Cellworks Group

Cellworks Group INTRODUCTIONMartin: This time we are in  San Jose,  in Cellworks Group office with Taher. Taher, who are you and what do you do?Taher: My background, my academic background is engineering and business school. Before starting this company Cellworks, I was almost 15 years in the semiconductor business in the automation engineering business. The idea behind Cellworks was to take those concepts and best practices and apply them into a completely different domain and make impact that industry.Martin: I totally understand that you have automation and modeling background, why did you choose health care?Taher: What happened is, in the semiconductor space for example, we have reached a point where every Christmas we would come up with a new product. We knew in 6 weeks development cycle. And every time we sent a design for fabrication to  Taiwan,  almost 99% certain that youll get working silicon.When we saw life sciences, everything was different. It took more than 10-12 years to get a produc t out. Less than 8% chance of success and the development cost was almost 10 times larger than in the semiconductor business.So from all the business vectors and so forth, it seemed like a very opportune industry in order to apply what we have successfully deployed in the chip business to the life sciences.Martin: So your basic hypothesis was, we, with our technology can minimize or lower the deployment cost in the health care industry?Taher: Yes, if I could increase the success rate, because the biggest challenge was the failures. So either you improve your success rate or if you have a bad product or bad architecture, you fail early. So our whole philosophy was that effectually through simulation and modeling, you can actually improve your probability of success in an industry where the success rate was extremely low.Martin: Great!BUSINESS MODELMartin: Taher, lets talk briefly about the business model. How does it currently work?Taher: Well, the core competency in the organization is the technology, which is the simulation technology. So every aspect of it is governed based on that. So we had few multiple choices in terms of business models.One was effectually being a service provider, and we early on decided that service is not something which we wanted to be in, as a services company.The second option was being a technology provider, being an enabler to the pharmaceutical industry. We decided to follow a business model where we would be our own consumers of the technology after we intensively validated and designed products using the technology which we can monetized by engaging with different players in the ecosystem.Martin: And now your first product is based on oncology. Can you tell us a little bit more about this, how this works and how is it applied in the typical work process of your client?Taher: Right, so cancer is one of our main focus areas. One of the main driving factors for selecting cancer is because of the whole moment of personalization of treatments. It so happens that in cancers, every patients genomic signature is completely different. And in that context, one size fits all approach does not work.So what traditionally used to be a block buster business model, where you take a drug product and deploy it to a large consumer base, that model falls apart based on personalization of medicine. And almost you start having a long tail business model.So in the context of cancer, the patients genomic signature is what we use to create a simulation of the cancer patient. And effectively then designed a treatment, which is personalized to that particular patient, and effectively either test it using the patients cancer cells outside the patient in the lab, or find a mechanism by which to take it to clinical transition pathway.Martin: Okay. I totally understand that once you have the date set, then you can make the predictions on what kind of treatment to use. But in the first place, how do you get this data set that you can r eally use to make a sensible prediction?Taher: So the core technology is simulation technology. Its effectively built based on, so all the last many decades there have been, in the science field, there has been tons and tons of publications, which effectively talk about this specific connection of the different biological players.So, the simulation technology is analogous to maps. You can think of it as Google Maps with the difference. So, it actually builds in a connection by connection that guides the pathways inside a cancer cell. With the added difference that you can actually simulate it, you can actually perturb any pathway, any  nerves  or any combinations of them, and see the effect of it propagating down the network.So essentially its been built based on information published out there, and aggregated together manually in a simulation model. So you can actually do what if analysis and run studies and experiments, to predict what would be the effect of this perturbation, eit her drug perturbations or other perturbations on the cancer end points.Martin: Is this the only data that is flowing in or did you also develop a machine learning algorithm that learns over time?Taher: We have induced machine learning. Effectually what happens in science that there are a lot of contradictory data sets. To avoid basically the phenomena of garbage in and garbage out, weve effectively used manual mechanism to aggregate the data and all kinds of automation is applied thereafter.So when youre building the model, we manually look at each and every research paper, the experiment conditions and all the parameters which determine the verification of the data, and then thereafter all the steps are automated.ONCOLOGY ECOSYSTEM GO-TO-MARKET STRATEGYMartin: Can you tell us a little bit more about the oncology ecosystem and