AI-powered Antibody Engineering Revolutionizes Disease Treatment

Posted by

LabGenius, a company led by James Field, is revolutionizing antibody engineering using AI-powered techniques at an old biscuit factory in South London. Instead of traditional equipment, LabGenius employs robotic arms, incubators, and DNA sequencing machines to transform the antibody discovery process.

Antibodies, which combat diseases and act as the body’s frontline defense, can now be designed synthetically to treat diseases like cancer or prevent organ rejection. However, the manual process of designing these antibodies is time-consuming, requiring human protein designers to sift through numerous amino acid combinations and experimentally test them.

Field established LabGenius in 2012 after observing a decline in DNA sequencing, computational costs, and robotics during his PhD in synthetic biology at Imperial College London. Leveraging these advancements, LabGenius automates the antibody discovery process with machine learning algorithms and automated robotic systems.

At their lab in Bermondsey, the algorithms design antibodies tailored for specific diseases, which are then built and tested by the automated robotic systems. The gathered data is fed back into the algorithm, minimizing the need for human intervention.

To tackle a particular disease, human scientists begin by identifying potential antibodies capable of differentiating between healthy and diseased cells. These selected proteins must bind to the diseased cells and recruit immune cells to complete the task.

However, due to the vast number of possible options, these proteins could exist anywhere within an infinite search space. LabGenius’ model addresses this challenge by initially selecting around 700 antibodies from a search space containing 100,000 potential choices. The selected antibodies are automatically designed, built, and tested to identify promising areas for further exploration.

LabGenius employs state-of-the-art equipment to carry out almost entirely automated testing processes. High-end instruments are utilized for sample preparation and to guide samples through various stages of testing. The antibodies, grown based on their genetic sequences, are subjected to biological assays using samples of diseased tissue they are designed to combat. While the testing process is overseen by humans, their role primarily involves transferring samples between machines.

“When you have the experimental results from that first set of 700 molecules, that information gets fed back to the model and is used to refine the model’s understanding of the space,” Field explains. The algorithm gradually develops a comprehension of how different antibody designs affect treatment effectiveness. With each subsequent iteration, the model strikes a balance between exploiting potentially fruitful designs and exploring uncharted territories.

Conventional protein engineering encounters a challenge wherein a promising molecule is subjected to numerous minor modifications to further enhance its properties. However, these subtle tweaks may have adverse effects on other crucial attributes like selectivity, toxicity, potency, and more.

Consequently, a considerable amount of time is spent optimizing a suboptimal solution instead of exploring entirely different areas. In contrast, LabGenius uncovers unconventional and counterintuitive antibody designs more swiftly, often leading to unexpected solutions that humans might never have considered.

Through its automated approach, LabGenius has achieved accelerated results, taking as little as six weeks from problem setup to the completion of the initial batch. The process is guided by machine learning models. Having secured $28 million in funding from Atomico and Kindred, LabGenius is now forging partnerships with pharmaceutical companies, positioning itself as a consulting service. Furthermore, this automated approach holds the potential to expedite other forms of drug discovery, transforming the traditionally lengthy and artisanal process into a more streamlined endeavor.

Field envisions this AI-powered antibody discovery as a recipe for improved healthcare. Antibody treatments developed through this automated process offer enhanced effectiveness and reduced side effects compared to those designed by humans.

The unique molecules discovered using LabGenius’ approach exhibit distinct properties that diverge from conventional designs, ultimately leading to better patient outcomes. According to Field, this streamlined approach has the potential to unlock molecules with superior properties, benefiting the field of medicine.

%d bloggers like this: