Whenever the immune system fights a pathogen, it has a memory of that experience, whether it's with an invader or a vaccine that mimics an infection, for example. Scientists have now developed a way to mine all of the rich history carried in a person's immune system, which may improve the ability to diagnose diseases like influenza or COVID-19, as well as other chronic issues like diabetes or lupus. In a new study reported in Science, scientists assessed the immune systems of 600 volunteers, including healthy individuals and patients with diseases such as type 1 diabetes. The investigators applied a machine learning tool called Mal-ID to analyze the structure and sequences carried in T and B cells and diagnose disorders in the study volunteers.
"The diagnostic toolkits that we use today don't make much use of the immune system's internal record of the diseases it has encountered," said postdoctoral researcher Maxim Zaslavsky, Ph.D. "But our immune system is constantly surveilling our bodies with B and T cells, which act like molecular threat sensors. Combining information from the two main arms of the immune system gives us a more complete picture of the immune system's response to disease and the pathways to autoimmunity and vaccine response."
Along with diagnosing some diseases that are challenging to identify, like lupus, the study authors suggested that Mal-ID might also be able to monitor immunotherapy courses or disease states to inform clinical decisions or treatment approaches.
The machine learning tools used in this study are based on large language models, but they look at antibody structures and the receptors on immune cells like B or T cells. Patterns in data can be identified by these computational tools to look for binding preferences. This can reveal what has activated an immune system to produce fighters like B or T cells that can attack invaders or deal with threats.
"The sequences of these immune receptors are highly variable," explained Zaslavsky. "This variability helps the immune system detect virtually anything, but also makes it harder for us to interpret what these immune cells are targeting.
The scientists used immune receptor signaling sequences to learn what the immune system has responded to over time. In this study, the researchers generated a dataset of sequences from over 16 million B cell receptors and over 25 million T cell receptors. This data came from almost 600 people who were either healthy, had COVID-19, lupus, type 1 diabetes, HIV, or were recent recipients of an influenza vaccine.
This work showed that lupus and type 1 diabetes could be seen through sequences of T cells receptors, while HIV or COVID-19 was identifiable with B cell receptor sequences. While only a few conditions were tested in this study, the investigators suggested that it could be easily modified to detect many other disorders.
Sources: Stanford University, Science
