CIBCB Day 2
The 5am start was because I wasn't sure how to get from the hotel where I was staying to the Renaissance downtown. Turns out that all metro systems are pretty similar, and I made it without too much hassle[1]. Today's schedule was two parallel sessions, so I didn't see everything, but I think I managed to catch most of the bits that interested me.
I chaired the first session, on gene networks, and gave the last talk in that section. Most of the talks were about reverse engineering gene networks from microarray data. Since I'm in the middle of a review which covers this area, I was particularly interested. It's a topic which is a natural for a computational intelligence approach, since the problem is desperately underdetermined; there's just not enough data to solve it. Added to which there are multiple networks which will produce the same gene expression pattern, the structure of the actual networks is unknown, and DNA-protein interactions are certainly not the whole story, and it becomes apparent that the problem is non-trivial.
My opinion, which was not substantially altered by today's talks, is that it can't yet be done. Current algorithms work pretty well on small, artificial networks, but don't scale to realistic network sizes, and the networks they produce cannot be properly validated. It's still an interesting area, though, and well worth working on. The insights and algorithms developed now may well be much more practically valuable when applied to the larger, more complete data sets which undoubtedly be generated over the next few years.
There were good sessions in the afternoon on Biological Theory and RNA structure and function. The latter was mostly algorithms for predicting RNA secondary structure from sequence data. Since we're starting to realise just how important RNA is, I think we'll be seeing a lot more of this sort of work in the next few years. And RNA structure might be easier to predict than protein structure, which is another largely-unsolved problem with huge biological implications[2].
Went to dinner tonight with a group of female bioinformaticians, an occurrence which is as rare as it is enjoyable. Then the train to Guelph, and, hopefully, some sleep in preparation for another day of talks tomorrow.
[1] Apart from the hotel desk clerk, but we won't go in that...
[2] Or then again, it may not.