Compugen’s Powers of Prediction




Episode Loading...




PharmaTelevision requires Javascript enabled and Adobe Flash Player to watch our programmes. If you do not have Flash installed, you can download it for free from the Adobe Flash homepage.

Improve your Internet experience and start watching exciting new video content.

Video title: Compugen’s Powers of Prediction
Released on: March 31, 2009. © PharmaVentures Ltd
Share/save this page:
Email
Bookmark
Facebook
Twitter
LinkedIn
Follow us:
RSS
Twitter
  • Summary
  • Transcript
  • Participants
  • Company
Do we really have a good understanding of how life works at a molecular level? Martin Gerstel, Chairman of Compugen, explains how the company has created a computer that can accurately model the behaviour of molecules, information that can then be used to predict drug targets.
Compugen origin and development.
Fintan Walton:
Hello and welcome to PharmaVentures business review here in San Francisco. On this show I have Martin Gerstel who is CEO of Compugen which is based in Israel. Welcome to the show.
Martin Gerstel:
Thank you very much, pleasure to be here.
Fintan Walton:
MartinGerstel,Compugen is a company that has its origins around computers, it's a computer based company. Tell us about the origin of that company and how it's developed into the Compugen as we know today?
Martin Gerstel:
I was introduced to the company about four years after it had been founded since though it was did something consulting for a venture firm, the company had created a special purpose computer a computer that could be one thing and one thing only and that is to analyze all of the biological data that was been created around the world this was the early 90's to analyze all of that data to try to make sense out of this in term of the human genome project and everybody is buying the sequencers and unbelievable amounts of raw data was being created but these were tiny little pieces of data the issue of how to put them together and make sense out of them and there were three kids from a very special research program in the Israeli Army called the Talpiot program that heard about this problem got some money from the government, $50,000 and in nine months created a computer that was a 1000 times faster than any in the world to do this project that it completed mispriced it. But they rapidly became the world leader with 90% of the market I believe to this day that we've patent that's filed in the United States with biological data's analyzed by a Compugen computer. They are very doing this they work with everyone with all of the leading centre's that were doing " doing this type of work and they rapidly discover that what they had done was they created a situation where it was raw data in garbage out of thousand times faster because the algorithms that were used to analyze this data just were very crewed, they were done by some Ph.D. student in the university and life science was changing everyday and understanding of it. So they decided that what they would do is hire some biologists and being with some of the best algorithm people in the world. Israel of course is known for that it's military unmanned, air surveillance aircraft and whatever. They were experts in this so they began to develop new algorithms to analysis the data and what they found out in doing this with the idea of creating computers they found out that the world would just -- was wrong in their understanding of how life worked at molecular level just absolutely wrong.
Fintan Walton:
So it was using algorithms to understand data?
Martin Gerstel:
Yes, yes exactly enormous, enormous amounts of data, dirty data from the standpoint that the machines that were creating that weren't that accurate and whatever and tiny little pieces and trying to put this thing together without and understanding of what you were looking for it's like doing a jigsaw puzzle but not knowing if you have all the pieces and having -- don't having that clue what the final picture should look like. All right and as I said they quickly discovered that the world was wrong in its understanding of how life work fundamentally wrong and that " that's when I came in as a consultant from this VC firm and they were going to use this knowledge to create another generation of computers. I looked that and said hey if this is for real and I wasn't in a position to judge it but I went around with some of my friends in the science world and found that " that this could very well be truth it is. They are one- already one of a leading life science companies in the world. We should forget about the computers and become a life science company and that's what we did. Venture capitalist got very upset whatever and we did something very unpopular in today's world we went in for long term research how this life work at the molecular level?
Fintan Walton:
But you mentioned the word that wrong mean that we, our interpretation of was wrong but maybe not as severe as that, it's maybe they've started to understand a little bit more about life than we did know?
Martin Gerstel:
It's pretty.
Fintan Walton:
And wasn't that fundamental?
