Talk:Synthetic biology

From Wikipedia, the free encyclopedia

I kicked off the key enabling technologies section. But it still needs quite a bit of work. Rpshetty 20:21, 6 April 2007 (UTC)

• the statement in the introductory paragraph, "There is significant controversy over whether synthetic biology as currently practiced is in accordance with the holistic approach," needs supporting references or it should be deleted.

• it might be good to break out new sections having to do with technologies that help with synthetic biology. sections might include:
  • modeling (including mathematical modeling)
    • quantitative modeling can be broken down into a number of categories, such as
      • Atomic models of biomolecules, including quantum (ab initio) calculations of enzymatic reaction mechanisms, molecular dynamics simulations of protein-protein, protein-RNA, and protein-DNA interactions, hybrid QM-MM simulations, and equilibrium Monte Carlo calculations of free energies
      • Course-grained models of biopolymers, including RNA and DNA, that describe their folding/loop dynamics
      • Simulations of 'network-level' dynamics or statics, which includes the usage of Boolean networks, ODEs, SDEs, and Master equations (or Continuous Time Monte Carlo simulations aka stochastic simulation), that describe how a system of proteins, RNA, and DNA molecules (and their binding sites) evolve over time. (I could write a primer on this topic. Wikipedia uses LateX, right? :) )
      • Bioinformatic models that use evolutionary information to infer structure-function relationships
  • fabrication (including de novo DNA syntheis)
  • measurement
  • others?

(I still think mathematical modeling should be under engineering. A simple question will determine the best location for it. Which occupation studies and uses more mathematics: any engineering discpline or molecular biology? I could write a longer justification, but the answer will be the same.) Salis 23:32, 22 September 2005 (UTC)

(Agreed that, at the moment, engineers tend to be the ones who are using the math, but... Most of the engineers/physicists who are now working in biology aren't working to engineer biology, rather they are working to help understand natural biological systems. The four sections are really meant to introduce the different types of work happening in synthetic biology, not how the work is done (i.e., what are the motiviations and applications of each faction?). E.g., much of Adam Arkin's work, Jim Collin's work, Alex van Oudenaarden's work is in service of analysis of natural biological systems, not engineering new synthetic biological systems. So, I was sort of viewing math and modeling as a technology that is applied in service of a particular application, in this case the science of biology, as opposed to the engineering of biology. And, since the math. and models. can be used all over the place, I still like the idea of pulling them out of the introductory text entirely and starting a new sections on technologies that enable synthetic biology) Endy 15:27, 23 September 2005 (UTC)

(That's true. Most of the current usage of mathematics is to study natural biological systems. The next step, which is really a very small step from the technology usage point of view, is to hypothesize synthetic biological systems that perform a specific function and then to build those systems in the lab, ie. design and construct. I think the tools for studying biological systems and designing them will be almost identical. For ODEs, there's Auto and XPP. For Master equations, nothing like that exists yet, but the research is moving in that direction. Then the only thing that separates the science from the engineering is the intent of the study. Do you want to understand how a natural systems works or propose systems that perform some desired function. Of course, understanding how the system works will make it easier to propose new systems that do something interesting. Salis 17:35, 23 September 2005 (UTC))

(Agreed. Also, note that understanding is essential for engineering, but perfect understanding is not. In fact, trying to gain a perfect understanding may prevent essential work on the actual engineering. E.g., the telegraph was invented and deployed ~30 years before Maxwell's equations got figured out. Historically, human's have been able to engineer useful artifacts prior to perfect understanding. In the early days of recombinant DNA technology, my guess is that the folks wanting to express human insulin in bacteria did *not* care about making an ODE, SSA, or any other type of formal mathematical model/simulation of their system. So, for me, intent is real important (i.e., intent is key). Finally, one foundational idea in engineering is insulation. Insulation, when it works, makes modeling easier. If Gerry Sussman were "here" he'd point out that you, and your computer, don't care about the complex magnetic field produced by the jumble of powercords running behind your desk -- even though we could try to model and simulate this electrical field -- instead, via insulation, we simplify the part of the physical world that we want to engineer (the wire) so that the modeling and model-based design are easier. Endy 18:25, 23 September 2005 (UTC))

(Right. All models are simplifications of reality and a perfect understanding of the system will never be achieved. The good models are ones that capture the important mechanisms and provide insight into how they work (and work together) to exhibit a certain behavior. Importantly, the knowledge gained from the model should supercede the specific model itself. So determining the effect of a feedback loop in one specific biological reaction pathway is ok, but using that knowledge to say something about all feedback loops in any system of such and such characteristic is much more useful. But that's why models are useful: you simplify the system down to the lowest common denominator of all interesting systems and then you see what it does. You can always add complexity, but good models are only augmented by additional complexity...not qualitatively changed by it. On the other hand, I am also of the mind that a "good model" is one which can be directly used to build something useful. So if the model is too simple and is completely disconnected from reality then it's not very useful at all. But that might be what separates a physicist's model from an engineer's. The engineer really does need to know what the numbers mean and how they quantitatively affect the system at the end of the day. As for the electrical field surrounding my computer, I'm sure the guy who designed my motherboard made it robust enough to tiny electromagnetic perturbations. He probably didn't make it robust against large electromagnetic perturbations, which is maybe why my friend likes to scare me with his Tesla coil. ;) Salis 04:15, 24 September 2005 (UTC)