By John Russell
Feb. 14, 2008 | Strand Life Sciences is hardly invisible, but its "Intel-inside" strategy means that many users don't know Strand is the engine inside their informatics tools. Even a near-fatal terrorist attack on Strand Chairman and CEO Vijay Chandru didn't produce the typical coverage that suddenly raises one's profile. He and three others were wounded and another speaker was killed in an attack during a science conference in India. The attack, by a lone gunman wielding an AK-47, occurred on December 28, 2005, when much of the world was in holiday mode and not paying a lot of attention. (For two accounts of the attack, click here and here.)
Chandru missed nearly a year of work, but the company he co-founded labored on. He eventually resumed his responsibilities, which today call for Chandru to be in the U.S. a fair amount of the time. If you use GeneSpring or PathwayArchitect, you are really using embedded Strand technology. Its strategy has been to develop solid technology and find partners to help bring that technology to market. Microarray tools were first, followed by pathway tools, then QSAR tools, and recently it's begun a solutions business that encompasses customized tools and research collaborations. Strand even has a system biology effort and has developed a liver model that's being evaluated by FDA and a few pharmas.
Chandru and co-founder Ramesh Harihanan are the key forces behind Strand, based in Bangalore, India. Chandru is an MIT-educated (Ph.D.) computer scientist with impressive credentials. He's a fellow of the Indian Academy of Sciences, has taught at Purdue, University of Pennsylvania, and the Indian Institute of Science. He recruited a talented team for Strand, including other prominent Indian researchers, and has successfully lured folks, for example, from Merck and Entelos. And yes, like most biotech upstarts, Strand wants to eventually get directly into the drug and biomarker business. That's where the big money is. While many biotechs have struggled, Strand's trajectory has seemed to be consistently up. SBNL editor John Russell talked with Chandru about Strand's journey and plans.
SBNL: Thanks for your time Vijay. Can you tell us a bit of Strand's beginning?
Chandru: I didn't do life sciences at first. I was a straight computer science [guy] and went back to India in the early '90s. I was at the Indian Institute of Science, which is probably one of the best research institutes for biophysics. G.N. Ramachandran, was there, the guy who did what are called Ramachandran plots (used to describe protein structure). So there was kind of a gradual learning of the interface in computational biology that happened. By the late '90s, our lab at the university was doing some stuff in computational biology. I was writing papers on protein folding and stuff like that.
SBNL: Did you ever participate in CASP -- the bi-annual competition to computationally solve protein structures from only their amino acid sequence?
Chandru: No. I wrote more theoretical papers on the Levinthal paradox and computational complexity and stuff like that. I was more of a logician than a software person. And then we spun out Strand as a small consulting outfit from the university. So in 2000/2001, I think biotech actually had money, and we were getting work from various companies to do essentially classification problems that were kind of tough -- problems in image analysis and biomarkers [though] it was not called that then.
We did some of that, we got paid well. We were actually a profitable little company. We had about 20 people. Then we decided that we'd build differentiating technology and then try to build products and solutions on top [of the technology platform], so we raised venture capital.
I think we were probably one of the few companies to get early-stage VC funding in bioinformatics, and Sequoia funded us. They had an India fund, and they wanted to try to do in India what had been very successfully done in Israel. So we were identified, and we raised our Series A by mid-2002, and then set about building the platform.
The platform we called avadis (access, visualize, analyze, and discover). Basically what we found was that Spotfire-like platforms had great applicability, but Spotfire was missing the analytics. We decided to do both visualization and analytics within the same framework, and by 2004 the platform was ready and we started spinning out products from it. Ramesh [Harihanan, Strand co-founder] was the chief architect of avadis. He took his PhD from Courant Institute, NY, and was on the TR-100 list in Tech Review 2002 - perhaps the only scientist in India to make the TR short list.
Anyway, Charles Cantor (CSO, Sequenom), Charles DeLisi (Boston University), Eric Eastman (former CTO, Gene Logic), and several other people were involved and on our Scientific Advisory Board. They suggested we look at array technologies. So we started building tools for microarray products, and in 2004 we entered the Asia Pacific with avadis for microarrays, and we got about 20 percent of the market.
SBNL: What was 20 percent of the Asian market back then?
Chandru: It was about 200-odd seats and we had maybe a half a million in revenues. We were sold heavily in Japan by Hitachi at that point.
SBNL: Isn't it hard to sell in Japan unless you have a local partner?
Chandru: Right. And Affymetrix liked us. So they would take us to user group meetings. They kind of identified us as the Asia Pacific provider for analytics, and we went along with Affy doing training programs all over China and Japan, and so on. We actually used Affy distributors in many of the Asia factories.
SBNL: How helpful was that?
Chandru: Very helpful. So that was our first product.
SBNL: Was price a big part of the attraction?
Chandru: Not really. We were about comparable to GeneSpring, and it was our big competition. We were considered scientifically as good as GeneSpring, but they, of course, had some years on us because of being in the market earlier.
SBNL: Did they also have more life science expertise?
