I prefer to play with blood and guts and living creatures rather than living on computers.

My first point is that I’m sure majority of synthetic biologists are really interested in understanding their systems, but judging entirely from the steps of mainstreamers, they don’t _seem_ to be. This was my point in answer to Ben’s comment – even if the knowledge from molecular biology (which is more or less my background) or other fields is reused, it’s not reusable. Usually I have no way of learning what makes particular part work with respect to its sequence.

Your second point is absolutely valid one. At this level, programming and engineering blur and you’re right about it being a hybrid and difficulties of debugging. I was rather thinking in biological systems (level or two up), which some people consider systems not synthetic biology.

I’ve just realized that probably the main difference in my way of thinking about synthetic biology, or particularly about biological devices is an acceptable level of deviation in the way a design is supposed to work. I am fine with imperfection, noise and other things as long as the efficiency is detectable. Others may be interested in having systems that are robust and 100% predictable (like the oscillator you’ve mentioned). So it may be even that I’ve misunderstood what the whole field is trying to accomplish, as Tom Knight wrote on syn-bio mailing list.

“Mainstream synthetic biology doesn’t seem to be interested in understanding how car and bike works – it’s interested in taking both of them apart as fast as possible, puting labels on the parts and pretend that now we understand how they work. And while this approach can be succesful to a certain extent in engineering, biology, especially synthetic biology, is not engineering, it’s rather a programming.”

Many of us in the field *are* interested in how the car and bike works, in fact this is the main reason why I’m personally interested in synthetic biology. However, “programming” cells is more like a hybrid of analog electronics and computer programming with an emphasis on the former. Cellular circuits are different from running computer programs. Cellular circuits are a mix of analog, some digital and a lot of noise. Another significant difference between traditional programming and “programming” a cell is the instant gratification we get from computer programming whereas building a circuit in a cell can take months to accomplish (i.e engineer). Check out the recent paper in Nature by Hasty on his group’s work to build a feedback and relaxation oscillator, I think it is a great example.

I’ve been a little harsh in this post, but that was on purpose. I’m sure that current approach is going to take us very far, both in terms of designed systems and in terms of understanding them. However, I see enormous pressure on making a design work, but almost none on the understanding. Synthetic biology parts do not exist in vacuum and were not created from the scratch – these are elements that were analyzed for years, but almost nothing of that knowledge is attached to particular part in any repository I’ve been looking at. An example: scientists collected over the years lots of information on mutations in regulatory elements. Could be that knowledge reused for some fine tuning of particular regulatory part? I think it could be possible. The other example: there’s almost no knowledge transfer from protein engineering field. We know quite a lot on various combinations of protein domains, including the ones that don’t occur (or at least not very often) in nature. Again it’s not reused in synbio field.

In my vision synthetic biology field is a place where all knowledge on biological systems at many levels will be put in a good use – a true inderdisciplinary field. I don’t see that yet and I’m not so sure it’s because we lack understanding how biological systems work. There’s a lot of smart people in the field and many of them are talking about something quite similar. However the speaker list of SB4.0 conference was not indicating that there’s a change coming.

Quick question on this post, perhaps my exposure to synthetic biology is atypical, but do you think that the — to use your analogy — taking apart of a bike and a car and seeing how the parts fit or don’t fit together is devoid of value in terms of understanding the underlying system?

I don’t think there’s any serious synthetic biologist who would disagree with what you’ve stated are the limitations of the current genetic-engineering-and-then-labeling approach, and I’m sure their ultimate aim is to achieve the “compiler” ideal that you’ve described. But, I think the current approach is adopted precisely because we don’t have sufficient understanding of the building blocks and their interactions with one another to get at that idealized compiler setup.

Anyways, I’ll get off my soap box — I’m interested to hear what you think.

My dream is to have (build?) a service in which you can approach designing a biological system from as many directions as it is possible. For example bioindicator could be constructed based on existing reporting systems, biosynthetic pathway on a basis of known metabolic pathway, while particular structural features could be built from a PDB files (I envisioned last example after hearing about haemolysin used in the new sequencing technology).

GenoCAD is going in the same direction isn’t it? I noticed a placeholder for public designs – it’s a great idea.

As software engineers and computer scientists, not biologists, we want to discover the logic behind the computation that is life, not the low level physics that impliments the processor/cell. Once we understand the logic operations that a cell supports, we can use those operations to write “Hello World” for that cell, and design non-biological processors that will execute the same logic.

Jim Shaw
CEO/CTO
FOCUS Biology