A common thread on the web these days seems to be the discussion around what bioinformatics is, where it takes us, and the role of bioinformaticians in biology. I write this as a theoretical chemist/biophysicist who somehow ended up in the world of bioinformatics when there were very few people who knew what a multiple sequence alignment was all about. I also write this as someone who’s seen the “field” evolve from barely having a name to one where universities offer programs in bioinformatics.
The two specific threads that resulted in this blog post are a question by Attila on Friendfeed and some soul searching by Shirley. As I read through the responses, some of which, especially in the Friendfeed thread, pretty much nailed my thoughts on what bioinformatics is about, one thought always came to mind. It’s what I call The Curse of BLAST.
While bioinformatics, and the practice of bioinformatics, has come a long way in the last 10 years, for a number of non-specialists, bioinformatics is still synonymous with BLAST, the poster child for commoditization of a scientific method. This is not a bad thing per se. It’s rare when a particular method proves this robust, but it does drive me crazy when you hear college kids pretty much equate bioinformatics with BLAST, and I suspect so do a large number of biologists, because that’s what they see. In other words, they think bioinformatics is just about sequence searching and alignment using a readily available tool. But the reality is very different. Bioinformaticians are developing new methods, complex pipelines, trying to develop classifiers, diagnostics, etc. Personalized medicine, pharmacogenomics, translational medicine, pick your buzzword du jour, all of them require quality informaticians, people who can develop methods, develop pipelines, and as the data get more complex, the challenges get more complex.
I like to think that part of a good bioinformaticians job (usually in conjunction with a good software team) is to eliminate their current job, by making complex methods available to biologists, and then move on to the next challenge. And there will always be challenges, and that’s always a moving target. When I finished grad school, my first job was as a scientific programmer at a bioinformatics company. If I left grad school today with the same skills as I had then, there is no way that I could cut it as a bioinformatician now, cause that bar has moved. There are people/companies that appreciate this role, and others who don’t. If bioinformatics a profession? If molecular modeler, so is bioinformatician. I don’t buy the argument that everyone does bioinformatics. Expert problems needs skilled practitioners. Now if only they were used as such. It’s the curse of BLAST
Related articles by Zemanta
- Zemanta API: Bukisa and Biological text extraction (zemanta.com)
- Context and puzzles revisited (mndoci.com)
![Reblog this post [with Zemanta]](http://img.zemanta.com/reblog_b.png?x-id=f05f46d2-164a-4af5-9377-386d310ac7d6)
8 Comments
But you are propagating the same opinion, that bioinformatics equals a specific set of tools – you just disagree on what tools. Is genetics about PCR? Is biochemistry about pipetting? Rather bioinformatics is about answering questions relating to the evolution of DNA, investigating such a thing, and testing hypotheses relating to the other thing. This is the interesting part for me. However, it does seem that practising bioinformaticians divide into two camps: the tools people (often with a comp. sci., or stats background) and the hypothesis people (often with a biology background).
Deepak,
I see what you mean and there was a time when I couldn't agree more with what you say.
However, the bench people who equate bioinformatics=BLAST also have a point. To them, BLAST is about the only useful thing coming out of bioinformatics. I might be slighly exaggerating here, but not much. Most of what is going on in 'real' Bioinformatics (TM) is of little concern to the molecular biologist. To them, the field looks like a bunch of eggheads trying to devise more and more complex methods to tackle meaningless questions. Knowing that some particular feature follows a power law doesn't help me to find out what my protein does (or how to cure disease XYZ).
Over the last few years, bench scientists have noticed that bioinformatics can be useful for them in a few instances: completing sequence fragments, finding splice variants, designing primers, predicting function based on homology, finding modification motifs, etc. For most of these questions, BLAST is the way to go. Or at least the only way they dare to go without special training, special hardware and locally maintained databases.
If bioinformatics wants to be recognized by molecular biologists, it has to cater for the needs of the pedestrian. Obviously, there is the other side of bioinformatics – work done and published by bioinformaticians for bioinformaticians. You can get much more sophisticated in that area, but you will have to do without the cheers from the masses.
I guess it is similar to the situation in (popular) music..
But you make the same mistake by defining bioinformatics by its methods. Genetics isn't PCR. Biochemistry isn't pipetting. You should define it instead by the realm of questions it can answer, e.g. bioinformatics is about testing hypotheses relating to evolution, etc. (Ok, for sure, there are two camps of bioinformatics researchers: those who develop methods, and those who use those methods to investigate biological questions.)
I don't disagree. The message I want to convey is that bioinformatics is complex, and has a lot of sides to it, and isn't just about aligning sequences. Just as there are molecular biologists and cell biologists, informatics come in all shapes and sizes with different skills.
I am not saying bench scientists require special training. BLAST is great. Anyone can use it, and get mileage out of it. But data are getting more complex and you just cannot simplify everything. Yes, methods developers, combined with good software engineers can abstract out a lot of the complexity, and give back information which the biologist can then examine. But in this day and age, it has to be a team effort. The informatician, the software developer, the biologist, all working together. They all need each other, and each has skills that the other can usually not attain.
I don't disagree. The message I want to convey is that bioinformatics is complex, and has a lot of sides to it, and isn't just about aligning sequences. Just as there are molecular biologists and cell biologists, informatics come in all shapes and sizes with different skills.
I am not saying bench scientists require special training. BLAST is great. Anyone can use it, and get mileage out of it. But data are getting more complex and you just cannot simplify everything. Yes, methods developers, combined with good software engineers can abstract out a lot of the complexity, and give back information which the biologist can then examine. But in this day and age, it has to be a team effort. The informatician, the software developer, the biologist, all working together. They all need each other, and each has skills that the other can usually not attain.
Great post, really help me alot. Thanks.
Cheers,
Blog Review
One Trackback
[...] A recipe-based approach is biggest curse Some time back Deepak Singh wrote an interesting post The curse of BLAST describing how BLAST has become synonymous to the bioinformatics. Although his article highlights a [...]