nodalpoint.org - Bioinformatics and Computational Biology - Comments

Up the complexity

mako — Thu, 01 Jul 2004 23:46:36 -0400

Although I hate to buck a trend, I do enjoy being the devil's advocate. Let's up the complexity with the terminology. Let me explain:

I actually prefer the term 'theoretical molecular biology' for my own research work. By introducing such additional delimiting terminology I simultaneously acknowledge the increasing complexity of biological research and differentiate the types of bioinformatics that support such research vs. that which exists for its own intrinsic value.

What I mean by this is that bioinformatics has become two-fold (at least) - that which is a service to be used to mine digital information by researchers, and that which is a discipline in its own right. The b-word, once overused and truly abused (by both practitioners and transients) actually has meaning with such a differentiation, where bioinformatics itself is concerned with the CS algorithms to store, retrieve, sort and manage extremely large data multisources (a true interpretation of information theory, which extends into the medical domain).

On the other side of this terminology are disciplines which are not in and of themselves of service to other researchers. Computational biology, for example, concerns itself with modeling, simulation, and development of heuristic algorithms to reduce and formulate biological phenomena. Theoretical molecular biology is the arm of biological research that is reliant upon such data (sequence, expression, structure, biophysical...) to make meaningful probabilistics inferences about molecular activities. This division may seem trivial, but such a schema (and resultant terminology) makes it easy to visualize and implement large scale working groups which is often necessary to meet the goals of high throughput in-silico experimentation.

Chris's comment is very apropos: "... it's about the intent, rather than the method." However, it is convienent to have meaningful terminology when working together in a global electronic environment, since with such titles I can identify people I need to augment my research efforts.

The major difference between a thing that might go wrong and a thing that cannot possibly go wrong is that when a thing that cannot possibly go wrong goes wrong, it usually turns out to be impossible to get at or repair.

Second that

chris — Tue, 04 May 2004 03:42:43 -0400

I'm a biologist. Specifically, a geneticist; but I work on biological problems. Computational techniques are a set of tools in my arsenal (the pros and cons I'll leave for another time) which I apply to problems that interest me. Nothing more. Just like I use recombinant DNA techniques to generate data -- to test a biological hypothesis.

So, my definition is: you are a biologist if you work primarily on biological problems, using any methods; you are a computational/information scientist if you apply computational theory to datasets that happen to include information of biological origin (ditto mathematician). In other words, it's about the intent, rather than the method. True, there are bound to be exceptions, but then there's no such thing as an iron-clad biological definition, is there?

Definitions

rohan — Tue, 04 May 2004 02:01:50 -0400

Here, here. Let's cut the crap and just get on with it. Personally, I can't stand either term -- "bioinformatics" is far too value-laden these days and I find "computational biology" simply pretentious. When people ask me what I "do" I tell them it's called "biological data analysis".

This kind of navel-gazing term definition is seemingly at odds with what is supposed to be a 'cutting-edge'' field: personally, I feel this is indicative of the highly routine nature of much of what passes for science in this area. Most of this kind of thinking seems to be generated by a subset of computer scientists who seem to be obsessed with problem-definition. I've had encounters with several of these individuals who seem to be unable to deal with an application-domain problem unless it fits neatly a given previously defined problem type. But frankly, what do you expect from a field in which many published papers seem to stop at toy examples that fit the assumptions of their allegedly novel method exactly.

Definitions

Greg — Mon, 03 May 2004 01:14:52 -0400

Define it however you want, you'll only know what bioinformatics or computational biology is if you're actually doing it :)

Bioinformatics and Computational Biology

kirubhai — Sat, 01 May 2004 10:57:08 -0400