Martin Gerstel:
It's. you know it's perhaps a slight over statement but not much because what life you know what I called the electricity of life what makes it all work. Our genes expressing transcripts becoming proteins and then proteins, peptides, pathways this is life, this is how everything, how we function. I'm talking now because proteins are kind of interacting and that amazing. Alright now at that time the central dogma of biology was that one gene expressed one transcript became one protein. Now people around the world were beginning to see exceptions to that, were beginning to see occasionally you would have a gene that would express two different transcripts and perhaps two different proteins, right? What their analysis came through was that almost the entire human genome did this called alternative spice, you get 10, 20, 30 different proteins from the same, from the same gene this was heresy. I mean, what I worked one when we took this to the leading pharmaceutical and research centers, we were nicely laughed at this can't be. Now there was only one recognition of work of our early statement and that is the economist that a story a few years ago on sort of something about life science and they mentioned in there about you know the tiny little Compugen identified this years earlier. If you remember, it's interesting, if you remember when the human genome project was completed there was a private project and a public.
Fintan Walton:
Sure.
Martin Gerstel:
The headlines were scientist astonished by how few genes there are considering how many proteins that's exactly what Compugen has said four-years earlier and no one believed us.
Compugen's basic research platform based on computational biology.
Fintan Walton:
Right, so it's the underlying platform, the differentiator what makes Compugen is it's computer part with algorithms in the context of life and " and understanding taking the data that comes from day-to-day information is coming out of basic research and then turning it into a way which we can identify new drugs?
Martin Gerstel:
Yes to be more specific what it is using computational biology or algorithms and past knowledge that we've accumulated to create predictive models of important aspects of how life works at the molecular level. How to genes express transcripts? How do transcripts become proteins? How do proteins become peptides? How do proteins fall?
Fintan Walton:
Yeah.
Martin Gerstel:
We the rest of the world in general is creating descriptive, trying to describe it and have a model of that describes it. We have no interest in the model that describes. We want a model that predicts what should happen and you know I think the record is very clear I mean the amount the discoveries that we are now making in so many different fields of therapy area are really quite astonishing.
Compugen experimental research business plan and it's products
Fintan Walton:
So how does that stand fit into a business because obviously -
Martin Gerstel:
Right. Yes.
Fintan Walton:
It's your team may win noble prizes in the future but what we worked were "
Martin Gerstel:
Right.
Fintan Walton:
Presumably you want to see is how you turn this into a business?
Martin Gerstel:
Right.
Fintan Walton:
That can actually develop parts.
Martin Gerstel:
Let me give you a one example. We have as so far we create things what we call discovery platforms which incorporate all of our capabilities targeted at a specific area, all right? let me talk about one of these platforms because it's a very important area. I mentioned that life proteins interacting and there are different families of proteins, by far the most important family of, protein through the pharmaceutical initiative what called the GPCRs.
Fintan Walton:
Yeah.
Martin Gerstel:
More than half the drugs in the world act on GPCRs. It's essentially every major pharmaceutical company has a program to try and find new Ligands for GPCRs, their membrane bounce or there are reasons why it's very difficult to find them by traditional experimental means. on average about two new Ligands for GPCRs are found each year by the industry alright.
Fintan Walton:
Alright.
Martin Gerstel:
All done experimentally which most of the research in the industry is done experimentally by that I mean you go to your lab and you run all these experiments in you either high throughput combinatorial or whatever to hope that you see something that is interesting, something that works. Here is what we did. We " with our knowledge of alternative displacing and other issues such as gene structuring, gene function we created a predictive model in our computers of all the transcripts that could possibly exist in a human. We then studied how do transcripts become proteins, post transcriptions added in other aspects of that and based on that knowledge we created in our computers and in in silico proteome all the proteins that could possibly exist in the human body. We then peptides the natural Ligands for GPCRs are typically peptides which are pieces of protein. The peptides are formed by proteins being cleaved being cut into individual pieces. We created a machine learning system that looked that a group of peptides that were known to be Ligands for GPCRs, a group that were known not to be and that created a system that if you gave it a peptide that it had never seen before, would say that has a 20% chance of being a Ligand, a 70% chance or no chance at all whatever, right? We took this machine learning system and applied it against our Insilico Peptidome and created a few 100 peptides, everything I have said is inside the computer in-silico not a single experiment has been done so far.
