Tuesday, July 20, 2010

Disordered ramblings

Of late I've become interested in so-called intrinsically disordered proteins (IDP's).* These are proteins that contain one or more "significant" regions of sequence that are unstructured. "Significant" can range from perhaps ten or so residues, up to the entire protein. There are experimental data suggesting that the disordered regions in some of these are vital for function. It is generally thought that disordered regions important for function might undergo a folding reaction when bound by another protein, or a nucleic acid, or even a small molecule.

I'm interested in IDP's** because two of my favorite proteins have disordered domains that are essential for function and that do undergo the kind of folding upon binding mentioned above. This makes these proteins more interesting to me intellectually (not that they would be boring without disorder), but can also make them significantly more difficult to study than your garden variety well-folded, globular protein.

The IDP field is populated by large numbers of bioinformaticists (spawning my last post). There are also experimentalists and computational biologists (of the molecular simulation kind), but much of the initial driving force in creating this as a field appears to have come from the bioinformaticists. A small group of them.

Who are seriously over hyping the field.***

The hype being based largely on predictions of disorder. Predictions. Not much data. A prediction is just a pointer to something that might (or might not) be interesting. It's pretty much meaningless without experimental verification.

This is a problem. Yes, we all need to sell ourselves and our research. We all need to convince others that what we do is important and should receive gobs of funding $$$'s. But what you're selling has to have some connection to reality. A track record. Data.

Right now the IDP field has all the appearances of an infomercial for some kitchen gadget that is promised to mix, knead, puree, blend, chop, slice, dice, julienne, fry, roast, bake, boil, steam, load the dishwasher, sweep the floors, put the children to bed, and polish your shoes. Only believable in the wee hours of the morning after a long evening consuming copious quantities of the alcoholic beverage of your choice.

For now I'm keeping my credit card in my pocket.










* There are many, many recent reviews on the subject. This one is okay (and free).

** I seriously dislike the name "intrinsically disordered protein." For a start, the majority of the IDP's that have been identified are mostly well-structured and only have a fraction of their sequences disordered. I saw someone use the term "intrinsically disordered region." That's an improvement.

*** Case in point: the many, many reviews. Many, many of which were authored/co-authored by this guy. Dude, enough already. Go spend some time in the lab generating new data.

6 comments:

JollyRgr said...

Somebody sounds a little annoyed...:-)

GMP said...

Sounds like the plea as old as science itself -- experimentalists think theory is crap/useless/worthless on its own and building theoretical/computational tools has no intrinsic merit.

I say have a little patience, Odyssey. When computational tools/techniques are good enough to compare to realistic experiment, I am sure verifications and predictions will come. But you have to allow some time and money...

Odyssey said...

GMP:
It's reasonable to assume that some experimentalists think little of computation, except what you don't know is I trained as a computational person and my experimental "reform" occurred fairly recently. And, in the dim and distant past, I dabbled in bioinformatics. I know enough to know the bioinformaticists in the IDP field can and should do better.

Also, over hyping is over hyping no matter who's responsible.

GMP said...

Odyssey:

Agreed -- my field is not bio related, but I think all fields have serious issues with overhyping. It's the fight among too many people for too few resources...

In my field it seems that the criterion for "maturity" of the subfield is when the thought of it makes people nauseated. It often happens that Nature or Science report a spectacular advancement, which no one ever repeats but hordes of students and postdocs try for years, and everybody and their brother jumps onboard. After much hullabaloo, three or four years later it turns out it was a crappy overhyped measurement and the real results are much more modest. But, all the hype propels the whole topic into the stratosphere and completely changes the funding milieu for a while.

To keep the funding, you cannot afford not to follow the latest hype, and it stalls the subfields where important, but slower and less flashy progress needs to be made. It's all quite exhausting, really. Sigh.

Anonymous said...

Hi,

as a theoretician/computational person myself I have a very good idea how dodgy some modelling can be so I sympathise and overhyping is a definite problem.

I am wondering, Odyssey, whether there's a relationship between protein disorder and surface activity (proteins at liquid interfaces being a particular interest)? Certainly there are some proteins, like caseins, which are quite disordered and are strongly surface active (though there are some counter-examples), so I'd be interested if you had any thoughts

Odyssey said...

I am wondering, Odyssey, whether there's a relationship between protein disorder and surface activity (proteins at liquid interfaces being a particular interest)? Certainly there are some proteins, like caseins, which are quite disordered and are strongly surface active (though there are some counter-examples), so I'd be interested if you had any thoughts

To be honest this is not something I've thought about before. I would imagine disordered chains at a liquid interface would behave differently (as compared to in bulk solution) and could have an effect on the interface itself. I'm vaguely aware that there is some literature on polymers at interfaces, but I'm really not familiar with that stuff. Food for thought...