PepDiscovery of Novel Peptide Drug Candidates by Artificial Intelligence Design. Interview conducted by Ben Woolf and Deborah Amsallem on 15th July 2019.
Founded in 2011, the company is involved in various successful discovery projects with academic and industry partners, ranging from uses in agriculture to immune modulators, diabetes, and several molecules in different R&D stages.
Pepticom raised 5 million to revolutionize the way we discover active peptides. Reducing molecule discovery from years of research and millions of dollars to only a few months and immense A.I computing power.
Dr. Immanuel Lerner, Ph.D., CEO of Pepticom, and Dr Maayan Elias Robicsek, Pepticom’s Business Development Officer, warmly received the JLM Biocity team at Pepticom offices, located at the Givat Ram campus. Dr Lerner answered our questions about his academic background, and his work as the leader of a promising start up.
What are Peptides?
Peptides have gained increased interest as therapeutics during recent years. Peptides are recognized for being highly selective and efficacious and, at the same time, relatively safe and well tolerated. Consequently, there is an increased interest in peptides in pharmaceutical research and development (R&D), and approximately 140 peptide therapeutics are currently being evaluated in clinical trials.
During the past decade, peptides have gained a wide range of applications in medicine and biotechnology, and therapeutic peptide research is also currently experiencing a renaissance for commercial reasons. In terms of value, Peptide based therapeutics are currently experiencing a renaissance in the global market with collective sales of more than US$ 20 Billion in 2017. Overall, the global peptide therapeutics market is expected to surpass US$ 50 Billion by 2024.
Naturally occurring peptides are often not directly suitable for use as convenient therapeutics because they have intrinsic weaknesses, including poor chemical and physical stability, and a short circulating plasma half-life. Some of these weaknesses have been successfully resolved through ‘traditional design’ of therapeutic peptides.
In addition, currently used peptide discovery methods limited capabilities in discovering novel peptide drug candidates. Consequently, most marketed peptide drugs are modified versions or derivatives of a natural peptide.
To help broaden the applicability of peptides as therapeutics it has now become necessary to explore new routes beyond traditional peptide design.
Pepticom aims to streamline peptide drug discovery by introducing Artificial Intelligence (AI). The algorithmic platform combines principles from economic and molecular mechanics in a reinforcement learning methodologies to generate a wide scope of active molecules with higher success rates. In so doing, Pepticom aims to lower the number of R&D candidates synthesized and tested with a similar reduction in time and costs when compared with traditional lab-based discovery methods.
The use of computational technology allows Pepticom to address complex targets unreachable by laboratory methods and to further enhance R&D. The platform utilizes the structure of the target protein to design innovative peptides at atomic resolution. It is capable of designing several peptides’ chemistry including linear, helical, and cyclic peptides and to incorporate D and non-natural amino acids in the process.
What is the professional background of Pepticoms founders?
Immanuel Lerner (I.L): Dr.Amit (Michaeli) the co-founder and I did a bachelor’s degree in biology, he then moved to computational chemistry and did a PHD with Maayan in same field. Moving forward I did my PhD and postdoctorat in biology in the field of Cancer and circadian biology.
How was the idea originally formed?
I.L.: We first started thinking about the company during our studies. When doing our PhDs, we were working on technology that could enable the binding of peptides to proteins. However, we didn’t have much attraction in the academy as people didn’t think it would be helpful. Then, while working on an assay regarding cancer and its relationship with the inflammatory environment, I was interested if one could block some kind of immunologically interaction, so we bought peptides out of our own pocket, that were designed using our own technology, and it worked!
At the beginning we had a difficult time to raise a penny. Then I had the good fortune of publishing my last paper with 2018 Nobel prize winner Michael Roshbash, and this supported our efforts for our first fundraiser, in which Maayan joined us as the first employee and has been essential to the business ever since.
Can you define your company’s technology?
There is a need to discover potential active peptides. The way it’s done today involves screening random libraries, billions of different molecules. But to discover the right molecule is it’s like winning the American lottery powerball 4 times in a row.
Furthermore, more than 85% of the costs incurred in pharmaceutical projects come from failed discovery and development costs. The very first discovery phase of a drug’s development takes 2-3 years before the drug can potentially get to trial, so when development fails – especially toward the end of a process that can last up to ten years – it represents a painful loss of substantial investment ranging from anything from $100m upwards.
Our technology enables discovery of the most advanced peptide-based drug candidates much faster, more comprehensive and successfully predict and rule out “problematic” or risky drug candidates. Internally we have built an A.I software and package software aimed at this; once you have a potential peptide in the computer it is then synthesized in real life and tested chemically. Dropping a process that takes 2 – 3 years, once optimized, to a few months.
You are talking about predicting risks, can you tell us more?
I.L.: The idea is to predict the risk in later testing. We can program the computer with very trivial tasks, a problem of solubility for example, there is no way to control a library of a billion molecules when you synthesize them, but when you use a computer you can put limits to determine which peptide attributes we are looking for.
Take a problem such as; toxicity, if you know where the problem is and can define it, then you can integrate it into discovery, this is important because what happens today is that a molecule is discovered from a library and has good activity, you then find toxicity issues so you need to throw it to garbage and repeat. Combining this process into our discovery will be extremely efficient so we don’t have to test a molecule which will be problematic in later stage testing. This ability is a big reduction of risk.
What differentiates your company and technology from your competitors?
It’s ideas from A.I that are implicated in computational chemistry, therefore we have a strong background in all these areas and the ability to apply the necessary tools and theories. This makes us different than other applications out there.
