Don’t collaborate; Work together
In their spare time, many engineers and biologists enjoy writing philosophical letters to editors of scientific journals, claiming that the age of collaboration and multidisciplinary research is upon us, even though we’ve been hearing this for much of the past 20 years.
In those letters, the authors express their concerns about the state of computational biology: “Engineers and biologists should collaborate more”, they urge. And collaborate they must: the current state of affairs is such that a lot of experiments in biology generate so much data that we aren’t able to process any of it with ease. Supposedly, there are oodles of science that we could learn from that data if only we had enough engineers and time, or computer cycles, to analyze it all.
One example that comes to mind is sequencing DNA. Currently, sequencing your complete genome would cost you around $10,000 and several hundreds of gigabytes of disk space. That said, companies like 23andMe.com will gladly sequence a small chunk of your genome. For roughly $200, they will mail you a kit that you spit in and then mail back. They extract the DNA from the sample, analyze it and send you the results. That said, they only look at certain regions of your genome known as Single Nucleotide Polymorphisms (SNPs, pronounced snips). These SNPs are locations in your genome where you find a mutated letter in the DNA sequence; such variations can be thought of as typos in your DNA.
While the price of sequencing a genome keeps decreasing (see our Q&A with NHGRI director Dr. Eric D. Green), the problems encountered with storage and data processing remain a big issue. Beyond that, the other issues that creep up include whether sequencing every human on the planet is feasible from the point of view of technology, whether it is useful in terms of the medical information we can reasonably extract from looking at sequence data, and whether it is even desirable in the first place, from the point of view of privacy.
But let’s go back to our multidisciplinary collaborators. On the face of it, it seems like such a simple problem: Gather biologists and computer scientists in one building, supply unreasonable amounts of caffeine and have them collaborate to solve all our problems.
However, a quick thought experiment would reveal that scenario to be ineffective. The much thrown-around idea that all we need to do is to build research facilities with floors that have both engineering and biology labs to increase interactions is wishful thinking. If the problem originates from differences in research culture, increasing the number of chance encounters in the building will get us nowhere.
Collaboration ≠ Working together
Working together does not mean having biologists conduct experiments and asking computer engineers to analyze the data later. It has been tried for years and has been the cause of much frustration and wasted time.
Biologists who want to plan an experiment correctly ought to have discussions with computer engineers before the experiment to have an idea of which experimental parameters will ensure significant results.
Conversely, computer engineers cannot develop data analysis tools without understanding the biology behind the experiment. Otherwise, how could they possibly know about the caveats of an experiment and how those show up in the data?
This fantasy world where biologists and computer scientists need only be near each other to foster an atmosphere of collaboration is becoming increasingly absurd
