Is Bioinformatics really hard?

I would like to share some thoughts that came to my mind today, after a specific event having to do with bioinformatics training. First of all, don't get confused by the title of the post. It might sound like a selfish, elitistic or even racist comment! No, it has nothing to do with selfishness... Let me explain.

During my MSc in Bioinformatics, I met four kinds of people.

  1. Biologists or other people coming from life sciences and bench work that wanted either to switch to bioinformatics or to get basic training but failed to do so because of their fear to sit down and face the evil "black screen" of a Unix command line, let alone other hierarchically lower demons such as basic statistics, or "for" loops, or hierarchically higher demons such as R, algorithmics, basic Perl etc.
  2. Computer scientists and/or mathematicians (like myself) that wanted to apply their background knowledge to a more "practical" and at the same time still scientific level, than predicting algorithm complexities, solving partial differential equations or wandering inside Banach spaces. However, the "application" turned out to be quite difficult as the the types of RNA, the thousands of genes with strange names and those blurry gel images seemed more noisy than an elegant solution of a differential equation or one more O(nlogn) algorithm optimization.
  3. People of the first kind that made it.
  4. People of the second kind that made it.

As a result, out of the 15 members of the MSc program, only 3 ended up in actually working in the field of Bioinformatics.

A few years and a PhD later, the history repeats itself. Armies of ambitious students, fascinated by the combination of biology and computer science, interested in switching fields, turn out to be afraid of those strange dollar symbols and a few pages of chemistry. What is my objections to all this?

Well, prospect and ambitious students that would like to switch from one field to the other should keep in mind that it will not be very easy. By no means I am saying that this is not possible. Of course it is possible as is currently seen by numerous examples. What I am saying is that it is a little more difficult as it requires a little bit more studying than remaining in one's field and just keeping up to date. And this, because all of a sudden you one has to keep up to date with two fields. It also requires a lot of patience. It is impossible for a person without any prior computer/statistical knowledge to come and demand to start working right away. It's true that a good motivation like a hot biological system is always wanted and stimulating but without spending some time to learn basic computer skills and concepts, the result will be at least dissapointing, first of all for the person that has devoted its time.

The other way around also applies. A computer person cannot suddenly start working on a bench without spending a lot of time in learning practical lab skills (which to my mind would take much more time that a biologist sitting down with patience in front of a computer, but this has been a subject of debate between myself and collegues for some years now). The result will be messy if not dangerous!

The take-home message of this post and basically what I want to say is that people wanting to expand their field should think twice. Not because they can't do it, but because it will be more difficult than originally anticipated and require tons of patience to withstand the error rate of a lot of trial-and-error approaches.


Pfern said...

I agree with most of this , but the title can be taken too far.
People that are intending to use Bioinformatics may get to see the title alone and get a wrong message.

I think that it requires a level of abstraction that is not rare but is also not to be found in everyone.

bioinfoman said...

Hi Pfern... You are right, the title gives the wrong message so I changed it a bit. But I wrote this at a time of a little frustration, facing once again a person from the 1st category :-)

Post a Comment

Copyright © Bioinformatics dance