Computational research, and medical degree - a match?

[Sonya]

Hi,

I want your opinions and thought on this issue. Could "computational methods" in research be along the lines of work someone pursues if they truly are interested in "medical" research. Anyone can do anything, but, would it be difficult to match interests in both medical research and computational research?

Let me define my terms.
Computational research/methods: this could be a lot of things. But, here, I am referring to theoretical predictions that utilize mathematical modeling. Of course, they probably have some basis in data, but they are also largely theoretical. I'm particularly interested in computational cellular neursocience/neurophysiology, but discuss any areas you are familiar with.

Medical research is, for the sake of this dicussion: basic research that has visible either a. medical uses or b. DIRECT implications for understanding diseases. Of course no biological research has no medical significance, but, i think you understand what I mean here.

The discussion ground is open.

Sonya

---

Members don't see this ad.

 

[mjs]

Sure.
Neuroscience and genetics come to mind immediately, but there's tons of other stuff. Sweet computer skills are becoming a neccesity in a lot of molecular biology labs. As far as immediately clinical stuff like in silico pharmacology or really whizbang models of disease processes, it's still a ways off.

Here are some examples I know of:

An older colleague in the <gasp> lay press:
http://seattlepi.nwsource.com/local/203984_brain16.html

David Altshuler develops methods to better associate complex genetics with disease:
http://www.hms.harvard.edu/dms/bbs/fac/altschulerda.html

 

[dl2dp2]

i'll give you my 2 cents. I have had the exact problem myself.

realistically, computational neuroscience, particularly at single neuron level, if you are talking about complex Hodgkin Huxley, multiple channel, biochemical, complex cable model etc, has a LOT to offer to experimental science. This is important because currently there's a wealth of data in single neuron physiology using things like high resolution calcium imaging with two-photon.

However, neither of the above has much relevance for medicine.

things like brain-computer interface, (as mjs posted), i think, have a lot of potential. However, I'm unclear as to how that is going to be implemented. If you think you want to do something like that I would strongly recommend looking into your biomedical engineering dept.

honestly, my friend, i think computational neuroscience is a very basic science that has hardly any medical relevance at this point. However, i don't think much of the basic science per se has any medical potential. so. it's more of a temperament thing than anything else. Can you see your self runing gels for 3 years?

From a practical perspective, you can still get a faculty position with a computational background. from that point, you can always hire postdocs to do the experiments FOR you.

 

[mendel121]

Epidemiology, and specifically genetic epidemiology come to mind. SOme one with this background would be well suited to design, carryout and devise methods for family based studies, and the hot topic in the field right now, genome wide association studies. THe study of complex diseases in genetics needs smart people educated at the interface of medicine and computational methods. Check out this program which has as its goal just this -

Center for statistical genetics/ Genome sciences training program -
http://csg.sph.umich.edu/

This includes a training program with worldclass faculty. Also check out the list of trainees, everyone from Evolutionary biology to human genetics to biostatistics to bioinformatics. Sounds like it's right up your alley.

 

[fantasty]

Epidemiology, and specifically genetic epidemiology come to mind. SOme one with this background would be well suited to design, carryout and devise methods for family based studies, and the hot topic in the field right now, genome wide association studies. THe study of complex diseases in genetics needs smart people educated at the interface of medicine and computational methods. Check out this program which has as its goal just this -

Center for statistical genetics/ Genome sciences training program -
http://csg.sph.umich.edu/

This includes a training program with worldclass faculty. Also check out the list of trainees, everyone from Evolutionary biology to human genetics to biostatistics to bioinformatics. Sounds like it's right up your alley.

I agree with mendel. U Mich is very strong in statistical genetics. (I took a genetic epidemiology class in grad school, and half of the readings were based on work at umich). At Iowa, we also have a public health genetics in the College of Public Health, which would be a viable PhD program in our MSTP. They have a lot of interesting research projects going on, too.

http://www.public-health.uiowa.edu/pphg/

And we've been hearing more interest from other parts of campus (computer science, etc) for folks interested in lots of different types of bioinformatics (proteomics, etc). Plus some of our MSTP students have been doing some interesting work with modeling protein structures and predicting active sites of enzymes within the biochem dept.

