Philosophy, Science

Here be dragons! Thomas Kuhn, Statistics and System Biology

Here be dragons!

Thomas Kuhn had physics in mind when he wrote Structure of scientific revolutions but his key ideas also apply to statistics and systems biology and can explain some of the confusion in the field.

Thomas Kuhn’s Structure of scientific revolutions desribes the history of science as phases of normal science separated by revolutions and paradigm shifts. During normal science, research is guided by  a ruling paradigm, which identifies feasible problems and routes to tackle them. Normal science is a period of puzzle solving. The better your paradigm, the clearer the puzzle, the better your chances to solve it and progress.

Here be dragons!

Think of a paradigm as a kind of map: it tells you which direction to go and how to choose a good path to get there. The better your map, the more successful your research. If your problem falls into an un-mapped region marked ‘Here be dragons’ it will be really hard for you to solve it.

That’s one of the reasons why the physical sciences seem to progress at a much greater speed than the medical sciences. ‘How to cure cancer‘ is a much bigger and less well defined puzzle than ‘How to find the Higgs boson‘ for which physicists know how to build big pieces of machinery.

Without a map, you are lost. This is what happens during scientific revolutions.

One of the indication of a revolution is self-doubt, when it’s not clear if a discipline is actually doing science at all. Or worse: people can’t even agree on how to name their own field.

In established sciences you don’t see that much of it. Physicists are prime examples of self-confidence and most of philosophy of science is modelled on how physics sees the world.

Biologists also feel pretty secure. My cell biology and immunology friends never show much self-doubt: it seems clear to them that they do important science and that any discovery they make will be scientific progress.

Only sexy with an X?

But what about the ofther fields surrounding biology in today’s interdisciplinary landscape?

At Simply Statistics I follow a discussion on self-doubt in statistics. Somehow statisticians seem to feel that their field is ‘unsexy’ and doesn’t attract students (maybe because it’s too darn hard). Some propose renaming statistics into statistiX (with a super-sexy X) and others discuss the differences to data science (based in parts on posts at reddit). But, honestly, as soon as you need Venn diagrams to figure out what field you’re in, you sure are in trouble!

Apropos data science: Whatever it is, it must be a complete misnomer: All science is about data; we are all data scientists! ‘Data science’ is a pleonasm. And the content seems quite old: data mining and machine learning have been around for quite a while.

Ok, let me calm down again … I think fighting about names and labels is pretty useless, even though I can appreciate practical consequences between identifying yourself as a statistician or a machine learner: it’s not only a difference in jargon, but also in salary. Obviously, when I negotiated my salary a few years ago I made it quite clear that I consider myself a computer scientist doing hip and cool machine learning — and thus need to get paid better than those boring statisticians.

Are you calling me a bioinformatician?!

Statistics is not the only field where people seem unsure how to call themselves and what their contribution to science is. In my own field we have similar discussions. Some of us call themselves bioinformaticians, others computational biologists (where it is of utmost importance whether you empasize ‘computational’ or ‘biologist’).

The new kid on the block is systems biology. And even that comes in different flavors and warring factions. In talks by US researches I see mostly big data and hairballs, whereas Europe seems to indulge more in ODEs and PDEs and all that dynamic modelling. This plethora of questions, approaches, and methods does not bode well for the coherence of the field.

And, indeed, systems biology as a field is pretty badly defined. In a conversation I had a while back with the senior editor of the main systems biology journal, he tried to define systems biology as ‘looking at all the components of the system, no matter whether that’s 3 or 10,000’. I nodded my head wisely, and avoided telling him that I didn’t feel this is a definition at all. For starters, it doesn’t explain at all what this ‘system’-thingy is everybody is talking about. And then … ‘all components’, what does that mean? To make it a definiton you need to give at least some criterion what ‘all components’ means. How do you make sure you got all of them or at least the most important one? Which criteria do you use to choose your level of abstraction?

In a field-defining journal you need to do better than that.

Systems biology is where the dragons live!


Image sources:

3 thoughts on “Here be dragons! Thomas Kuhn, Statistics and System Biology

  1. I would say that systems biology is the opposite of the reductionist approach to a problem, and so at this stage could be defined as what ‘reductionism’ isn’t. That’s just shifting the goalposts, of course, because now you need to define what ‘reductionism’ is.

    Perhaps more to the point is that right now, ‘systems biology’ is as you say, where the dragons are. Maybe in 5 or 10 years people will have a formal definition. But surely until then, in the famous words of another case of subjectivity,


You gotta talk to me!

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s