AI accelerates the development of sequences recognized by the immune system for biomedical and technological applications

A new tool makes it possible to screen millions of small protein fragments and select those that can be recognized by the immune system

CIC biomaGUNE thus speeds up the construction of a collection of hundreds of thousands of molecules and the study of their possible applications in medicine, pharmacology, or biotechnology.

The Cooperative Research Center in Biomaterials, CIC biomaGUNE, has developed, in collaboration with the company Multiverse Computing, epiGPTope, a system that generates and classifies epitopes based on machine learning.

The presence of viruses or bacteria in the organism activates the immune system. The antibodies that are generated recognize a small part of these viruses or bacteria, the epitopes, and trigger an attack strategy. These epitopes are small fragments of protein recognized by antibodies or by receptors of immune cells. Therefore, discovering new epitope sequences that target specific antibodies is essential for the development of diagnostic tools, immunotherapies, and vaccines.

The Biomolecular Nanotechnology laboratory of CIC biomaGUNE, led by Ikerbasque professor Aitziber L. Cortajarena, is creating a library or database of hundreds of thousands of synthetic epitopes with the help of this technique based on artificial intelligence. This method for creating biologically viable sequences allows the research group to generate and select synthetic epitopes more quickly and cost-effectively, as well as to classify them according to their viral or bacterial origin, thus facilitating their application in biotechnology and biomedicine.

“From among millions of different possible combinations, we detect synthetic epitopes, very similar to natural epitopes, that can be recognized by antibodies,” explains Aitor Manteca, associate researcher of the group, “in order to see what applications these molecules could have in medical research, drug development, or biotechnology. In addition, we are able to discern whether an epitope is from a bacterium or from a virus.” In this way, they manage to build “a rational library of epitopes of hundreds of thousands of units (instead of hundreds of millions), which are stored in the laboratory. These are physical collections of molecules for experimentation,” he adds.

Point-of-care diagnostic devices

However, these epitopes do not remain in small laboratory tubes. Real applications are sought for them. Once this first screening has been carried out, the protein fragments are analyzed using microfluidic systems. This is a technology “that allows testing a single epitope against a specific antibody in a very precise, fast, and inexpensive way, obtaining many results in a short time,” explains Manteca. Thanks to microfluidics, experiments are carried out in very small droplets that function as individual reactors, with very small quantities of molecules. “It is possible to study millions of different combinations at the same time in a short period of time,” adds the CIC biomaGUNE researcher.

Thus, it is possible to know in advance “which sequences will generate immune responses and to advance, for example, in the generation of diagnostic techniques and new Point-of-Care devices capable of measuring the presence of a bacterium or a virus in an organism, in blood, in water, etc.,” explains Dr. Aitor Manteca.

These developments are especially relevant for their transfer to the industrial environment, as in the case of the company Taldeki Biosolutions, which exploits a detection technology licensed by CIC biomaGUNE. In this context, the ability to generate and select epitopes quickly and rationally makes it possible to significantly accelerate the identification and validation of recognition elements for their integration into new sensors.

This approach is expected to have a direct impact on the development of advanced diagnostic solutions, since the rapid development and selection of epitopes will allow significantly expanding the scope of these sensing technologies to a wide variety of detection applications. This includes not only the biomedical field, but also environmental and biotechnological environments, strengthening the connection between fundamental research and industrial application.

The use of machine learning algorithms and artificial intelligence is revolutionizing all fields in which it is applied. Biotechnology is no exception and, in fact, it could be said that bioinformatics is one of the first to adopt many of the new technologies that are being developed.