Fintan Walton:
Right, right.
Martin Gerstel:
Alright and what you have is a prediction on top of a prediction on top of a prediction on top of a prediction on top of a prediction, alright? In each one of those predictions is of a massively complicated function of life, genes expressing transcripts in an amazingly complex phenomenon each one of them, all right? So now we have our experimental, we have our few hundred, we took and we picked 33 of them. We synthesis them, we threw them against a battery of important GPCRs, 8 of them were real, all right? 8 out of the 25% of our predicted molecules existed of course may be they were been in this may be just the computer garbage of it, they existed and they modulated GPCRs, this is- was an astonishing result and it's one of the results now is getting us on trade of pharmaceutical companies because they can see that we can do things that people would have guessed just was not doable.
Fintan Walton:
Right, so these, these peptides are once that mimic existing peptides or "
Martin Gerstel:
These are natural.
Fintan Walton:
So these are --
Martin Gerstel:
Peptide these are "
Fintan Walton:
You have found through computer installation.
Martin Gerstel:
The natural peptide Ligands eight of them and we are now in the process we are licensing them out to it.
Fintan Walton:
Okay.
Compugen's business model and competitive advantage
Martin Gerstel:
I should say a little bit about our business model. Our competitive advantage is solely at this level, all right? Very early stage.
Fintan Walton:
Right.
Martin Gerstel:
Identifying a product that and expert in the field would look at and say you know that's relatively early but that's exactly the type of product that I would like to put in through my pipeline.
Fintan Walton:
Right.
Martin Gerstel:
And we want to do that under royalty agreements that's our business model.
Fintan Walton:
Right. Okay. So you don't go beyond the synthesis of the peptides themselves?
Martin Gerstel:
I hope that we will not have to. This is the critical issue for us as a company right now because "discovery early stage discoveries are all over the place.
Fintan Walton:
Right.
Martin Gerstel:
Now the fact that we can create them with these platforms and if you -- the way I have described them what that means is, the next time -- and we've done this now twice with the GPCR peptide platform. The second time we were on the same platform we get better results, more results and better results because it's improving our models because of course the number of them operate.
Fintan Walton:
Yeah, more information coming in.
Martin Gerstel:
That improves the model and then you just over time you know get better-and-better and that's one out of ten, one out of ten platforms. So we can create all of these early stage compounds. We take them now to the industry and the business development people in general and industry are looking for product candidates, alright? So many people have suggested that they had platforms in the past and they all failed that the industry is very low on the hands of the platforms any more, they want to see products, they want to know they work. So when they come to us they see our product in such an early stage and they say look this looks great. Take it the next step and if it works we will license it in a minute, all right?
Fintan Walton:
That's the classic response.
Martin Gerstel:
But the problem is we have something that no one else has. We have 100's of these early stage so which ones do we take to the next stage.
Collaborations,Partnerships and Future strategies
Fintan Walton:
So who are you partnering with then?
Martin Gerstel:
Well, we are only beginning now to partner with some of the companies. It took us ten years. I should mention, it took us ten years and we've published 50, 60, 70 peer reviewed scientific papers of our aspects of our work. So this is world class science that's going on there, that --
Fintan Walton:
How much money has been pumped into this?
Martin Gerstel:
Well over a 100 million has been spent and that we've created this capability it's quite astonishing our market Cap today is about $20 million. You know, it's interesting to think about that from the standpoint that in the beginning I wanted to do both human and agriculture because we are talking about life, more than you know life it's not just humans it's everything that's living. So we started a program internally also with agriculture to use exactly the same concepts of predicting whatever. Along the way for financial reasons we decided we have to spin it out. So we spun it out it's called Evogene [PharmaDeals ID = 11604] as a separate company and it's just continued to do " to do it's research. This is back in 2002 I think we spun " we've spun it out. They had more trouble then we did in trying to get the attention of the bio companies because they have been so burned by spending billion of dollars for these platforms in the past that have never worked, all right?
Fintan Walton:
Eventually a negative period?
