Search magnifier

חיפוש

עכשיו קל ונוח יותר למצוא כל כתבה ברוכים הבאים לאתר אקסטרה המחודש - עדכני ומגוון בשפע כתבות ומאמרים מקצועיים

World Leading Player in the field of Systems Immunology

World’s first machine-learning model, which predicts immune responses for drug development faster and more accurately than ever

צילום: shutterstock

CytoReason was established to tackle two major issues in the fight against disease. Firstly, the immune system – a vast, dynamic and complex cell-based system. It is the body’s checkpoint for hundreds of diseases and their treatment – and poorly understood at a system (cellular) level. Secondly, to drive more effective and productive drug discovery and development – a process that can cost two decades and two billion dollars just to bring one new drug to patients.

Put the complexity of the immune system together with the fundamental challenges of drug discovery and development and you have a challenge of epic proportions to overcome. This is the sweet spot for the application of machine learning approaches, with key answers laying at the intersection of big data and biotechnology.

At the root of the company is more than a decade of research by Prof. Shai Shen-Orr, of the Technion, aimed at leveraging massive and diverse amounts of data with new computational methods to drive unparalleled levels of understanding and prediction through the creation of unique disease models at a cellular level – CytoReason’s Cell-Centered Models.

These models act as a kind of Waze GPS for the immune system, creating an accurate and fast roadmap that is capable of understanding the direction of cellular activity, why it is going in a particular direction, and what specific cells and interactions between cells, cytokines and genes are causing the hold-ups. This ability to see cause and effect at a cellular level is critical in advancing drug discovery and development.

In 2016, another intersection – the one between CytoReason’s CEO, David Harel, a digital health veteran, and Prof. Shen-Orr – led to the birth of CytoReason. They, along with their four other cofounders, Dr. Elina Starosvetsky , Yuval Kalugny, Dr. Ksenya Kveler, and Dr. Renaud Gaujoux, saw the potential for what they were working on, and set about turning it from academic endeavor into commercial reality.

Young but Mature – CytoReason Found its Feet Quickly

The company, headquartered at the Azrieli Center, in Tel-Aviv, celebrated its second birthday in September. It now has 25 employees, compared to only 6 last year – with an average age of 43 and mostly coming from the fields of data science and biology with extensive pharma or biotech experience.

The company is already making its mark and has been revenue generating from day 1. Today, CytoReason is working with some of the biggest and best pharmaceutical organizations in the world, and is set to grow significantly.

Can your technology replace clinical trials?

David Harel"Our platform is not designed to be a substitute for clinical trials – and its application goes beyond just impacting clinical trials, having use across the discovery and development spectrum.

"But, it can significantly and positively impact clinical trials in a precise way by defining sub-populations more accurately, and by showing who will react to the drug. If you know exactly who will respond to the drug, the clinical trial has a better likelihood of success, reducing cost and potentially speeding up approval.

"Moreover, instead of identifying an appropriate medication for the patient by trial and error, which can cause suffering and result in loss of valuable time, it will be possible to match a drug that has the highest likelihood of working to an individual patient. This will enable pharma companies to come up with faster and more effective solutions".

How does the process work?

"We replicate biology to generate biological insights. The process starts with building a model of a given disease and a specific tissue. This makes it possible to begin asking the relevant questions of the model: If I change something in the cell, what happens to the cells next to it? How will a mutation effect a biological process in the system? What happens in the entire tissue? How does the immune system respond in an inflammatory setting?

"A drug makes a change at the biological level. By mapping each cell in the tissue and the molecules they interact with, we can look at the changes and compare all the models of diseases that we store in our systems. This allows us to see a change that indicates whether or not the drug is working on the biology of the disease. Developing a drug and bringing it to the market can cost up to a billion and a half dollars and take more than 15 years. If we can provide a projection that strengthens or weakens assumptions along the way, it is possible to save a lot of time and a lot of money, while providing validation".

What makes you unique in the industry?

"Today we are the leading player in the world in the field of Systems Immunology. This is the field that deals with the response of the immune system as a whole, rather than the response of individual cells. In this field, we are well ahead of the game. We don’t sell any software or files, we sell scientific discoveries, directions for new drugs, or drug-disease pairs.