what is your go-to market strategy?Taher: Well, we have 2 ways of approaching it.One is in the process of modeling different patients and different segments of patient population. We are coming up with our own  predict programs, which we are patenting and validating. So we are in-house building a library of patterns of  predict  programs, targeting a very specific patient segments. So that’s for medium to long term strategy, in terms of collecting a library of patterns.More for a short term to medium term go-to market strategy is, were looking at how this personalization can be deployed as a decision making algorithm in the clinical practice. So in that capacity, weve started talking to clinical groups who effectively are collecting patients tissue samples and getting it genomically profiled. And working with them, and collaborating with them, how we can actually build the next layer of personalization, where they can actually translate that information into actionable insight.Martin: Okay. Taher, can you tell us a little bit about the players in this oncology market?Taher: In general, in pharmaceutical business and also of course in oncology business, there are multiple players in the ecosystem. This is probably one of the industries where typically examples of multi-sided business model. This is truly a multi-sided.In this business, the clinicians are the final arbitrators in terms of making decision, in terms of what treatment interventions. The consumer or the patient effectively take the direction from the physicians and does not pay for the services. So the group which pays for it is the insurance companies and the payers. And the suppliers into the ecosystem are the pharmaceutical companies, the diagnostic, genomics companies.The whole industry is then regulated by FDA for example, and other regulatory body. So you have a multi-factorial ecosystem.In terms of the strategy for launching this go to market strategy, we have to start working with the clinical groups in terms of building validation, credibility and getting them on board. Once that process is ongoing, then you effectually go to patient advocacy groups as well as working with different members, like the pharmaceutical and diagnostics. You indirectly or directly influence the payers to start paying for this technology and solutions.Martin: Okay. Great!MARKET DEVELOPMENTMartin: Lets talk briefly about the market development, especially in oncology, for example. I mean this is quite a big industry in the health care industry. Can you tell us a little bit more about the technology drivers over there and maybe some trends you have identified over the last years?Taher: I think the personalization of cancer treatment has become mainstream. This is something which is widely accepted and deployed in all major, major clinical centers globally, have the initiative, in terms of being able to find the right treatment for the right patient. I think that is the main focus.The idea behind the simulation is the fact, there are tons and tons of big data which is being generated as a consequence of the next sequencing for example, which comp anies are deploying. So the cost of sequencing, which used to be in thousands and hundred of thousands, has dropped down to less than a thousand dollars now.The ability to sequence this cancer tissues, be able to generate this big amount of data in gigabytes, is effectively creating a situation where you have the data but you dont have an actionable pathway to intervene or influence the technical decisions. Thats where the predictive technology fits in beautifully, in terms of being able to take advantage of the move to big data, creating a patient specific of  thought,  and then through the modeling approach coming up with a scientific rational and a treatment, which is actionable and something which a clinical group can actually utilize for intervention.Martin: Great!ADVICE TO ENTREPRENEURS In San Jose (CA), we meet the founder and CEO of Cellworks Group, Taher Abbasi. Taher shares his story how he came up with the idea and founded his company, how the oncology market is currently working, as well as he provides some advice for young entrepreneurs.The transcript of the interview is provided below.INTRODUCTIONMartin: This time we are in  San Jose,  in Cellworks Group office with Taher. Taher, who are you and what do you do?Taher: My background, my academic background is engineering and business school. Before starting this company Cellworks, I was almost 15 years in the semiconductor business in the automation engineering business. The idea behind Cellworks was to take those concepts and best practices and apply them into a completely different domain and make impact that industry.Martin: I totally understand that you have automation and modeling background, why did you choose health care?Taher: What happened is, in the semiconductor space for example, we have reached a point where every Christmas we would come up with a new product. We knew in 6 weeks development cycle. And every time we sent a design for fabrication to  Taiwan,  almost 99% certain that youll get working silicon.When we saw life sciences, everything was different. It took more than 10-12 years to get a product out. Less than 8% chance of success and the development cost was almost 10 times larger than in the semiconductor business.So from all the business vectors and so forth, it seemed like a very opportune industry in order to apply what we have successfully deployed in the chip business to the life sciences.Martin: So your basic hypothesis was, we, with our technology can minimize or lower the deployment cost in the health care industry?Taher: Yes, if I could increase the success rate, because the biggest challenge was the failures. So either you improve your success rate or if you have a bad