Martin Gerstel:
Exactly, so we couldn't even get in the door. But then there is a tremendous advantage in Agbio over what we were doing because when they if you remember I said we have reached a stage where we end up with 10, 20, 30, 40 possible candidates and then you have to go to the lab to see if they work and if it's a therapeutic it " each one of them you have to all kinds of studies may be 5, 6 years later you will be in the clinic with the first one. In agriculture when you have those 30 to 40 you immediately throw them into the plants, you don't care if half the plants die, you are not so, you don't worry about them methodical sampling, all right, So you immediately get final proof positive that, that "
Fintan Walton:
Sure.
Martin Gerstel:
Not only are the predictions good in, but so -- we showed that about a year ago, within a month we had a half of dozen of major bad bio companies sending teams of people to Israel to review this and the company just signed an agreement with Monsanto on four major crops for two what they you know two in two traits, what they call traits in the Agbio business with about a $100 million of committed funding including $30 million of equity at a price which is seven times, is seven times Compugen's current market value and clearly the world of " what we are doing in human therapeutics will dwarf what they are doing.
Fintan Walton:
Right.
Martin Gerstel:
But we cannot go from our computer predications and throw our drugs " our predicted drugs into humans and even if they could, I wouldn't. most of them are going to work.
Fintan Walton:
Yeah.
Martin Gerstel:
Alright?
Fintan Walton:
So you need to collaborate them with the pharmaceutical industry and --
Martin Gerstel:
Well, yes and that's.
Fintan Walton:
And you started to set up your own agreements, you got deals with Merck?
Martin Gerstel:
Well we've done with Merck, [PharmaDeals ID = 29391] with Roche, [PharmaDeals ID = 29331] with Merck Serono [PharmaDeals ID = 31756].
Fintan Walton:
Yeah.
Martin Gerstel:
A number of the biomarker companies because keep in mind our discoveries isn't just therapeutics it is also biomarkers and we have with four of the leading 10 biomarker companies.
Compugen's future business and relationship strategies.
Fintan Walton:
So how do you place now Compugen and bring it into the future? You've built this capability, you are on out producing tangible potential candidates that "
Martin Gerstel:
But that's the best way to say. We are presenting -- I mean what we are doing is you know giving some " giving some pharmaceutical company a 20 lottery tickets that we have through our analysis have said you know in there are three winning tickets, we don't know which ones they are, alright? The problem we've run into is that the typically pharma companies don't want to bring in 20 candidates to go through early stages but they want " is they want to see one that is going to next stage.
Fintan Walton:
Yeah.
Martin Gerstel:
And so our, most of our business development to-date has been to try and is to license out these individual " individual discoveries that we've made.
Fintan Walton:
Right.
Martin Gerstel:
We now have decided that both for financial reasons, business strategic reasons it really doesn't make any sense. It doesn't capitalize on the discovery capability we have. So we have now be going to approach the major pharmaceutical companies to see if we can enter into very substantial strategic relationships where we " they see us as the feeder for whatever discoveries they want.
Fintan Walton:
How do you define that? they have been different therapeutic areas that you do it focus? they could buy up basically or get access, exclusive access to particular therapeutic areas?
Martin Gerstel:
Mostly now we are just beginning this process, most likely it will relate to " what I call these platforms because each platform and we have a platform for therapeutic, proteins one for GPCR Ligands, I mentioned one for targets for monoclonal antibodies which is a very, very hot field and we are partners there with Medarex [PharmaDeals ID = 26359] and some of ours, we are looking, looking little further, so --
Fintan Walton:
Are you going to set them up as joint ventures? Strategic?
Martin Gerstel:
We are at a very, very early stage. my guess we are going to take the next, I would say three to four months to talk with a number of companies. We've started that process and through that process I think we will decide that we were going to work with one company, two companies. My guess right now is that the best would be the one to two companies in therapeutics, One to two companies in diagnostics and you know, it's that the " given what we have created to-date that their capabilities, the platforms, the tools we now every time we approach a new problem it's easier than before. When we decided to go after GPCR Ligands the day we started the program was 80% finished.
Fintan Walton:
Right.
Martin Gerstel:
Because we relied on all of these pieces that have been done over the years.
Fintan Walton:
Sure.
