Monday, April 22, 2013

Returning to Standard Life

My psyche may not be quite settled again, but it is now officially time to return to a more standard existence.  And by that, I mean the establishment of scientific and engineering standards.  I'm on my way, right now, out of the country over to the UK for the next SBOL standards development workshop.  SBOL expands to Synthetic Biology Open Language, and it's a community-driven effort to make it easier for practitioners of that science to exchange information about the systems they are building.  This applies to person-to-person communication, but also (perhaps more importantly) to communication from one computerized tool to another.  This latter is so important because building new organisms involves a lot of different types of information and processes, and it's hard for humans to track all of it correctly except for the very simplest of designs.  It also allows engineering services, like fabrication of new DNA, to be taken on by specialists who can take advantage of economies of scale to do them efficiently and cheaply.  The group at Newcastle is hosting this meeting, and we'll spend three days showing off our newest tools to each other and delving into such weighty topics as prioritization of working groups, standards body governance, interoperation with related standards, and extensions to handle additional types of knowledge.

Not long after I get back, I'll be hosting another standards discussion, the Workshop on Metrology for Mammalian Synthetic Biology that we're organizing in conjunction with the first Workshop on Mammalian Synthetic Biology.  At that one, we'll be asking a lot of really basic questions that don't yet have good answers, like "Who wants to know how a biological part behaves in a cell?", "What do we need to measure about its behavior?", and "How do we tell whether we got the measurement right?"  I expect the discussion will be much broader than just mammalian cells, but we needed to narrow the scope enough to have a productive discussion.  Mammalian cells also have some interesting properties that make them an attractive target: they're relatively big and tough, which seems to let them get less messed up by adding our new biological computing circuits into them, and there are also starting to be a lot of good mammalian biological computing components to choose from, which means we might soon be able to try to build much more complex systems.

In one sense, this type of work is boring.  After all, most of the process of getting things done is about having long dry arguments about tiny points of detail, and there's nothing inherently sexy about building better rulers or trying to agree on a data format that doesn't leave anybody too dissatisfied (consensus and satisfaction are necessarily opposed qualities).

That first impression is misleading, however.  Every scientific endeavor, no matter how sexy the topic may seem, has a lot of tedious work underlying it that is necessary to building a solid foundation.  Debugging code is never sexy.  Building makefiles and regression tests is never sexy.  Feeding mice or culturing cells or staring blankly at columns of data in Matlab is never sexy.  But you can't get where you need to go unless you do the hard work.  The alternative is to practice "slash and burn" science, where you go just far enough to get proof-of-concept results, then publish them with grandiose claims and move on to the next thing, leaving little for those who follow you to build on and yet making it difficult for them to do the work you ignored: "Didn't Smith and Jones already do that?"

When you get the unsexy parts right, you enable great things.  And standardization is a huge, huge, part of that.  Standards, and especially standards of measurement, are civilizational infrastructure.  Consider, for example, a 2x4 in your local hardware store: that humble piece of wood represents an absolute revolution in the construction and remodeling of houses.  Likewise, the next time that you chance to use a tape measure, think about how hard it is to make tape measures the same length.  How is it that you don't get different lengths of tape measures from different stores?  Why doesn't the length of tape measures drift over time as the machines used to make them slowly wear out? They do vary, of course, but a remarkably complex system of engineering mechanisms, professional associations, and government bureaucracy combine to ensure that you always have as many significant digits of length available as you need and are willing to pay for.

I didn't get into synthetic biology out of a desire to work on standards.  I got into synthetic biology to practice sexy cool mad science with living organisms.  Getting there from where we are today, however, will require that we build that depth of infrastructure that is so much needed and so often unappreciated.
Post a Comment