Stephen Wolfram on Bioinformatics


Stephen Wolfram is the creator of the popular Mathematica program, the author of A New Kind of Science, and a keynote speaker at O'Reilly's upcoming Bioinformatics Technology Conference. Tim O'Reilly recently conducted a brief interview with Wolfram about his research, his new book, and its connection to bioinformatics. We're looking forward to a more substantial opportunity for a Q&A with Stephen at the Bioinformatics Conference.

O'Reilly: What is the link between A New Kind of Science and bioinformatics?

Wolfram: Bioinformatics has traditionally concentrated on handling genomic data---mostly in the form of strings. What's now starting in computational biology is the study of actual processes that operate in biological systems. The things I've discovered in writing A New Kind of Science suggest a framework for thinking about such processes. The key idea is to think in terms of simple programs---and to use those to capture the essential mechanisms in biological systems. From genomics we know the lowest level code for biological organisms; now we have to have models for how that code actually runs.

O'Reilly: How can finite automata be used to model biological processes, such as evolution?

Wolfram: You probably mean "cellular automata." They're my favorite example of simple programs. And they are good models of some specific biological processes---like the ways complex pigmentation patterns are formed, and probably the way some complex cell morphologies are produced. I've looked at lots of other simple programs too, and many of those are more broadly relevant for biology.

About evolution: One of the key issues is what it takes to get the complexity we observe in biology. Our usual intuition is that to get complex things requires a lot of effort. And people have assumed that a complex process of evolution must be involved. The things I've discovered suggest that actually it can be very easy to get complexity: it takes only a simple program--and it's easy for such programs to arise in biological organisms.

O'Reilly: What kinds of computing do you envision will be required to make "a new kind of science" happen? What kinds of computational hardware, software, networking, and so forth?

Wolfram: What's made it happen for me is Mathematica. I built Mathematica to have a general and very high-level computational environment that would let me do experiments and set up models easily. And it's worked fantastically well. I would never have been able to discover more than a tiny of fraction of what I have without Mathematica. From a technical point of view, a key idea is symbolic programming, which is fundamental to Mathematica. In terms of hardware, it's always been more important to run the right program. In the 1980s I used massively parallel computers and such. But over the last decade or so, I've really just used quite generic computer hardware for the things I've done. Except perhaps when I have occasionally called into service a few hundred PCs instead of just one.

O'Reilly: What can biologists and computer scientists do to contribute to this new kind of science?

Wolfram: Lots! Given the idea of looking at simple programs, there's a huge territory to explore. Both at the level of basic abstract science, and at the level of applications. I think practically every page of my book brings up a bunch of new questions to address. Soon I'll be putting an explicit list of "open problems" on

O'Reilly: Who else is a leader in this new kind of science? In the six months since the book came out, who has "got it" and is actively pursuing your ideas?

Wolfram: I had expected it'd take about six months for there to start to be real development on the basis of my book. And it seems that timescale is more or less right. We're starting to be deluged with very interesting things that people are doing on the basis of the book. It's quite overwhelming actually. There are people in almost every area of science--and beyond--starting to do things.

O'Reilly: Will you be demonstrating any approaches to biology and bioinformatics at the Bioinformatics Conference?

Wolfram: I certainly plan to.

O'Reilly: How do you respond to criticism that your "breakthrough" is mostly speculative, and based largely on the unacknowledged work of others?

Wolfram: I'm not sure how looking at simple programs can be viewed as "speculative." The programs just do what they do. There's no speculation involved. When it comes to applying ideas about simple programs to questions like the fundamental laws of physics, then, yes, that's definitely speculative for now.

About history: I did a lot of historical scholarship for my book--vastly more than most scientists ever do. And all the evidence is that I got the history right. History is hard, you know. I was amazed at how many times I had the vague impression that the history went a particular way, but when I went and looked at actual documents, and pushed hard enough in talking to people, it turned out something quite different was true. In the end I was rather pleased with the extent to which I managed to unravel histories of all sorts of things.

O'Reilly: Given that you have specified copyright (all rights reserved) on all of the images in the A New Kind of Science book, how can generating an elementary automata and printing the results in a paper or Web page not violate that copyright?

Wolfram: The copyright notice in A New Kind of Science is a bit more sophisticated than that. It's an attempt to unravel some of the complex intellectual property issues associated with simple programs and their outputs. We definitely don't have all the answers yet. I am very committed to trying to help people make the best use of what I've done, but I want them to do it in ways that are fair and appropriate. Defining all of that in clear-cut legal ways is quite a challenge.

O'Reilly: What's next--for you, your institute, your work, Mathematica, and so on?

Wolfram: There are two big things for me right now: taking the next steps with Mathematica, and setting up an infrastructure to support A New Kind of Science (NKS). There's a lot of exciting things that I've wanted to do with Mathematica for years, which I'm now finally having some time to concentrate on. The response to NKS is quite overwhelming, and I'm trying to put in place structures so that people can work productively on NKS, and communicate well about it. One thing we're just announcing is the NKS 2003 Conference, to be held in June 2003. There'll also be an NKS summer school. By the way, I just have a company, not an institute, at least right now.

O'Reilly: Where do you see the biological sciences going in the next 10 to 20 years? What about computing?

Wolfram: Well, lots of exciting things are surely going to happen--at least a few of them perhaps made by possible by things I've built.

O'Reilly: What role, if any, has serendipity played in your career?

Wolfram: I've wondered that myself. It's hard to know. At some level I've been very systematic in pursuing my goals for nearly a quarter of a century. But there are plenty of specific ideas and discoveries that it feels like I could have missed but for some piece of luck. Of course, after the fact they often seem quite obvious. And it seems quite outrageous that one might have missed them....

Return to