Martin Gerstel:
So as a company from a financial standpoint it takes us -- we only spend about $10 million a year to do this work which is a drop in the bucket in the pharmaceutical industry and the, this you know and we've you know like this GPCR program you are talking about a million dollars from start to having 8 ligands in hand.
Fintan Walton:
Right.
Martin Gerstel:
It's
Fintan Walton:
Can I mean just want to try and rap the interview now.
Martin Gerstel:
Yes.
Prediction about Compugen over the next five-years.
Fintan Walton:
So last question you talked about predictions and particular with science and so forth. May be you could give me a prediction of what Compugen will be like in five years time?
Martin Gerstel:
Well it, to some degree it's easier to talk about what the in -- I mean the industry in five-years time the research aspects of the industry will be largely what we have today. I mean the experimental ways, everyone knows that show these curves of you know increasing expenses were as decreasing returns.
Fintan Walton:
Yes opportunity.
Martin Gerstel:
No one ever says why that is, that's because their finding things experimentally and you can use high throughput and everything but sooner or later you hit a brick wall you get the easy things first and then it's get harder and harder. With what we are doing, once you actually have proven that it works its get easier and easier because now the scientific method is moving up and that's known every pharma company knows that this is the future. Question is how do they make these transitions and win. So the problem for Compugen is we clearly are leading the way, I am not gonna have any doubt that we are the top the best in the world in what we are doing right now across the board alright In all these different fields. The issue for us is how to " how to capitalize on that for the benefit of our current shareholders. Our stuff cannot go away, alright? What we are doing is science which people talk about that you know we have technology but we have no technology in our company.
Fintan Walton:
Right.
Martin Gerstel:
We have science and understanding of the science out of which we can create these predications this is not going to go away. If we were called unsuccessful we may end up being bought by big some pharmaceutical company. If we are successful I think that within 5, within -- in that five-year time frame we could be leading the world in " in the number of our product candidates that are in the pipelines of other companies being developed at their expense where we will have share in their revenues and that's our goal as a corporation.
Fintan Walton:
Martin Gerstel, thank you very much indeed for coming on the show.
Martin Gerstel:
Thank you very much. Appreciate the opportunity.
Martin Gerstel
Chairman
Martin Gerstel Pissarrahas been chairman ofCompugen since 1997. In addition, he serves as chairman of two Compugen affiliated companies, Evogene Ltd and Keddem Bioscience and is a co-founder and co-chairman of Itamar Medical (an Israeli based medical device company). Prior to relocating to Israel in 1994, Mr. Gerstel was co-chairman and CEO of ALZA Corporation, which he helped found in 1968. He currently serves as a director of Yissum Ltd, Yeda Ltd and the US Foundation for the National Medals of Science and Technology.MrGerstel is a member of the Board of Governors and Executive Committee of the Weizmann Institute of Science and the Board of Governors of the Hebrew University. He is also an advisor to the Burrill Life Sciences Fund and the board of the Israel-U.S. Bi-national Industrial Research and Development (BIRD) Foundation. Mr.Gerstel holds a B.S. from Yale Universityand an MBA from Stanford University.
Compugen
Compugen's mission is to be the world leader in the discovery and licensing of product candidates to the drug and diagnostic industries under milestone and revenue sharing agreements. The Company's increasing inventory of powerful and proprietary discovery platforms is enabling the predictive discovery " field after field " of numerous therapeutic and diagnostic product candidates. These discovery platforms are based on the Company's decade-long focus on the predictive understanding of important biological phenomena at the molecular level. Compugen's current collaborations include Biosite, Medarex, Inc.,,Merck & Co., Inc.,, Ortho-Clinical Diagnostics a Johnson & Johnson company,Roche,Siemens Healthcare Diagnostics, Inc., and Teva Pharmaceutical Industries.In 2002,Compugen established an affiliate Evogene Ltd . (TASE: EVGN.TA) to utilize the Company's in silico predictive discovery capabilities in the agricultural biotechnology field. Compugen has been publicly traded on Nasdaq (NASDAQ: CGEN) since August 2000 and on the Tel Aviv Stock Exchange since January 2002.