ProjectBlog: Immune 2.0 — Getting us Ready for the Next Pandemic

Blog: Immune 2.0 — Getting us Ready for the Next Pandemic

Killer diseases have been invading our lives for all of human history. If we go back to the fourteenth century, we can see the horrors of the Black Death. It killed 60% of Europe and an estimated 150 million people. Although the mass destruction, no viable therapies were created because many thought of it as god’s anger on the world.

This seems like something that would never happen nowadays but you would be surprised how close we have gotten to a world pandemic. Our reaction to it is something that I always think about:

Will we really be ready if an epidemic came? How fast would we be able to create a vaccine and distribute it to the world?

It is not only for a world pandemic, so many diseases catch us off guard and leave us scrambling for treatments. I really think this is not acceptable. Late diagnosis and our reactive healthcare is a serious problem with huge impact.

If you have a car, you will most likely not wait to get into an accident to change your brakes. You will most likely check your brakes and prevent it before it happens. If we are treating an inanimate object with such care, we should treat our bodies better but we are waiting for pain and discomfort from a patient to diagnose.

I believe in a future without all of these diseases and a large part of this future is enhancing the most adaptable system in our body, the immune system.

The immune system is our line of defence, keeping the world of viruses and deadly bacteria from killing us. It is kind of like the doctor inside of us, constantly checking up on ourselves and solving something if it isn’t right.

In the future, we would like to produce our own immune system, but for now we should solve some of our current problems so we can get there in the first place.

That is why we created Immune 2.0, using machine learning to improve immunotherapies.

A large standing problem with our therapies have been their predictability. The immune system is the greatest system of the body in my opinion because there is so much variability. This variability comes from a structure called T-cell receptors (TCRs)

The T-cell receptor is what binds to the pathogens/antigens in our body. They are able to recognize a specific antigen and decide to either kill the cell, or leave it. Kind of like police officers or detectives.

Our immune systems have to be able to bind to every antigen in our body without missing one or else we will get some sort of infection or virus if exposed. As a result, we have a possibility to create over 1 000 000 000 000 000 000 different T-cell receptors. Although it is great to know our body is working hard to adapt to prevent possible diseases, this makes it hard to predict what will happen with immunotherapies and when certain viruses enter our body.

With Immune 2.0, we are trying to change that.

We are utilizing the power of machine learning models to predict these responses. Our model is able to receive a specific amino acid sequence from a T cell and spit out the correlating part of an antigen that the T cell binds to.

To be exact, we input it a section of the T cell called the complementary determining region (CDR). CDRs make up the region of the T cell that actually bind to the antigen. Therefore, they are the most variable and vital part of the system. A set of these CDRs make up what is known as the paratope. This paratope binds to the part of the antigen known as the epitope. This is the other part of our data.

We used a large dataset to train our model on known paratopes with corresponding epitopes for a variety of human leukocyte antigen (HLA) types.

We now have a model that can take the factors that go into T cell binding and learn the pattern by itself and predict this binding. Let’s discuss a few places where it matters.


Autoimmune diseases happen when the immune system gets confused and starts attacking the own body instead of foriegn pathogens.

We have these antigens in our body called self-antigens which are located on normal cells that are not really causing any problems in the body. We have a couple of these self-reactive receptors in our massive inventory that can sometimes be produced. But usually, they are usually killed before they are active to prevent us from having T cells that switch teams and fight against it.

The process from when a pathogen is in the body to the production and binding of a T-cell is complicated and when any one of these mechanisms mess up, we have the possibility of that T cell to attack ourselves. This is what we call an autoimmune disease.

There are many diseases that fall on these diseases including arthritis, type 1 diabetes, multiple sclerosis, addison’s disease, and celiac disease. These problems affect so many with one out of five people in America suffer from one of the autoimmune diseases.

Much is needed to learn about these diseases including even the basics of what causes them! With new technology coming out in sequencing, it is likely that we will see an influx of epitope and paratope sequencing data for autoimmune diseases.

Knowing the map of T-cell epitopes in different autoimmune diseases has massive implications for not only the future of research but for autoimmune disease treatment. With Immune 2.0 you can input self-antigen sequences and be able to know the TCR sequences opening the door to create immunotherapies. A large limit for current therapies for these diseases are the lack of ability to associate antibodies with a clear target which can also be solved with Immune 2.0.


One of our biggest applications is improving our reaction and understanding to a worldwide spread virus where time is so vital.

Many believe we are approaching ‘Disease X’ the new pandemic which we will all be very unprepared for. At Immune 2.0, we think that is very preventable. With our system we can increase the speed of creating a vaccine for this pandemic.

When we find what antigen/pathogen is causing the pandemic, we can sequence it and enter it into our system. Outputted is our T-cell sequence that has the highest affinity to bind and kill this pathogen.

Within seconds, we have the T-cell which can prevent this world pandemic which again, opens the door to innovative treatments.

After we have eliminated the first wave, in order to eradicate a disease, we must understand it. Immune 2.0 can again be used to research the adaptive immune response to different viruses and infections.


Obviously, one of the most pressing diseases that have been affecting humanity for centuries is cancer. Cancer is an interesting disease in terms of our immune system.

Something to know about cancer is that it is really just a part of our body. We get tumors all the time but our body is usually able to defeat it. We get cancer when these tumors interact with the mechanism of our immune system and are able to hide from them.

Over recent years, we have seen new innovative therapies trying to re-expose our blind immune system to the troubles of cancer. These include CAR T-cell therapies that use genetically engineered TCRs specifically targeted to cancer cells to help destroy them and recruit more T-cells. We have also seen developments in neoantigen vaccines which assist our immune system in targeting cancer cells accurately.

At Immune 2.0, we want to fit in a lot of places in this space where is needed. A major one is obviously prediction. Due to cancer being a part of our body, the way everybody response and develops cancer is very different. The need for caring to the patient rather than the population is high. We can sequence people T-cells in their blood and antigens on their cells to give them a good look at how their body is dealing with their disease and where the windows for innovative treatments are.


With such an impactful start. It is super exciting where we can go with Immune 2.0. We have multiple ideas for the future.

A main one is expanding our data to microbiome data. Our microbiome is the set of microorganisms living in our body, mainly living along our gastrointestinal tract. Researchers seem to finding new ways our microbiome controls our life. They have found it effects everything from cellular processes such as metabolism to aspects of our life we thought to be uncontrolled like our behaviour. We want to expand patients knowledge about their lives by analyzing their gut bacteria and how it can affect them in the future.

We also want to see about predicting binding with conformational epitopes. Conformational epitopes are epitopes that are recognized by T-cells not only by their sequence but their protein folding. Our model currently only recognizes linear epitopes which facilitate binding based on their sequence. The main limitation of this prediction of conformational epitopes is because nowone has figured out the true process of protein folding and how to predict it. But helping answer that question and unlocking more potential to revolutionize immunotherapy.

We are truly excited for this future. We live in such a great time to be alive. Not only do we live for longer than ever but we have the tools to make so many people’s lives better. That is what we want to do with Immune 2.0.

To learn more visit our website at!

Source: Artificial Intelligence on Medium

Leave a Reply

Your email address will not be published. Required fields are marked *

Back To Top

Display your work in a bold & confident manner. Sometimes it’s easy for your creativity to stand out from the crowd.