Almost a month has passed since I published an opinion piece called “All biology is computational biology” in PLoS Biology.
In my paper, I envisioned a biology that explicitly and clearly acknowledges how much it has changed over the last 20 years, how much its questions have changed, and how much the practice of doing biology has changed. I envisioned a biology that gives credit broadly and fairly to everybody who contributed to key insights – regardless of what tools they used.
As intended, my paper provoked many responses from the community, and in the following you find my thoughts on some particularly interesting comments.
“But everyone already knows how important computational biology is.”
You are correct. No one I have ever met has disputed that computational work is useful and important for many applications.
Usefulness, however, is not the issue here. The widespread frustration I observed stems from the fact that usefulness and importance have not translated into a fair share of credit. This problem is due to the old-fashioned power structure of the life sciences–as represented by hiring committees, funding committees and editorial boards.
I illustrate this issue with three examples at the beginning of the paper. Two examples are anecdotes from my career, but the general feedback suggests that they are representative of many other peoples’ experiences.
“But experimental work is also important.”
You are correct. I highlight how much biological knowledge today depends on computational foundations. This claim does not imply that computational research should or will completely replace experimental research (come on, stay real!). It also does not imply that training in computation can replace training in more traditional biological subjects.
All scientific examples I discuss in my paper depend on synergies between experimental and computational work. They are about establishing context for experimental results (through databases), condensing large datasets to guide follow-up experiments (through statistics) and making hypotheses testable in experiments (through rigorous modeling).
“But there were many quantitative approaches even before computational biology came along.”
You are correct. Computational biology did not develop out of the void. Quantitative approaches in genetics (eg RA Fisher) or pattern formation (eg Turing) predate the modern field of computational biology. I have not written a historical survey, which is why you won’t find a discussion of these historical approaches in my paper.
And while they all have been very important, none of these historical approaches have had the impact on biology that technology-driven changes had over the last 20 years. These recent changes were dramatic and make it necessary to redefine what biological research is and how credit is being distributed – and this is what my article is about.
“But I develop computational methods and don’t want to be a biologist.”
No worries! I don’t want to force you to be one. “All biology is computational biology” is compatible with “Some computational biology is not biology.”
Indeed, I see two possible paths to a successful research career in comp bio: either you are very good at the computational side –at making the fastest algorithms and the most powerful statistics– then it is completely OK if you don’t have a specific biological question you work on.
Or, you pick some biology that excites you and use your computational skills to understand it better. This is the path I follow: I work in a cancer biology institute and a substantial part of my group works in the wetlab. I am definitely a biologist and I wrote my article from this perspective.
“But this is just like saying something silly like >all biology is chemistry<.”
No, it isn’t. I can see why you might think that from the title, but there is a major difference.
You are referring to reductionist efforts to base a field on some lower level of explanation: Biological processes use chemical molecules, so biology is actually just chemistry. Chemical molecules are based on physical laws, so biology is actually just particle physics. I never found these intellectual party games very interesting, and I am definitely not doing anything like this in my paper.
To the contrary, instead of reducing biology to computation, in my paper I broaden the definition of what good biological work is. I want biology to embrace its computational foundations and to give credit where credit is due.
“But why are we having this discussion at all when the only thing we should care about is doing good science?”
Every time this argument is being made, you can see all heads nodding wisely.
It is a good argument to shut up a discussion that makes you feel uncomfortable: Who disagrees with the importance of doing good science? No one, or course. Then let’s stop talking and go back to work.
A powerful argument – but deceptive and naïve.
Think about it: What is good science? There is not a single criterion. Whether your science will be judged as good or bad depends on the intellectual background of who is judging you – the people facing you in hiring committees, funding committees and on editorial boards. This is true for all fields, but experience shows that for a long time the deck has been stacked against computational biologists.
This is exactly the message of my paper: Computational biology should be judged as good biological science.
I am positive for the future.
There are many people pushing for a redefinition of what good biological research is. For example, just a few days ago in Nature Genetics, Casey Greene and friends described the outcome of the first round of Research Parasite Awards to celebrate creative and insightful computational re-analyses of existing data. This is their inspiring response to the NEJM editors ‘research parasites’ insult that I mention at the beginning of my paper.
Several visible journals like Nature Genetics and Nature Biotechnology have a very good track record of publishing computational work. From my own experience, journals like PLoS Medicine do a good job at catching up.
But we are not there yet. The strong reactions I got show that many residual biases remain. Serious improvements will only come over time through generational change, not from opinion pieces, no matter how thought provoking they are.
In the short term, however, I hope my article has given computational researchers arguments to defend the foundational role they play in modern biology. I hope it has motivated them and reduced their widespread frustrations.