When a company, pharma is starting a project, they have 90% failure rate and need 10 years to develop, or sometimes longer (10-12) Clinical, and claim its around $1.5b taking into account all failures and takes into account a lot of variables but still cost around $250m.
Now the first step, starting from scratch, you know you want to inhibit something in the body, it takes 3 years to have a lead molecule, something that shows activity and can be pursued, and even if you have a molecule there and you spend a few million to develop it, you still have a huge percentage of them failing later on because of different problems encountered.
The idea is taking these 3 years and making them much more efficient, something like half a year, much cheaper and take some of the risks that could make molecule fail later stages and reduce them.
It may not be 0 but even to increase the success rate by 10% is a huge impact in the industry.
I.L.: Using computers as an aid is a no brainer, everybody understands you need to persuade and show your capabilities. It took us a long time to get where we are, with few resources, and now we can do much more very fast. I think our technology has many more advancements to undergo and other A.I applications are not rival with Pepticom even now.
Video of Pepticom’s Technology
Can you give a short description of your evolution and achievements?
I.L: We are enthusiastic about taking the A.I platform and fusing it with micro array technologies, usual methods require synthesis up to a month and costs $100’s per peptide, whereas we can screen 1000’s of active peptides at little cost.
Until today computer tells us we need to synthesize 20 peptides in order to test something, so you synthesize, and it takes a month. So it’s still 100 times more efficient than other technologies, but still takes time. And to maintain success rates we need to screen a few dozen. We have started with interesting results, using micro arrays, which is printing, a peptide micro array, every active peptide shows as a red dot on the system, and less active peptides are a lighter shade, the more red, the more active toward what we are looking for, this is before optimizing and screening 1000’s of peptides with lots of activity. The cost is 1/1000 of synthesizing in a different way and much faster as you can test a few thousand in one go.
If you would screen random peptides the problem is 10 to the power of 30 you would spend your whole life screening, with A.I you can understand the problem, bring the solution much faster, this two technologies are fitting like a glove to a hand. This allows us to understand the problem and bring a solution much faster so we can develop our own peptides to be licensed or developed later on the pharmaceuticals.
Maayan Elias (M.E.): “20-30% of peptides that we predict are active, which is very high compared to industry standard, and is only a starting point to further optimization. We are talking about the next generation of peptide. The chemistry and non natural amino acid, and other modifications that increase their stability in the body which can increase the binding affinity. This way we can answer the demands of the industry of oncology and immunology”.
What are the current challenges in the development of the technology?
I.L: First is general development and discovery of the best molecules that applies to most biotech companies, this involves developing from scratch, optimizing and regularly updating our in-house platform.
Secondly is penetrating the pharmaceutical market, building business to bring serious revenues, collaboration and creation of new IP.
M.E.: regarding Pharmaceuticals, it’s a conservative market and need to show that they will benefit from innovation in both the short term and the long run. “
What is the culture of the company regarding human resources and relationship amongst the team?
I.L.: I want to work with people who are happy and enthusiastic when they come to work and promote engagement. Working with ambitious people with exceptional capabilities is vital to delivering the vision of top management. Other necessary requirements are to have good interactions with those you work with for 10-12 hours a day, and more important than filling a quartile is delivering high quality deliverables and building trust through open communication.
M.E. “It’s very important to us that people who come and join will have the same enthusiasm as the company, we want to enjoy the people we are working with, and that they fit in as one big family”.
Is there anything you do to specifically nurture this culture?
I.L: Nothing formal, but the whole company is around this. If you are in the mentality of not being involved you will not fit in. It’s the privilege of a small team, once we grow then we will need an ordered way to improve and sustain a nurtured culture.
We have a hierarchy but we desire that people are comfortable to express problems to whoever they want, the benefit with a small team is that problems do not fall between gaps and there is always someone to report to.
An issue that can arise is when someone is having a hard time to solve his problem but does not share with others, this can become a disaster and we seek to avoid this.
We have found that our team are very dedicated, we do our best not to micromanage and give people flexibility and ownership of their work.
Do you have any advice for future Phds and entrepreneurs?
I.L: I think you need to look in the mirror and see that there is a delicate decision in the beginning, that 99% of the time it’s not easy to know if you really believe in what you are doing, the technology, application and team. Ask yourself if you keep going or quit. If you talk to an industry professional who you know well and trust that tells you something is a waste of time then you must take this and break it down in an analytical way, and if they are correct, then there is a big problem.
M.E.: “surrounding yourself with good people is not easy, mentors, team, consultants, but also service providers, your lawyer and CPA are most important at the beginning because you will encounter a lot of issues that you have never heard about, so you need help and input from someone you trust, as a wrong contract can break the company, trivial but very important.”
What makes Israel such a good place to succeed ?
I.L: I think that Israel is a good place to start a start-up, when mentioned that we are an Israeli company to people from the U.S, Europe and the far east, it’s a very good position, if anything, only one step below Silicon Valley.
In some aspects, government funding in Europe is very good as they receive a higher return for R&D expenses, however this is a factor that can hurt companies, because they may receive more funding but be destined for failure, in Israel you do not have this opportunity, you must survive or be closed.
If you can survive in Israel, you can survive anywhere !
What about working in Jerusalem specifically?
I.L: It is very important to have a company in Jerusalem, though it is not straightforward as we compete with Tel’ aviv, Moadin and Rehovot, but Jerusalem is strong and there are growing ways to help companies in the city.
M.E.: “There are less VC’s and angels here, the structure of biotech, medical centers and university support is very strong. Therefore the potential is huge and continually improving”.