Chandru: Well, we had a fair amount of life science by then. Now we have about 35 life scientists in the company and about 60 computer scientists. I think by the end of 2004, we were starting to be noticed. There was this Japanese investor who came and took a stake in the company, a group called MediBIC. They're a small consulting group in Japan. They started representing us in Japan. And suddenly we got a call from Nick Roelofs at Stratagene saying, "We have this product, ArrayAssist, and it's been in the market for a while, but we've seen your product and think we should retire ours and take yours on." It took about a year or two to maneuver all of that, but avadis became ArrayAssist.
SBNL: Is it still sold as ArrayAssist?
Chandru: Yeah. It's sold as ArrayAssist and it says "powered by avadis." We have the "Intel inside" kind of [strategy.]. It was a good relationship, because Stratagene really understood what the research biologist wants. In many ways we had over-engineered, we hadn't understood the usability issues as well as they had, so it was a good partnership.
SBNL: The collaboration improved the product?
Chandru: Absolutely. Stratagene had also built a good relationship with [Affymetrix], so we were the first in the market with exon tools, we were first in the market with copy number. We were doing well but Stratagene certainly has much more of a focus on the academic market segment than pharma.
SBNL: Is that where Strand was focused as well?
Chandru: Yes. The next event is almost a repeat of this (Stratagene experience). Nick Roelofs is at Agilent [now], and so he thinks of Strand again. Agilent acquired Stratagene (June 2007), so for the last six months, we've been moving GeneSpring over onto avadis.
SBNL: Will the brand name is retained but the engine be avadis?
Chandru: Yes. GeneSpring has a large following, so it made complete sense to keep that brand name. What you'll see in the market with GeneSpring 9.0, which will come out at the end of the year will be avadis. (GeneSpring 9.0 officially launched January 15, 2008).
SBNL: So Strand is the embedded system supplier of the array software.
Chandru: It's a good story. I think we've sort of risen to the top, and that's taken a little while. And it's joint ownership of the product (with Agilent), it's a revenue share. It's not fee for service.
SBNL: You said that Stand had 20 percent of the market back when you first went into Asia. What do the sales figures look like today? How big is your market?
Chandru: The combined user base for all the "avadis inside" products will jump from between 1000 and 2000 users in 2007 to a 7000-plus user base in 2008.
SBNL: Your user base suddenly just jumped and with every revision that comes out of GeneSpring now it's like getting a new customer.
Chandru: Yes. We also provide first- and second-line support, so we'll be talking to the customers. We've set up that infrastructure.
SBNL: Is it a 50/50 split on the revenue sharing?
Chandru: I'm not really at liberty to discuss that. But it's put us in a very nice position, let me just say that.
SBNL: I thought you said you were profitable all along?
Chandru: No, no. Once we went into the platform development, we were not. That's where the VC funding came in. In 2004, we started earning again, and in 2005 we were breakeven. In 2006 we dropped a little in the sense we invested a lot more in people and development, because we also got into chemistry, and that's the other story. In 2007, we are breaking even, hopefully will be profitable.
SBNL: So far all the products are in life science.
Chandru: We're completely life science-focused.
SBNL: So the chemistry story?
Chandru: Yes, the chemistry story. So to back up now, in 2003/2004 we had the platform, [and] we started using the platform internally to build models. And in chemistry we find that there's an opening for good QSAR type models, so we started building models and in 2004, our QSAR Admet models were, by arrangement with BioRad, bundled and sold as part of the KnowItAll cheminformatics platform.
SBNL: You used the same strategy to find somebody who's already in the market?
Chandru: Absolutely. Our advantage, of course, is the scientific and technical talent in India. Our disadvantage is the distance to market. So we've been working always with partners who give us that visibility into the market.
SBNL: Who would you bump up against in the chemistry market?
Chandru: We were bumping up against ADMEnsa. We just recently launched our products. It's one thing to sell models, but we've also now built a platform for modeling and also deploying models into medicine and chemistry. We call this platform Sarchitect, and it's an SAR modeling platform, and we've been successful in getting it into pharma. So it's still early days with the chemistry product, but Gilead is now our first customer.
We have an innovative approach of providing open access to the models within Sarchitect Designer to permit the computational chemist to examine the training and validation sets, modify them and get a sense of the chemical space that the models are localized in. We have generally adopted an "open source" approach to empower computational chemists and the community website -- www.qsarworld.com -- that we created has a Google rank of 3 on "qsar" and has attracted over 8000 visitors, many of whom find the curated datasets on the website a boon for testing modeling techniques.
SBNL: It seems like Strand has a full plate. What's next?
Chandru: Early in 2007 we also started up what we call a solutions practice, so we hired a senior vice president out of Merck. He's a senior vice president for us. At Merck, he was like a team project leader in the IT. So he spent 10 years in Boston working at Cereon and then Biogen and then Merck. His name is Sai Subramaniam and he's moved back to India, heading up this unit. So the idea there is we have all these products.
Oh, by the way, there's one little twist to the Stratagene story. Along with the ArrayAssist, they asked us to go with them on PathwayArchitect. We have a full, natural language processing engine for generating protein interaction databases and the full visualization, and so on and they wanted to use it.
SBNL: That's available now?