 

[mendel121]

I agree with mendel. U Mich is very strong in statistical genetics. (I took a genetic epidemiology class in grad school, and half of the readings were based on work at umich). At Iowa, we also have a public health genetics in the College of Public Health, which would be a viable PhD program in our MSTP. They have a lot of interesting research projects going on, too.

http://www.public-health.uiowa.edu/pphg/

And we've been hearing more interest from other parts of campus (computer science, etc) for folks interested in lots of different types of bioinformatics (proteomics, etc). Plus some of our MSTP students have been doing some interesting work with modeling protein structures and predicting active sites of enzymes within the biochem dept.

I have a UM bias, but like dante said, there are others - check out U. Washington, Hopkins, Harvard, Wash U - all have similar programs I think.

 

[crackerj]

Well, I think one of the most relevant examples where computational research is directly related to medicine is visualization of data acquired from MRI, CT scan, etc., in medically meaningful ways, especially by reconstructing a 2D set of images into a 3D model. Although it may sound simple, a lot of work have been done in the area of 3D visualization. You can refer to this website: http://public.kitware.com/VTK/.

Probably 3D visualization alone may not suit your interest but I think I know at least one example where you can use "theoretical predictions that utilize mathematical modeling" in the context of 3D visualization and medical research - "Probabilistic Brain Atlas". If you want to explore this area in detail, try Google or PubMed. I think a group of researchers from UCLA is fully engaged in this work.

Good luck..

 

[Reimat]

Sonya,

I'm a rising college senior applying to MSTP programs this fall with the desire to computational neural modeling work for my PhD. I have communicated this desire with a few MSTP programs, and in all the responses I have received so far I have been told that I appeared to be a good match for their programs.

I don't know what kind of modeling or how much you have done, but this summer I'm working with a Professor who is well know in the field of neural modeling and has contributed much to it. Many of the models he has created are helping to produce a better and deeper understanding of how the brain works (and- for example in the case of amnesics- fails to work). It is research that directly pertains to how these patients should be treated- not in drug form, but suggesting they can learn in ways not previously known.

Those are my opinions/experiences with modeling, and while I'm slightly biased, I think it's a good PhD choice for MSTPs

 

[Sonya]

i'll give you my 2 cents. I have had the exact problem myself.

realistically, computational neuroscience, particularly at single neuron level, if you are talking about complex Hodgkin Huxley, multiple channel, biochemical, complex cable model etc, has a LOT to offer to experimental science. This is important because currently there's a wealth of data in single neuron physiology using things like high resolution calcium imaging with two-photon.

However, neither of the above has much relevance for medicine.

...

honestly, my friend, i think computational neuroscience is a very basic science that has hardly any medical relevance at this point. However, i don't think much of the basic science per se has any medical potential. so. it's more of a temperament thing than anything else. Can you see your self runing gels for 3 years?

From a practical perspective, you can still get a faculty position with a computational background. from that point, you can always hire postdocs to do the experiments FOR you.

That is the type of stuff I see myself doing, in terms of computational.. channel biophysics/single neuron/few network neuron simulations.

Of course, I can easily see myself doing purely basic research for the rest of my life (along with patient care), i know many physicians who do it. I will also, if i do computational stuff, learn experimental methods in gradschool. I've read a lot of computational papers. The ones done by purely computational folks are usually crappy...

I'm already in gradschool. The question i posed has been in my head a long time, and I generally agree with sluox that it has limited relevance. but, then i see your point that a lot of basic science has limited relevance... so, why not do it?

The question came up in my head recently, because I just talked to a particular professor at my school who does basic science, ion channel research. He studies genetics of mutations that lead to epilepsy, and e-phys functional biophysics of the channels. He is *purely* an MD. And, thus, when I met with him, I was really suprised to learn he had a relatively sizable portion of his lab doing computer modeling. As an md, he obviously wasn't a basic science nut, who did not care of medical implications - not at all!

I posed a question similar to this to him, and he said, basically, computers have use everywhere. He thinks, it's like the next molecular bio... gonna become the thing of the future.

But, i just thought.. well... hey, of course HE says they will be the future... i mean, he's probably got the biggest computational cellular neuroscience group on our campus (we don't have a lot of CNS people here... oh well, at least we got GOOD researcher, thought not a lot). But, I wanted a less biased opinion.

Thanks