I'll start with a story: after news of the post-9/11 anthrax attacks in the fall of 2001, an older colleague told me that he didn't think the attackers were very sophisticated bioengineers. "After all," he said to me, "even you know how to make antibiotic resistant bacteria." It's true - I, as a very green, inexperienced graduate student, knew how to make antibiotic resistant bacteria - it's one of the few things I could do easily in the lab.
My colleague was in essence saying that it's not really that hard to hack life in some basic ways. The idea of biohackers has been floating around for some time now. After all, if a teenager sitting out somewhere in western Russia with a computer and a lot of time on his hands can hack the Pentagon or write viruses that cause billions in damage to the world's IT infrastructure, why can't a future biohacker, again some bright but twisted teenager sitting in his underwear in the basement (it always seems to be a guy who does these things - girls are equally skilled, but are maybe less likely to unleash havoc upon the world out of sheer boredom), in the not too distant future, make an equally infectious biological virus?
The question is essentially, how hard is genetic engineering?
Physically, it's easy and getting cheap too. Our ability to physically synthesize and manipulate genes, and to place them in foreign hosts (Craig Venter's whole genome transplant experiment being the most extreme example) has progressed tremendously in the last 30 years. Much of this is stuff that a Russian adolescent truly could do in his basement.
But manipulating genes is one thing - making them do what you want is a very different problem.
In fact, in all except for fairly trivial (but sometimes effective nonetheless) cases, making genes do what you want is extremely hard. If you compare the building of a modern, suitcase-sized nuclear weapon to writing a computer virus, I think biological engineering is much more like building a nuclear weapon than it is like writing a computer virus.
Why aren't people building nuclear weapons in their basements? Well, ok, I'll admit it's somewhat dangerous (but hey, people build meth labs... about equally dangerous to the individual in the immediate vicinity) and the materials are hard to come by. But why is building a nuclear weapon so hard for even nation-states that are willing to make a serious effort? Again, getting highly enriched uranium is tough, I admit - but why not enrich it yourself?
Here's something to try at home: head on over to Borders and grab a book from O'Reilly or Peachpit Press on algorithms or game physics or MySQL. Then go over and find one on fluid dynamics, the easiest-looking one you can find, open it up and compare it to the O'Reilly Manual. There is a huge difference.
Computer science can get very sophisticated and technical - I'm not denying that. But the fact is, you don't need to know much to start doing a lot: do-it-yourself computing is widespread. Do-it-yourself fluid dynamics is not.
Why? The concepts of fluid dynamics are much, much more abstract, and require a rigorous foundation of background knowledge that is very hard to acquire on your own.
So even if the materials were cheap, designing your own nuclear bomb (I'm talking about a suitcase-sized implosion bomb, not the shoot-a-peg-through-a-uranium-doughnut kind, which didn't even need to be tested before it was used in WWII) would still be much, much more difficult than designing your own computer virus. Or designing your own plane from scratch. Yes, you can buy kits of pre-made airplane parts with a given design, and maybe you can customize a little, but you're not really designing an airplane.
So what about biohacking? Is it as hard a fluid dynamics? We can play around a fair bit with existing designs, the ones produced by evolution. You can get bacteria to make insulin, get yeast to glow green in response to a human hormone, make a weed that metabolizes TNT, or, in the most heroic efforts, engineer a multi-step metabolic pathway to produce a drug by adding enzymes to the beginning and ends of an existing metabolic pathway.
This is today's bioengineering. Don't get me wrong; we can do a lot with it.
But my inspiration (and those of many other systems biologists with whom I've mulled this problem over) has been the computer-aided-design of the Boeing 777. Keep in mind that this amazing feat of aircraft design was achieved after over 100 years of work in aerodynamic engineering, so of course we shouldn't expect the same thing for biology any time soon.
But there's one thing aircraft engineers (and, for that matter, electrical engineers) have had that I don't believe biologists have yet - a rigorous, quantitative physical theory as a foundation for their efforts. Major advances in fields like fluid dynamics and solid state physics have gone along with engineering progress over the years.
So where are biologists going to find their theoretical foundation? My prediction is this: progress in using computer models to engineer biological systems from scratch will only advance with progress made by modeling reverse-engineered, existing systems. We need to develop our theoretical know-how by building models that predict the behavior of existing complex biological systems.
Modeling existing systems is occurring more frequently in biology, but it has yet to transform the whole field. Genomics has changed the way almost every biologist does research, but modeling and systems biology have not.
So, tune in tomorrow for my next installment of aimless speculation - I'm going to talk about how modelers might go about revolutionizing biology and laying the theoretical foundation for much more sophisticated bioengineering.
In the mean time, here's a sobering quote from the 777 design story I linked to above:
"While the Boeing 777 experience is exciting for the VE enterprise, we should recognize just how limited the existing CAD tools are. They deal only with static solid modeling and static interconnection, and not—or at least not systematically—with dynamics, nonlinearities, or heterogeneity."
"Note that if we introduce uncertainty in our description of the components, it can drastically increase the computation required to do pairwise checking and make the bounding box approach even more attractive. Actually, while the basics of solid modeling are well-developed, there is no standard approach to uncertainty modeling even here and many open questions. Once we introduce uncertainty in a general way, then exponential growth in evaluating all the possibilities becomes a worry."
Biology is all about uncertainty (because it's hard to make all of the relevant measurements), nonlinearities and heterogeneity. Of course engineers do have special tools for those nonlinearities, and biologists are envious.