The antibodies They are the body’s response to disease. They are proteins produced by the immune system of the body when it detects harmful substances, called antigens, from which the body must defend itself.
Since the 1980s, science has used Antibody therapies to treat diseases such as cancer.
However, designing these therapies is a slow process for humans: scientists must explore millions of possible combinations of amino acids to find the proteins that will fold exactly the right way and then test them all experimentally, to adjust some variables or improve some characteristics of the treatment.
Now, by using the artificial intelligence (AI) it is possible to process large volumes of data and accelerate the design of next-generation therapeutic antibodies.
To do so, Robots, computers and algorithms They are capable of processing large volumes of data and building highly effective molecules that humans can’t even imagine.
The British James Field He started his company LabGenius in 2012 when, while studying for a PhD in synthetic biology at the Imperial College London, imagined that the costs of DNA sequencing, computing, and robotics could be reduced.
“If you want to create a new therapeutic antibody, somewhere in this infinite space of potential molecules is the molecule you want to find,” he said in a recent interview with Wired.
On the company’s website it is presented as “a pioneer in the development of an intelligent robotic platform (EVA™) that is capable of designing, performing and, most importantly, learning from its own experiments.” “We believe this approach will radically improve the process of Drug Discovery, accelerating the generation of advanced antibody therapies,” he said.
Using LabGenius DNA sequencing, computing, and robotics Automates the antibody discovery process. In Bermondsey’s lab, a machine-learning algorithm designs antibodies to attack specific diseases, and then automated robotic systems build and grow them in the lab, run tests and feed the data back to the algorithm, all with limited human supervision. There are rooms for growing diseased cells, developing antibodies and sequencing their DNA.
To begin with, human scientists begin by identifying a space of Search for potential antibodies To address a particular disease: They need proteins that can differentiate between healthy and diseased cells, attach to diseased cells, and then recruit an immune cell to finish the job. But these proteins could be located anywhere in the infinite space of search for possible options.
This British start-up developed a Machine Learning Model that you can explore that space much more quickly and effectively. As Field explained, “The only information that is given to the system as a human being is, here’s an example of a healthy cell, here’s an example of a diseased cell; And then you let the system explore the different (antibody) designs that can tell them apart.”
As published by the magazine specialized in technology and American science Wired, “the model selects more than 700 initial options from a search space of 100,000 potential antibodies and then automatically designs, builds and tests them, with the goal of finding potentially fruitful areas for further investigation.”
“When you have the experimental results from that first set of 700 molecules, that information is fed back into the model and used to refine the model’s understanding of space,” Field said.
That is, the algorithm begins to build a picture of how different antibody designs change the effectiveness of treatment, and at the same time, in each subsequent round of antibody designs, carefully improves and balances the exploitation of potentially fruitful designs with the exploration of new areas.
While the Humans oversee the process, their job is mainly to move samples from one machine to the next.
According to the publication, “LabGenius’ approach produces Unexpected Solutions Humans May Not Have Thought of and finds them more quickly.” And they stressed that it only takes six weeks from the time a problem is established until the first batch is finished, all driven by machine learning models.
LabGenius began partnering with pharmaceutical companies it serves. For Field, the automated approach could also be implemented in other forms of drug discovery, turning the long “artisanal” process of Drug Discovery into something more simplified.
