hope that AI will help us design brand-new antibodies saving time and cost, meaning we will be able to go from idea to treatment faster.

Teva is developing its own AI as well as testing AI tools from different suppliers to test this technology. These AI are still in early stages of development, and I’m cautiously optimistic about the potential for using AI to build new antibodies at the moment.

Trend three: AI-driven antibody improvement - precision engineering for better medicines

Once we find an antibody that works, the next step is to make it work even better: improve its efficacy and making longer-lasting medicines for patient. That’s where AI comes in again. The final trend I’m excited about focuses on using AI as a master craftsman, helping scientists fine-tune antibodies to be the best version they can be: more selective, more durable, and more effective. With just subtle, precise changes—such as adjusting a single amino acid, the basic building block of a protein—AI can have a dramatic impact.  

It can enhance how well the antibody binds to its intended target, minimize interactions with non-disease-related proteins, reduce the risk of side effects, increase its half-life in the body, and make it easier to manufacture at scale. It’s like a tiny twist of a string on an instrument: a small adjustment can transform the entire sound.

This level of optimization is less about overhauling the antibody and more like an art form—preserving its core function while refining its properties for better performance.
In this and for the other two trends, the AI is like having a super smart assistant to help us design better molecules and medicines 02 . As with any medical research, it takes time to develop and there is much trial and error. But I think the result will be worth it.

Footnote :

  1. Back to contents.

    Nature. ‘It will change everything’: DeepMind’s AI makes gigantic leap in solving protein structures’, 30 November 2020, https://www.nature.com/articles/d41586-020-03348-4  

  2. Back to contents.

    Springer Nature, Biomarker Research, ‘Revolutionizing oncology: the role of Artificial Intelligence (AI) as an antibody design, and optimization tools’, 29 March 2025, https://biomarkerres.biomedcentral.com/articles/10.1186/s40364-025-00764-4  

Find out more

Share this article: