OAKLAND, Calif.--BioInform recently spoke with John Couch, the former Hewlett-Packard and Apple Computer executive recently named CEO of bioinformatics software provider Pangea Systems here. Couch replaced Joel Bellenson, a founder of the company, who is now serving as director of Pangea's scientific advisory board and as its chief strategist. Prior to his appointment as CEO, Couch had been working with Pangea for a few months as an executive-at-large for Mayfield Group, the venture capital firm that led the company's financing. The article will conclude in the November 24 issue of BioInform.
BioInform: What special skills do you bring to Pangea?
Couch: My background is in computer science. My degrees are from UC Berkeley in 1969, masters degree in 1971, and I did about a year and a half of PhD work. The industry was so hot that I then went to Hewlett-Packard, where I worked on the HP 3000 and eventually ran the software for that di vi sion; worked for Ed McCracken, who's at Silicon Graphics now. Then I went to Apple, where I was director of new products and vice-president of software, and then v-p and general manager of the personal office systems division, which did the initial graphical user interface.
With my computer science background I really look at bioinformatics as the application of advanced computer science to biological research. At HP I witnessed the hand-held calculator introduction. I remember the research people coming in and telling Bill Hewlett not to build the HP 35, because nobody wanted it. He said, I want it and furthermore I want it to fit in my shirt pocket. I realized then and there that when you're creating a market, sometimes a lot of market research doesn't really do much for you.
When I went to Apple, as the fiftieth employee, we had the challenge to create a new market for the personal computer, because at that time most people didn't believe in it. In fact, Hewlett-Packard asked if I would return once the Apple "game machine" failed. I learned a lot in that process. We grew from zero to a billion dollars in sales in five years by creating a new market; a paradigm shift took place from the time-shared world to the desktop. So when I was working with Mayfield as an executive-in-residence, and they asked me to take a look at a number of their companies and one of them was Pangea, I started to see very similar things. Everybody talks about the industrial revolution that has taken place given the human genome project, and the need for computer science, particularly in the integration of a lot of different types of data that are spread all over the world in different formats and in different spreadsheets or flat files or lab notebooks or whatever, and I could see some of the same things happening all over again that I saw in the early Apple days.
There's an opportunity to define and create a bioinformatics market, and the application of advanced computer science is the process that needs to take place. It's like in the banking industry. Originally I saw a lot of the banks try to do their own software, only to realize that they really weren't in the software business, they were in the financial business, and that standard, industrial-strength software components--applications and platforms--came into play. I saw a very similar scenario starting to play a role in the bioinformatics area.
So the two things I'm bringing to Pangea are my background in computer science and developing software platforms, and having managed large groups in an environment where a market is being created.
BioInform: You've obviously been in on the ground floor of some really big things. Do you think bioinformatics is going to be another one of those?
Couch: I do. I'm sort of a visionary; I like challenges. Right now bioinformatics is in the early adopter stage and the research realm, but I see it moving to the development phase and eventually over into the clinical market, clinical research organizations. I just attended a CEO conference and Jan Leschly, the CEO of SmithKline Beecham, spoke. One of the things he was talking about was the fact that the channel has yet to be created, from the data and the results of the genetic makeup of the individual to the individual. He was thinking five, six, seven, eight years out, but people are already starting to think about how to get this information about personal genetic makeup to the individual, to the physician, etc.
BioInform: As you staff up to move the company to the next level, what's your approach to recruiting the right people? I keep hearing about a shortage of qualified people.
Couch: That's probably true, especially if you're looking for a combination of computer science and molecular biology experience. The key centers around two things. One is a vision and a passion that's contagious. The second is an executive team where people see the opportunity to grow. The fact that I was one of the original execs at Apple and was able to be part of Apple's growth from zero to a billion and really define an industry is an aspect that, I think, people find exciting, because they want to work at a company that is well managed and has a vision and has a set of directions and a specific set of steps to move in that direction.
This is a very exciting industry, it's brand new, it's to-be-defined, and the really creative people thrive in that kind of environment. I come out of many years at HP and many years at Apple--I've been around in this valley a long time and know a lot of people, so we can lean on our resources. It's sort of contagious in the sense that once you get the top management team, it's easier to get great people because a lot of people tend to follow that management team.
