John Wikswo and the Understanding of Biology

Today we had lunch with Dr. John Wikswo from the Vanderbilt Institute for Integrative Biosystems Research and Education. After lunch he gave a seminar with the title: “Can we comprehend the full complexity of our own biology?”. Who would miss a talk with such an appealing title!!

The talk was brilliant. Definitely one of the top 10 talks I have ever witnessed live. It was full of uber interesting technical aspects such as the fact that a mathematical model of a complex biological system might require Avogadro’s number of differential equations (apparently this is called a Leibnitz). How can we tackle such a problem? Dr. Wikswo, his team and collaborators are trying to use robot scientists and genetic algorithms to try to find the right equations that fit the data. He also talked about the concept of having biological invariants, just like in physics, which blew my mind away. Again lots of fascinating concepts and technicalities that I might talk about in future posts.

The most interesting thing for me, however, was the philosophical question that he enounced in the title. He separates the problem of understanding biology into two problems: an ontological one and an epistemological one. The ontological problem might be solved by all the interesting technicalities mentioned above. It’s a matter of getting enough data so that we get a picture of what ‘is’ (the ontological meaning of ‘is’).

The epistemological problem however is the one that worries Dr. Wikswo. And I have to say I don’t share his worry. I would argue that depending on how we define ‘understanding’ we don’t need to worry about it. Other scientific areas such as quantum mechanics and astronomy deal all the time with concepts that epistemologists might consider beyond our ‘understanding’. Do we really know/understand what entangled states are? Do we know/understand what’s the size of the universe? I think these are the same type of epistemological problems and yet we are able to use our knowledge about them and thus one might argue that we do understand them.

Anyway, I am no philosopher and I am nowhere near the level of understanding of Dr. Wikswo so I should stop writing before I embarrass myself talking too much about what I don’t know. But wait… how do I know I don’t know what I don’t know? Good question.

Holger Becker on Microfluidics

We met Dr. Holger Becker today. After a short visit to our lab he gave a very interesting talk about the state of the microfluidics industry.

If you are a member of the microfluidics community you probably know Dr. Becker from his regular focus articles in RSC Lab on a Chip. He began writing those as a way to engage the community on a debate to try to close the gap between academic applications and industry. I recommend reading them, he does a great job.

In his talk he started by introducing his company Microfluidic ChipShop and their fabrication capabilities. They have a wide set of capabilities as far as fabrication is concerned. They also pay special attention to standardize their fabrication. All their chips have a standard interface and layout that complies with mainstream existing ones such as 94 well plate dimensions.

Other aspects of their fabrication are their techniques and the materials they use. He reminded us that there is a plethora of materials to choose from and that even in an academic setting we should not limit ourselves to a couple of choices, if we do, we are limiting the manufacturability and mass production of our microfluidic devices. We all know this is true… but then again most of us are not material science people and sometimes we stay in our confort zone (a.k.a PDMS). In any case I came out of the talk thinking I should expand my knowledge in this regards. Hopefully, if I do, one day I’ll be hot embossing my devices.

During the Q&A someone asked about the need for a ‘killer app’. This is actually a recurrent topic among engineers in our field. Usually when most people talk about ‘killer apps’ they want to see a complete solution to a big problem and, if possible, completely solved by the almighty powers of microfluidics. They also tend to omit applications widely spread such as printer heads.

In my opinion we don’t need a ‘killer app’. Actually let me rephrase that. We do need a ‘killer app’, but we should not be looking for one. Killer apps are usually stumbled upon, not engineered. Instead we should be looking for killer microfluidic components. Components that work and that are standardized, like the standards in ChipShop for instance. I always say we engineers are grown up children who like to play with LEGO. Give us a bunch of pieces that we can snap together and interchange and we will come up with a ‘killer app’. In this case, however, we need to start by making those ‘pieces’.