Immunai is mapping the immune system at unprecedented scale and granularity. The map, paired with machine learning, looks at how the immune system will respond to drug targets, offering an affordable way to prevent expensive drug failures.
The ultimate goal is to market immune treatments for diseases like cancer faster than ever before.
In this week’s conversation, Noam Solomon, CEO and co-founder of Immunai, covers the data gap in drug discovery and how machine learning (ML) can solve it, how to de-risk early-stage drug discovery, predictions for artificial intelligence (AI), and more.
Table of contents
About Immunai
Immunai was founded on the belief that immunomics and machine learning could unlock the next generation of immunomodulatory drugs.
The company has leveraged its technologies and more than 25 academic partnerships to build AMICA (Annotated Multiomic Immune Cell Atlas), the world’s largest data atlas of clinically-annotated single-cell immune data.
This atlas, combined with Immunai’s team of immunologists and computational biologists, provides insights into underlying biology that drives disease etiology, allowing the company and its partners to discover novel targets as well as analyze and develop existing pipeline compounds.
Pioneering immune system mapping for disease insight
Solomon said the company started more than five years ago to fully map the immune system.
“The immune system is really implicated in every health condition: Our response to therapeutics, how we cope with viral and bacterial infections, how we age. We understood early on that if we want to improve the therapeutic development process, we want to study the common denominator to all disease, which was the immune system.”
Immunai is taking a step back from drug development taking on diseases and conditions to see what the steps leading up to the disease look like.
“When you study a cancer patient and you study the biopsy, it’s already when it’s a very late stage. If you can study the immune system, you can study it even before the patient had cancer. And you can also look at the common denominator because there are similarities and relevance between patients.
“Studying the immune system is allowing us to really look at the root cause analysis for disease, and then we can start figuring out different mechanisms.”
The role of AI in mapping the immune system
Artificial intelligence has taken a prominent role in many fields, none more so than in biotechnology and drug discovery. But just how important is AI in developing new treatments?
“We believe that by mapping the human immune system, by mapping thousands and then tens of thousands, and then even millions of patients, we will be able to map the human immune system and then apply computational models and AI to try to decode, again, the root cause analysis and figure out the immune intelligence that is needed to make better decisions and develop better therapeutics,” Solomon explained.
Amica: “A Google map for the immune system”
AMICA is the core IP at the company.
“AI or ML means nothing if you don’t have a database that is large enough and smart enough and clean enough to run complex and expensive models on it,” Solomon said.
“We’ve been measuring the immune system first from human patients that are being treated with immune-modulating drugs for oncology, for autoimmunity, and for other immune-mediated diseases.
“That means that we are looking into a cohort of patients, different time points before and after a drug has been administered, and we’re trying to compare or mine or dissect the differences between responders to the drug and non-responders to the drug, or patients that have a toxic event. We want to understand why.
“The ability to leverage computational models to find the features that underscore the difference between the responders and non-responders, between toxic events and non-toxic events, between healthy people and disease patients, is allowing us over time to better understand how to improve the drug discovery and drug development process.”
“What we are doing (is) kind of Google Maps for the immune system.”
Immunai’s data sources
Solomon said the data come from collaboration with a large network of academic partners, hospitals all over the world, as well as biotech and biopharma companies. Another source of data is preclinical work Immunai does in the lab. Data also are sourced from what is published in the public domain.
“I think about AMICA as a knowledge sphere that is growing. But in reality, there are a lot of holes in this field. So maybe more like a Swiss cheese. And over time, we’re able to identify the holes or the gaps and then infuse more data to bridge the gaps and close the holes.
“As the data grows and the knowledge gaps are closed, we’re able to find relevant insights that will explain why certain patients will respond to drugs and others won’t, and how to find the next generation of therapeutics that are going to overcome the resistance mechanisms.”
Harnessing AI to improve clinical trial success
Solomon said that with success rates and return on investment (ROI) falling, pharma and biotech companies are moving away from investment in drug discovery, preferring to invest in later stage assets already cleared in phase 1 and 2.
“I think that if this is going to continue, we’re going to see less innovation, and that’s not good for patients. I hope that with platforms like Immunai, where we are aiming to improve the statistics, to improve the phase 1 success rate, to improve the phase 2 success rate, it will lead to better ROI.
“Then we’re going to see further investments in the early phases, because that’s very important to invest in innovation and to improve patients’ outcome from the beginning.”
Solomon said AI will significantly improve the statistics of clinical trial design. He noted factors such as optimal dose and schedule, the right combination of drugs, or who should and should not be given a drug, will be informed by AI.
“On top of this, I think that there is a momentum to start looking at molecules earlier on the preclinical stage, analyze how the molecules are interacting with cell-to-cell interactions or interactions, perturbations in earlier preclinical models, in vitro, ex vivo, and in vivo, and be able to predict or deduce what will happen later in the clinical stages.
“I believe that by creating the right data sets and applying the right tools, we are going to see a transformation in this direction.”
The potential of immune system mapping in preventive medicine
Immunai works with five of the top 15 pharma companies, which Solomon said gives the company access to cutting-edge drugs and clinical trials.
“We really believe that the platform gets better every time we inform a decision, because more data is flowing into AMICA. I hope over time, with the statistics being published and readily available, we are going to really inform decisions, end to end.”
“Our platform can really inform these decisions. So, for example, if you are trying to develop a drug and you don’t know which of the top 10 oncology indications you want to go after, and we tell you, you need to go after those three and ignore the seven others. If you trust us, and we are right, we saved you a lot of time in the process, not just money and not just better statistics.”
However, it is in preventive medicine that Solomon hopes AMICA can really make a difference, although he admits the company isn’t there yet.
“The immune system is the core reason for what we are able to do differently before people are getting sick. If you are able to identify someone before they get cancer, let’s say a year before, maybe you can give preventive care. It’s going to be much, much better. I think preventive medicine with this approach of immune monitoring patients is really where the platform can add value in the long run.
“I am convinced that with more data monitoring people when they are healthy there is so much to be done about preventive care. It’s not just about diagnostics for patients that have early signs of cancer. I’m talking about being able to start identifying small variants in parameters, even a year or two or three before people are going to get sick,” Solomon said.
“A few years ago, I described what we are doing as kind of Google Maps for the immune system. But in Google Maps it’s not just to say where you are, but also if you want to go from point A to point B. If we have the human immune system and you are in point A and you want to get to point B, if we are able to understand this, we should be able to guide you in the right direction.”
Solomon believes that the hope of the past few years will be transformed into success stories over the next few years.
“These success stories are going to, I hope at least, create a momentum and a way for more companies to invest in novel therapeutics, novel diagnostics, and preventive care. I think it’s going to be a very interesting decade,” Solomon concluded.
To learn more about this topic
Here are some links to more articles on the subject of AI and drug discovery.