"There are many companies that use public data and repositories of measurement data, but there is no company that has a learning model of the immune system, a critical component of every disease. We provide pharmaceutical companies with an ability to see a broader, fuller picture of their drug in the context of disease. With every collaboration we conduct, with every institution, our system becomes more and more precise, giving unparalleled insight to drug companies".

What challenges are you facing today?

"The field of AI is vast and filled with a lot of hype – many claim to be able to cure the world but fail to deliver. We have a clear focus and we deliver – and we do not apply a “black box” approach of throwing data into an algorithm and expect the results to be trusted. Everything we do, at every step of the process, follows a biological rationale and is made fully transparent.

"It is however, a challenge to cut through the hyperbole and noise. Given the success we are having though, we are getting our message across, so the next obstacle is the ability for companies to incorporate machine learning approaches into their R&D – from a technical and from an ethical perspective.

"The market is based on trust. The reputation and experience of companies in the field is very important. Key concerns are privacy of patients, data and intellectual property. Again, we have this covered, the highest level of confidentiality and ethical treatment of the data is of paramount importance to us. We never keep raw data and we never reveal the questions asked, or the answers generated, from our model to other collaborators. What our system retains is a summary – the big picture learning and context provided by every new data set".

Working with the World's Largest pharma companies

Is it possible that your solution may help eliminate animal experiments?

"There is clearly a difference between humans and animals – the immune systems of humans and animal models of disease specifically are quite different. For example, experimenting on mice, which traditionally has been seen as the best way of validating data, is not efficient in the case of immunological drugs and often the results in mice do not translate to humans.

"We are trying to solve a difficult problem in an area where animals do not provide a sufficiently effective solution. Our approach when it comes to translating mouse data to human data is based on the same learning abilities as our human cell centered models. It is unlikely to replace mouse experiments but it will make them more targeted, more accurate, more informative and thus more effective – in the same way that our platform can across the entire discovery and development process".

And what about people?

"If you can make your clinical trials more “personalized” – more targeted – you can make them more effective. If you already have an understanding of who will respond to a treatment, you would be able to cut out a whole group of people who will not respond. This can make the trials smaller and even more informative which can speed up development and get new drugs to patients more quickly. Already the FDA, which leads the regulation in the field pharmaceuticals, has begun to set up for these changes and to start using smart diagnostics".

Who are your customers?

"Since the company was founded, our first customers have been the world's largest pharma companies, and organizations such as the Parker Institute for Cancer Immunotherapy (Parker was the founder of Napster and President of Facebook). Since then to the present day, we have been working with a growing list of the world's largest pharma companies. We also work with biotech investment funds, who come to us to help their portfolio companies".

The Largest Database

Is there a clear success story you can tell us about?

"Six employees during the training period helped prepare a presentation to one of the world's largest pharmaceutical companies. About a month before the presentation, we ran the system on a certain disease and discovered that the mechanism of action of the drug was drastically different from the mechanism that the pharma company believed was at play. We presented the results to that company, who, unbeknown to us had also been further investigating the mechanism for some considerable time.

"The project leader happened to be present at the meeting and immediately confirmed that our results were correct. Our results were so accurate that they thought that they had been leaked to us and pressured us to reveal who leaked them – they simply could not comprehend how we had discovered in just 4 weeks what had taken them three years and $50 million to uncover".

Could you spell out your vision?

"We want to build the most accurate system for predicting immune responses. We are building the largest database and the most accurate models of the immune system. CytoReason is re-defining the context of the immune system at a cellular level in order to better understand disease and drive more effective drug discovery and development. Our leading-edge machine-learning approach identifies “cause and effect” of the cellular relationships that lie at the heart of treating disease. Faster and more accurately than ever before.

"CytoReason’s unique platform computationally replicates biological processes in creating cell-centered models – blueprints of immune activity within the context of disease, affected tissues and drug therapy. This unique, and continually learning process, will become the gold standard that will help ask, and answer, the right questions that will drive the development of countless life-saving and life-improving innovative new therapies for the patients that so desperately need them”.

טוב לדעת