DNA from hagfish being used to synthesize bulletproof slime (opens in new tab)

Biological abstraction is simply not a thing. You must understand that the fundamental difference between digital logic, software, etc. and biological systems, reaction mechanisms, signaling, and the like is not simply about sorting out how the different things are connected together. The biggest differentiating factor is by far is how deeply fundamental high-order nonlinear dynamics and complexity are within biological systems. The topology is an implementation detail and isn't want causes us problems.

Consider the nature of digital logic for algorithms of a low computational complexity. There might be a simple state machine, some kind of input, and one or more terminal states. Let's reduce this even further to the level of logic gates. NOT, AND, OR, and all the like are extremely simple and trivial to calculate. As far as the physical implementation of these goes, I have to admit that my knowledge of EE is fairly poor. I know there are some rather involved mathematics. AFAIK, however, the ultimate determinate of "on" or "off" state is a thresholding function. It's really not that bad of a problem.

Now let's consider one of the lowest possible levels of what we would consider to be a biochemical system: two chemicals involved in a reaction. For our sake, we would probably make it an acid and base problem, since acid-base chemistry is prevalent in biochem and it's quite easy at the freshman level.

http://en.wikipedia.org/wiki/Chemical_equilibrium

As you can see, the treatment is a little bit more complicated than Boolean algebra. There are terms for Gibbs free energy, chemical potential, activities, solubility potentials, etc. to consider for a system sitting in some glassware. I won't get into the details, but do make note of the math. Can you think of how to compute these quickly in silico or by hand? Probably not too hard.

We're just getting started though. Things get rather noisy as we start to consider more than a few chemicals. As we trend up towards approximating an in vivo system, we also have to developing a model for interaction. To do that, we have to understand each component. At both ends, we enter into the realm of statistical thermodynamics. Here you'll encounter much more complicated models and concepts such as canonical ensembles:

http://en.wikipedia.org/wiki/Grand_canonical_ensemble

You know can't simply consider aggregate systems as pairwise interactions or summations or any other trivial model. For one, there are too many incredibly localized chemistries within solution that produce rich behaviors. For example, here's a decent figure that presents a chemical transport process: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2980713/figure/f...

All of the familiar, easy math typically involves equilibria, and equilibria systems are nice and easy systems to study. But to be alive is to not be at equilibrium--one facet of biology is to push against that state as long as possible. Nonequilibrium systems present further computational and modeling challenges for us. http://en.wikipedia.org/wiki/Boltzmann_equation

I've presented a bunch of math, and I'm not done yet. It's known that Metabolic pathways, the interplay interacting metabolites, happen to be NP-hard. Typical of graphs, fairly orthodox. The computational complexity of the topology isn't really isn't expressive enough to convey the beauty of a graph edges themselves--that they are in constant flux, connected through a chemical calculus. Chaotic, transient metastabilities that arise and collapse. Digital bits blink, but chemicals dance.

Do you know the breadth and depth of this machine? Here's a wonderfully illustrative map of the known metabolic pathways--at a glance it'll give you an instant perspective if you didn't already have context: http://www.cc.gatech.edu/~turk/bio_sim/articles/metabolic_pa... (When observing the forest, be careful not to forget what the trees actually represent. There is unfathomable complexity to be found in the small interactions.)

It probably dawns on you now that we're not going to live to see the answers. I've only really introduced metabolites here. Gene regulation, proteins... I never even mentioned the word "cell". That's far too much.

If it doesn't make sense why bioengineering is slow-paced, then I did a poor job at conveying how hard the problem space is. Yes, we can make certain gene products and achieve simple little results that still dazzle us. It's not quick, it's leagues away from what we want, and it's definitely not abstractly or generically portable. (I don't refer to the assays/methods.)

I suspect the only way to enter into the biological renaissance requires that we get better at everything else: math, low-level logic, algorithms. I imagine we'll have to find ingenious new ways of greatly contracting the problem space, reducing dimensionality, operating over compositions of subgraphs, etc. in order to deal with computing the complicated nonlinear dynamics at interaction level. Integration into an acceptable holistic model is mind-boggling. How to do this and get meaningful results, I have no idea. Whatever far future tools that might be effectively applied to this will appear to have bent the Matrix.

Thinking of today's "big data", I can't help but feel dwarfed at a truly astronomical scale. The depth of ignorance we'll eventually have to cross in order to solve the problems presented by a single cell... it puts everything into perspective.

Sorry for the long post. I hope I illustrated that the core tendencies of electrical engineering and computer science towards simplification and logic reduction are simply not very amenable to bioengineering. You can't simply tweak one thing or edit the binary. Everything touches everything, and you can't adjust the weights and levers without understanding the ultimate runtime behavior. You already know that line from Jurassic Park about the weather...

As we learn more, I agree that higher level approaches will certainly find application. But to the idea of creating industrially useful quantities of spider silk, what's really going on is engineering an entire production workflow. For example, we've heavily selectively bred Bombyx silkworms, to the point that they are true domesticates: they cannot survive (or even eat) without direct human maintenance. (Note: there are many varieties of wild silkworms as well, many tended for fiber use.) This is incredibly labor intensive, and the fact that we continue to do this speaks to the high value placed on silk. We could engineer another domesticate biofactory as a spider, but that seems likely to also be incredibly labor intensive. But who knows, maybe an arachnid equivalent of Bombyx mori could spawn an entire spider silk industry?

Still, the desire to use bacteria as protein factories holds obvious allure for industrial scaling. Of course, that means we need to engineer not just protein secretion but the entire process from gland to spinneret.

I would expect so, since these processes are working towards using the same protein structure organized in the same way as produced by the living creature.

Also note that TFA and my earlier discussion is all about protein fibers and attempts to recreate those. Protein fibers include animal fibers such as silk, spider silk, wool, alpaca, and many more. Another approach to novel fiber qualities might arise through bioengineering and process improvements around cellulose fibers. In fact, this is exactly what's happened with rayon fibers. Rayon is a now-broad category of cellulose based fibers. Lyocell (branded Tencel) is one variant that was specifically designed to be pollution-free amongst other qualities, eliminating some serious environmental concerns with the chemical production process of early rayons. Likewise, you'll find "bamboo" yarn and fabric, which is a kind of rayon based on bamboo cellulose vs. the more usual wood chips.