Plus you have to admit, having IVP, Kleiner Perkins, and Mayfield as investors doesn't hurt us in terms of attracting people, either. I think this is the first time they ever invested together in one company. Out of the 86 investments they have jointly in the biotech and life sciences areas, we are the only investment in the bioinformatics area. The reason for that, I think, is that they looked at each one of their 86 companies and saw that every one of them was being forced to do software because there was no common platform, no common set of tools, no commercial- or industrial-grade set of tools that they could purchase. So they sat there and said, gee, does it really make sense if the core competency of these biotech companies and these life science companies that we've created is really science, that we spend all of this time doing computer science? What we really need to do is invest in a company whose core competency is computer science, particularly data management, and build those tools that then can be supplied to these other companies. And that, I think, was their original motivation for investing in Pangea.
BioInform: So you don't need to find people who are both biologists and computer scientists, you can just create tools that biologists can use without having to be experts in computer science?
Couch: The best approach is to have biologists focused on defining the tools, and computer scientists focused on implementing them. Then you have specialist-level knowledge bearing down from both sides. It's critical that we form relationships with the academic and scientific communities and provide them a set of tools that they can use. To me it's no different from going back, in the early Apple days, and saying gee, what I really need to do is build an interface that emulates the way people solve their problems, and then I need to build the generic functions that they do, the word processing, the spreadsheeting, the data storage. This model is one we've been through with other industries.
BioInform: What sort of advances on the hardware side would you like to see, from a bioinformatics perspective?
Couch: It's really interesting. If you take Pangea's model of bioinformatics from the generation of data to the capture, to the filtering, to the analysis, to the mining and the modeling, all sitting on top of integration, and if you draw that as a ladder, it gets interesting. The bottom end of that ladder, in terms of the generation and the capturing of the data, really is occurring on an NT-like box. It's when you get up into the modeling and the simulation that you start to see the need for the Silicon Graphics machines or some of these specialized hardware boxes. I've seen the results of 40 clustered NT's. I believe there's going to be a day when you have a scalable, expandable NT network that gives mainframe performance. In fact, I've seen the research results right now where a clustered set of NT's handles one billion transactions a day or a hundred million hits on a web site a day. That tells me that there will come a day when you will have a hardware solution that spans this entire ladder that is NT-based; multiple Pentium processors or whatever. So I see the technology of expandable, scalable, network-distributed NT's as a real possibility. I've seen some technology at UC Berkeley--it's called the NOW project, which is the Networked Office Workstations--where they have the concept of a proxy server such that when the load on the network gets to be at a particular level, the network itself goes out to the complete network at UC Berkeley to find which PC's are not being utilized, and brings that computational power to bear on solving the problem. There is some interesting stuff coming down the pike. I've never been an advocate of dedicated hardware; it seems to me that Moore's Law in terms of machine performance continues to increase and so I could always put together a network of processors that will scale up to meet the computational needs.
BioInform: How will you and Joel Bellenson work together?
Couch: Joel brings to the party some very practical knowledge by virtue of the fact that he ran the sequencing lab at Stanford, and from there, between Joel and Dexster and a few of the original Pangea engineers, they developed many custom applications for people over the years, starting back in 1991, with the clinical genetics database at Stanford, and then the sequencing laboratory information management system at Stanford, and then the yeast genome map lab system, then they did a compound screening project, then they moved into sort of product tracking, and then they did a human gene expression database, they did some chemical stuff, protein expression database for OGS.
So he brings to the party a very wide range of experiences, and that really plays a key role in defining the data schema of our platform, and the data layers of the warehousing software. He and the Pangea engineers bring to the party an awful lot of unique experience across different types of applications. He's working right now with me on ensuring that the data schema is robust, along with some of the strategic partners that we have formed relationships with. He's sort of the visionary looking out at how do we make the transition from the research phase to the development phase to the clinical or diagnostic phase. I'm freeing him up from most of the operational aspects so that he can be a student again, he can stay on top of the literature, attend a number of the conferences, work on some of the relationships that he's had with people since the beginning of this industry. And I count on him very heavily for his domain knowledge. Joel's challenge is to keep the company's domain knowledge accurate and current.
Coming in the next issue of BioInform: John Couch talks about his specific plans for the future of Pangea Systems.