product or bad architecture, you fail early. So our whole philosophy was that effectua lly through simulation and modeling, you can actually improve your probability of success in an industry where the success rate was extremely low.Martin: Great!BUSINESS MODELMartin: Taher, lets talk briefly about the business model. How does it currently work?Taher: Well, the core competency in the organization is the technology, which is the simulation technology. So every aspect of it is governed based on that. So we had few multiple choices in terms of business models.One was effectually being a service provider, and we early on decided that service is not something which we wanted to be in, as a services company.The second option was being a technology provider, being an enabler to the pharmaceutical industry. We decided to follow a business model where we would be our own consumers of the technology after we intensively validated and designed products using the technology which we can monetized by engaging with different players in the ecosystem.Martin: And now your first produ ct is based on oncology. Can you tell us a little bit more about this, how this works and how is it applied in the typical work process of your client?Taher: Right, so cancer is one of our main focus areas. One of the main driving factors for selecting cancer is because of the whole moment of personalization of treatments. It so happens that in cancers, every patients genomic signature is completely different. And in that context, one size fits all approach does not work.So what traditionally used to be a block buster business model, where you take a drug product and deploy it to a large consumer base, that model falls apart based on personalization of medicine. And almost you start having a long tail business model.So in the context of cancer, the patients genomic signature is what we use to create a simulation of the cancer patient. And effectively then designed a treatment, which is personalized to that particular patient, and effectively either test it using the patients cancer cells outside the patient in the lab, or find a mechanism by which to take it to clinical transition pathway.Martin: Okay. I totally understand that once you have the date set, then you can make the predictions on what kind of treatment to use. But in the first place, how do you get this data set that you can really use to make a sensible prediction?Taher: So the core technology is simulation technology. Its effectively built based on, so all the last many decades there have been, in the science field, there has been tons and tons of publications, which effectively talk about this specific connection of the different biological players.So, the simulation technology is analogous to maps. You can think of it as Google Maps with the difference. So, it actually builds in a connection by connection that guides the pathways inside a cancer cell. With the added difference that you can actually simulate it, you can actually perturb any pathway, any  nerves  or any combinations of them, and see the effect of it propagating down the network.So essentially its been built based on information published out there, and aggregated together manually in a simulation model. So you can actually do what if analysis and run studies and experiments, to predict what would be the effect of this perturbation, either drug perturbations or other perturbations on the cancer end points.Martin: Is this the only data that is flowing in or did you also develop a machine learning algorithm that learns over time?Taher: We have induced machine learning. Effectually what happens in science that there are a lot of contradictory data sets. To avoid basically the phenomena of garbage in and garbage out, weve effectively used manual mechanism to aggregate the data and all kinds of automation is applied thereafter.So when youre building the model, we manually look at each and every research paper, the experiment conditions and all the parameters which determine the verification of the data, and then thereafter all the steps are automated.ONCOLOGY ECOSYSTEM GO-TO-MARKET STRATEGYMartin: Can you tell us a little bit more about the oncology ecosystem and what is your go-to market strategy?Taher: Well, we have 2 ways of approaching it.One is in the process of modeling different patients and different segments of patient population. We are coming up with our own  predict programs, which we are patenting and validating. So we are in-house building a library of patterns of  predict  programs, targeting a very specific patient segments. So that’s for medium to long term strategy, in terms of collecting a library of patterns.More for a short term to medium term go-to market strategy is, were looking at how this personalization can be deployed as a decision making algorithm in the clinical practice. So in that capacity, weve started talking to clinical groups who effectively are collecting patients tissue samples and getting it genomically profiled. And working with them, and collabora ting with them, how we can actually build the next layer of personalization, where they can actually translate that information into actionable insight.Martin: Okay. Taher, can you tell us a little bit about the players in this oncology market?Taher: In general, in pharmaceutical business and also of course in oncology business, there are multiple players in the ecosystem. This is probably one of the industries where typically examples of multi-sided business model. This is truly a multi-sided.In this business, the clinicians are the final arbitrators in terms of making decision, in terms of what treatment interventions. The consumer or the patient effectively take the direction from the physicians and does not pay for the services. So the group which pays for it is the insurance