As we have already reported, Bence Kiss has received the PTE Inventor Innovation Award during the PTE Innovation Days organised on May 14-15. Mr. Kiss is an assistant professor at the Pécs Medical School, and he has also received the special award of the Chamber of Commerce and Industry of Pécs-Baranya. The topic of his application was “Immunogen epitope selection aided by deep learning 3D-antigen visualisation”. We discussed the background of the topic and the future opportunities of his research.
written by Rita Schweier
“We call matter that causes an immune reaction in the body immunogens. These are mostly proteins, but there are lipids, polysaccharides and nucleic acids that are immunogens as well. If there is also an adaptive immune response against a given immunogen, then we call the molecule (protein, polysaccharide) an antigen. Within antigens, we differentiate epitopes, which are areas of the molecule activating B-cell receptors (antibodies) and T-cell receptors. Which parts exactly of a molecule form an epitope could only be determined by experimenting, by the so-called crystallography method, which is a long and costly process, and it also presumes the presence of the antibody. However, when designing a new protein vaccine, it is useful to predetermine the protein sequences that are most likely to become epitopes. This has become possible and more widely available with the help of artificial intelligence, and we also use this method at the University of Pécs, and we test the process on animal studies” – he summarised the main points of his application.
As he said, the topic interests him because the future belongs to personalised medicine, and vaccine development and immune-biotechnology are fields specifically relevant in the field. As we know, the immune system – due to its nature – has a very high variability. Creating an in silico protein model is from the amino acid sequence forming proteins is possible with artificial intelligence, allowing a high percentage of immunogen sequences to be predicted. This could be useful for monoclonal antibody production, or immune therapy treatments and synthesising biomarkers. This know-how can be requested and used as core facility and service.
“The novelty of the topic is the deep learning algorithm developed by DeepMind (DeepMind Technologies Limited – British-American artificial intelligence research laboratory, a subsidiary of Google), which allows us to model 2D-protein systems from simple amino acid sequences. To my knowledge, we are the only ones using it in Pécs, and for us its usage is now routine. The other approach is the HLA (human leucocyte antigen) preciction system, developed by the Danish Technical University (DTU), providing the personalised nature of the research. Everyone is born with a unique immune system, which is genetically determined and does not change throughout our lives. Which genes are dominant in a patient can be determined with the new generation sequencing. Once this is done, we can use the so-called predictive algorithm to optimise the vaccine to their patterns. Before manufacture, we review the predictions with the highest percentages on a 3D model, then we finalise the most suitable sections in the special configuration. The innovative value is given by the routine usage of this method” – he added.
The project was started during the coronavirus pandemic in 2020 by the UP MS Department of Biochemistry and Medical Chemistry; the head of the research group is dr. Antal Tapodi. In the meantime, they have also established a spin-off company, Neo-Antigen Biotechnology Kft.
He mentioned about the future that they currently have two vaccine development projects running, both of which have the aforementioned bioinformatics process as their hearts. He believes that the field of bioinformatics is underrepresented in Hungary, which makes research on the topic even more important. Their research is currently aimed to deepen existing knowledge, following along with new opportunities and developments and to use new methods.
“This award is a huge honour. I am happy that this year, the call for applications was artificial intelligence, therefore topics connected with informatics and specifically bioinformatics could come into view. Efficient methods using artificial intelligence are indispensable in long term medicine” – he said.
Photo:
Szabolcs Csortos/UnivPécs