Chandru: That's available now. PathwayArchitect is on the market. I think the user base is about 500. So that goes into the stable now with Agilent as well. So we have Pathway [tools], we have the array, and then the chemistry stuff, which is just getting into the market. On the chemistry side, we'll build the solutions to pitch to pharma, but we are on the lookout for a channel to also sell the chemistry product and we've been having conversations with the obvious channel partners, and we'll look for a similar relationship. So the business model for Strand now is starting to get a little more complex. We have products we do the development and support and application science -- but we work with partners for sales and marketing.
SBNL: But the solution group is a little different?
Chandru: Yes, solutions is different. It's just starting out, but we're looking at value-added solutions, which are value-add to the product suite to begin with.
SBNL: So if I wanted to have a version of PathwayArchitect customized for a particular therapeutic area, I'd go to the solutions group?
Chandru: Exactly. The other aspect of our solutions [group] is it now also describes the third business unit within the company, which is our R&D group. We pulled out the head of immunology at Entelos and brought him back to India. He is Kas Subramanian.
Kas came back to India and has built a systems biology team, and we've built a model of the human liver.
SBNL: Aren't there several liver modeling efforts going on?
Chandru: There are a few announcements. I think we're a little further along. We started back in 2003, and were funded by The World Bank, actually. They gave us a million dollar grant to do this work. So we've been at it for some time now.
SBNL: What have you done to validate the model?
Chandru: At this point it's all based on public data, but the model is exhibiting homeostasis. I don't believe anybody else has a model that's reached that [point]. And the FDA is taking a look at it.
SBNL: Doesn't FDA also have a collaboration with Entelos to build a liver toxicity model? Wouldn't that be competitive?
Chandru: There's one issue with these models, which is they're basically based on disease models. And the fact is that the extreme stress you're putting on the animal models in that case and the gene expression profiles that you're going to get out [from them] are going to be so all over the place. [It's] not clear what value that model is going to have. At least the toxicology experts that we've spoken to are not too confident that that [approach] will actually lead to a good predictable approach.
You need a normal liver model to begin with, I think, to be able to then capture the effects [of compounds]. So we've shown it to a few key people, [such as] Steve Tannenbaum at MIT and a few others. A couple of major pharmas are very interested in it. The comments have been, "It's nice to see that you guys have actually been working on this for quite a long time and it's ready for prime time. So, we want to test this out and if we like what we see, then it would absolutely be a great tool for us."
SBNL: How would such a tool be deployed? Would it be provided as a service or would you sell the model?
Chandru: I think at this point we are looking for collaborative relationships with pharma. Whether we go with exclusive relationships or with a consortium approach, I think, isn't [clear].
SBNL: You could use pharma's experimental data to help validate the model.
Chandru: Actually the model is a hybrid model. You actually have to run some assays, determine some parameter value, and then plug those in. So you would need to also set up the assays. So that's an exciting part of our R&D. So when you talk about solutions, you are also talking about collaborative contract research.
SBNL: Here's a question. Nearly everyone selling biotech technology hits a growth limit.
SBNL: You can't grow really big unless you're somehow own a piece of the drug or biomarker. Is it Strand's goal to eventually own a piece of the drug?
Chandru: Absolutely. Not drugs necessarily right away, but biomarkers for sure. We are already in collaborative relationships. We're working with hospitals with clinical trials in groups, and so forth.
SBNL: So there is a kind of progression for the company?
Chandru: Absolutely. We're not happy with the valuation. I think life science just requires a little more gestation time, a little more validation. If you really wanted to just make money, we should have been applying data mining to financial markets. We could very well have done with this same platform, but that's not [the strategy]. Spotfire, for example, has gone into business intelligence.
SBNL: That's one reason they're able to grow. They have a variety of industries they sell to.
Chandru: What we hope to do on the technology side is slowly transition from the research to have a role in translational medicine.
SBNL: How many people are in the company now?
Chandru: There are about 110, and 95 are scientists We have a small marketing team in India and business development here and tech support in San Diego The majority is in India in Bangalore.
SBNL: In the midst of building Strand, I understand you were caught up in a terrorist attack and severely injured. Is this something you're comfortable talking about?
Chandru: Well, it's part of my academic life. Although I'm an honorary professor now, I ran a conference in decision sciences at the Indian Institute of Science and was president of the Operational Research Society. We were at a conference in December of 2005, and a terrorist walked in and opened fire, and I was hit by several bullets, both arms and the chest, but luckily it missed the lungs.
SBNL: This was in India?
Chandru: Yes, in India, December of 2005, yeah. It happened just after Christmas. So it was carried in the International Herald Tribune and so on. But I think a lot of people didn't see it because it was the holidays at that time.
Yes, one professor died and there were several injuries. It's pretty amazing. So [it's] another reason to be committed to the healthcare community. They've taken good care of me. And so I was actually flown out to Mass General. And everything was refitted here. So they did bone regeneration.
SBNL: How long was your recovery?
Chandru: Technically I went back to the office after nine months, but I think it was about a year to really get back. So 2006, I was pretty much out of action. I was very lucky, but now I am back, and we're very optimistic about the future.