Presenting an Interesting Exchange. Any Thoughts?

Presenting an Interesting Exchange. Any Thoughts?


Source: Reddit/Bioinformatics

Question from Xinlitik:

I’m an MD student interested in possibly doing a PhD in bioinformatics. Can anyone offer some advice?

Answer from synthespian:

Don’t. First of all, you will make little money - much less than as a clinical physician. Second, bioinformatics is a bandwagon on which a bunch of computer science types eagerly jumped on over these last 10 years. The field is festered with research that is just a mindless algorithmic void, with little regard or respect for biological facts, obtained the hard way, on the wet bench. It has become a playing field for the paper-publishing game. Doctors can’t play that game, and the computer science types think they instantly grok sophisticated fields, like immunology. Because of that, with some exceptions, we have a tons of research of questionable real value, in terms of improving clinical conditions, about which neither mathematicians, physicists or computer science majors haven’t a clue. You would not believe how dysfunctional a dialogue between physicians - who are basically unable to understand a single thing being said - and bioinformaticians is. If you don’t believe me, think about this: it is 2012 and we have not substantially altered clinical medicine because of bioinformatics, have we? We still work with “shift to the left” for leukocytes, and have to wait a week for bacteria to grow, and basically make educated guesses based largely on epidemiological data. This is because there has been very little understanding of what is really needed to advance patient treatment. Besides, you probably don’t have the heavy math chops to really work on bioinformatic algorithms. Math is hard and very time consuming. If you haven’t got ‘em by now, chances are you’re not going to be able to have the time for it. Being “java savvy from years ago” is not nearly enough. It would be nice if you understood a lot about number crunching - which is usually a hard area to master - even CS majors know little about numerical computing (case in point: Java). Now, if you ask me, an area that is in dire need of physicians with some notions of computers and engineering is in patient management. I don’t know how it is where you live, but where I live the health system is bordering collapse, and this is mainly due to mismanagement. Hospitals need to be administered the same way engineers administer airports and factories (plus the human touch, of course). But what I mean is that we need invest heavily in stuff like scheduling algorithms and be able to work out some optimizations for patient flux. Physicians are very ignorant about these things and if you put one in some management position, chances are the system will be mismanagement. CIO (chief information officer) positions are on the rise for physicians. You might want to look at that. A real revolution is just waiting to happen in the area of patient data.

He/she continued:

My last 2 cents. In fact, a warning: do not give up on medical residency. I have made that mistake, and now I am back in the rat race.

Here’s why: you get no respect from the people who are from computer science (etc.), and no respect from fellow physicians, who don’t understand your choices. You are not as sharp as they are at their field (because, frankly, you have no time for puerile things such as the next math Olympics); and your fellow physicians think you must’ve have hit your head at some point in your life and are crazy-cuckoo (besides, they are basically ignorant about your interest).

Take it from me. You will not believe how baffling it is to work with the people at the physics, math, etc. department, who think they can read a little something on the side about a complex area specialist area and then immediately come up with imbecile solutions for non-existent clinical problems.

Your best scenario is to wind up surrounded by biologists. The problem is, biologists often know - or care - little about the clinical conditions that haunt a physician’s mind. You see, to be a physician is to have a mission. You are committed to the betterment of human life. Do not forget this. It is a beautiful, important thing. Do not give up on this. This is why society has such a high regard for physicians.

So, you need to become a specialist.

Not only you will make much more money, you will get respect, and you will get to simple tell the “geniuses” what to do, because it’s all a paper-publishing game to them, they know nothing about real-world problems in the filed of medicine. If they do, they know very little. This is understandable. But, you know, the arrogance…It’s been said: the problems of biology are enormously hard. Physics deals with bouncing balls and particles. Biology deals with a monkey jumping from branch to branch, a far harder problem.

So do not give up on a residency program. Think about it this way: you will have more time for self-learning the computer stuff you love. Later on, you will do what you want, but you will do it from another level, higher up in the food chain. And, if and when you get tired you can just recoil to a profitable private practice.

Hear, hear - my advice does not come cheap, it’s coming from someone who has walked your path. ;-)

Question for readers - is the above person presenting a proper ‘outsider’’s perspective on bioinformatics, or is he too full of himself?

Few thoughts -

(i) Negative: The doctors are the primary interface between ordinary people, who are supposed beneficiary of research, and the researchers. So, if doctors do not see any value in bioinformatics, then neither are the patients going to see any use. The situation is somewhat similar to Toyota salesmen telling customers that Toyota cars are no good, while Toyota engineers are busy perfecting the engine.

(ii) Positive: It is too early to turn all skeptics in favor, but compared to where it was in 2000, or 2005, bioinformatics did come a long way. Ten, or five years back, even most biology researchers did not see any value in bioinformatics beyond having few web-accessible BLAST-like tools.

(iii) Negative: It is a valid criticism that most bioinformaticians are excited about the problem they can handle best (genome comparison), whereas the part that really matters in finding cure (immunology) gets very little attention. That was one of the reasons we started writing on uses of NGS to study immune system.

iv) Positive: Even though doctors themselves do not see any benefit from bioinformatics, the drug companies certainly do. Medical industry is R&D; heavy, and informatics plays a big role in both disease detection and finding cure. Especially, there is significant attempt going on to reduce costs of clinical trial through proper drug targeting, and removing the patients for who drug is less likely to work based on genetic information.

Admittedly, the question is about whether to quit MD to do PhD in bioinformatics, and the author suggests better alternative careers like epidemiology, electronic medical records, medical image analysis, etc.



Written by M. //