Artificial intelligence (AI) is increasingly being used to aid scientific research, and one area where it is making a significant impact is in the discovery of novel antibodies. Researchers are using AI algorithms to search through vast amounts of genomic data to identify potential antibody candidates that could be used to treat a range of diseases.
Researchers created an artificial intelligence tool that could help speed up the development of new high-affinity antibody drugs. University of California San Diego School of Medicine researchers have developed an artificial intelligence (AI)-based strategy for discovering high-affinity antibody drugs.
The approach was used in the study, which was published in Nature Communications, to identify a new antibody that binds a major cancer target 17-fold tighter than an existing antibody drug. According to the authors, the pipeline could hasten the development of new drugs for cancer and other diseases such as COVID-19 and rheumatoid arthritis.
An antibody must bind tightly to its target in order to be a successful drug. Researchers typically begin with a known antibody amino acid sequence and use bacterial or yeast cells to produce a series of new antibodies with variations on that sequence. The ability of these mutants to bind the target antigen is then tested. The best-performing antibodies are then subjected to another round of mutations and evaluations, and the process is repeated until a set of tightly-binding finalists emerges.
With our machine learning tools, these subsequent rounds of sequence mutation and selection can be carried out quickly and efficiently on a computer rather than in the lab.Wei Wang
Despite this lengthy and costly process, many of the resulting antibodies fail to perform well in clinical trials. UC San Diego researchers created a cutting-edge machine learning algorithm in the new study to help accelerate and streamline these efforts.
Similarly, researchers generate an initial library of about half a million possible antibody sequences and screen them for affinity to a specific protein target. Instead of repeating this process, they feed the dataset into a Bayesian neural network, which analyzes the data and uses it to predict the binding affinity of other sequences.
“With our machine learning tools, these subsequent rounds of sequence mutation and selection can be carried out quickly and efficiently on a computer rather than in the lab,” said senior author Wei Wang, Ph.D., professor of Cellular and Molecular Medicine at UC San Diego School of Medicine.
One particular advantage of their AI model is its ability to report the certainty of each prediction. “Unlike a lot of AI methods, our model can actually tell us how confident it is in each of its predictions, which helps us rank the antibodies and decide which ones to prioritize in drug development,” said Wang.
Jonathan Parkinson, Ph.D., and Ryan Hard, Ph.D., project scientists and co-first authors of the study, set out to design an antibody against programmed death ligand 1 (PD-L1), a protein highly expressed in cancer and the target of several commercially available anti-cancer drugs. Using this method, they discovered a new antibody that bound to PD-L1 17 times better than atezolizumab (Tecentriq), the wild-type antibody approved for clinical use by the US Foods and Drug Administration.
The researchers are now using this method to find promising antibodies against other antigens, such as SARS-CoV-2. They are also developing additional AI models that analyze amino acid sequences for other antibody properties important for clinical trial success, such as stability, solubility, and selectivity.
“By combining these AI tools, scientists may be able to perform an increasing share of their antibody discovery efforts on a computer rather than at the bench, potentially leading to a faster and less failure-prone discovery process,” Wang said. “There are so many applications for this pipeline, and these findings are really just the beginning.”