companies and the payers. And the suppliers into the ecosystem are the pharmaceutical companies, the diagnostic, genomics companies.The whole industry is then regulated by FDA for example, and other regulat ory body. So you have a multi-factorial ecosystem.In terms of the strategy for launching this go to market strategy, we have to start working with the clinical groups in terms of building validation, credibility and getting them on board. Once that process is ongoing, then you effectually go to patient advocacy groups as well as working with different members, like the pharmaceutical and diagnostics. You indirectly or directly influence the payers to start paying for this technology and solutions.Martin: Okay. Great!MARKET DEVELOPMENTMartin: Lets talk briefly about the market development, especially in oncology, for example. I mean this is quite a big industry in the health care industry. Can you tell us a little bit more about the technology drivers over there and maybe some trends you have identified over the last years?Taher: I think the personalization of cancer treatment has become mainstream. This is something which is widely accepted and deployed in all major, major clinical centers globally, have the initiative, in terms of being able to find the right treatment for the right patient. I think that is the main focus.The idea behind the simulation is the fact, there are tons and tons of big data which is being generated as a consequence of the next sequencing for example, which companies are deploying. So the cost of sequencing, which used to be in thousands and hundred of thousands, has dropped down to less than a thousand dollars now.The ability to sequence this cancer tissues, be able to generate this big amount of data in gigabytes, is effectively creating a situation where you have the data but you dont have an actionable pathway to intervene or influence the technical decisions. Thats where the predictive technology fits in beautifully, in terms of being able to take advantage of the move to big data, creating a patient specific of  thought,  and then through the modeling approach coming up with a scientific rational and a treatment, which is actio nable and something which a clinical group can actually utilize for intervention.Martin: Great!ADVICE TO ENTREPRENEURSMartin: Taher, we always try to help our first time entrepreneurs learn from great entrepreneurs like you. What advice can you give them, especially related when somebody comes to you and says, I would like to start a health care company. What would be your advice?Taher: I think the biggest thing to watch in a health care business is effectively, if youre going to be regulated, so for example the regulatory aspects are key considerations in the health care business. So if youre doing something which is beyond that, for example if youre coming with a health care application on mobile phones. So those things would be less regulated and the ability for you to go to market is going to be easier and faster.If the interest is to be able to get to market in a couple of years, you need to watch on the regulatory side of things, what are the implications are.Beyond that I thi nk which is true for health care or any business right now, pretty much you dont have all the answers when you started a venture. Looking for all the answers on day 1, I never have seen that happened in my personal lifetime. In terms of knowing all the answers. In that sense, you have to take the leap of faith out there.The third and final point which is very critical for most ventures to start is the financing aspect. If you are, if the business require some kind of external financing down the road, if theres a strategy to build a prototype through some kind of a bootstrap method or some other technique, that would be ideal because the prototype with the idea is a good inflection point for the external financing.Martin: Great! In terms of health care, isnt it quite expensive to bootstrap? Lets say if I would like to try to develop a cancer treatment?Taher: Absolutely. I think not only its expensive and difficult to develop the treatment, also the aspect of taking it through the cli nical study is prohibitively expensive. So, I  guess if were looking at large sciences, you may have to look at those peripheral areas. If intervention is 1 aspect of a treatment, 1 aspect of it, it could be in areas for example, electronic health records, different kinds of bioinformatics solutions and other kinds of technology solutions which sort of sit on the periphery.I guess, the closer you move towards intervening and impacting the treatment, the higher the bar, in terms of regulatory as well as the development process.Martin: Taher, what would you advice somebody who is currently working as an employee and just thinking about should I start a startup or not. What would you advice him?Taher: It seems like, given the recent and all trends and the excitement towards new ventures, it’s a good time, at least right now, for the last couple of years it has been a very good time for starting. If you have the right idea.The main thing, in many of these starting a new venture, you c ant expect to get all the ideas. In fact almost always, the business model of the plan which started out with and the plan which you actually end up executing, turns out to be quite different. So given that aspect, when the right idea is there and the ability for you to sort of, and you think that you have the right team to execute on it, its a good time to start a new venture.Martin: Okay. Great! Taher, thank you very much for your time. Cellworks Group is a very good example of how technology can help all of us become healthy and maybe have less cancer. So next time you think about starting your own company, think about how technology can influence your industry.