Many readers will have heard of Rosetta@Home. It's a distributed-computing approach to protein folding problems, which is certainly an area that can absorb all the floating-point operations you can throw at it. It's run from David Baker's lab at the University of Washington, and has users all over the world contributing.
A reader sends along news that recently the project seems to have come across a good hit in one of their areas, proteins designed to bind to the surface of influenza viruses. It looks like they have one with tight binding to an area of the virus associated with cell entry, so the next step will be to see if this actually prevents viral infection in a cell assay.
At that point, though, I have to step in as a medicinal chemist and ask what the next step after that could be. It won't be easy to turn that into any sort of therapy, as Prof. Baker makes clear himself:
Being able to rapidly design proteins which bind to and neutralize viruses and other pathogens would definitely be a significant step towards being able to control future epidemics. However, in itself it is not a complete solution because there is a problem in making enough of the designed proteins to give to people--each person would need a lot of protein and there are lots of people!
We are also working on designing new vaccines, but the flu virus binder is not a vaccine, it is a virus blocker. Vaccines work by mimicking the virus so your body makes antibodies in advance that can then neutralize the virus if you get infected later. the designed protein, if you had enough of it, should block the flu virus from getting into your cells after you had been exposed; a vaccine cannot do this.
One additional problem is that the designed protein may elicit an antibody response from people who are treated with it. in this case, it could be a one time treatment but not used chronically.
The immune response is definitely a concern, but that phrase "If you had enough of it" is probably the big sticking point. Most proteins don't fare so well when dosed systemically, and infectious disease therapies are notorious for needing whopping blood levels to be effective. At the same time, there's Fuzeon (enfuvirtide), a good-sized peptide drug (26 amino acids) against HIV cell entry. It was no picnic to develop, and its manufacturing was such an undertaking that it may have changed the whole industry, but it is out there.
My guess is that Rosetta@Home is more likely to make a contribution to our knowledge of protein folding, which could be broadly useful. More specifically, I'd think that vaccine design would be a more specific place that the project could come up with something of clinical interest. These sorts of proteins, though, probably have the lowest probability of success. The best I can see coming out of them is more insight into protein-protein interfaces - which is not trivial, for sure, but it's not the next thing to an active drug, either.
1. bad wolf on April 30, 2010 8:07 AM writes...
Since so many people were interested in immigration issues effecting science on this blog recently, you may be interested in the current immigration reform proposal.
"This proposal will reform America’s high-skilled immigration system to permanently attract the world’s best and brightest while preventing the loss of American jobs to temporary foreign labor contractors. At the moment, high-skilled workers are prevented from emigrating to the Unites States due to restrictive caps on their entry. In order to accomplish this goal, a green card will be immediately available to foreign students with an advanced degree from a United States institution of higher education in a field of science, technology, engineering, or mathematics, and who possess an offer of employment from a United States employer in a field related to their degree. Foreign students will be permitted to enter the United States with immigrant intent if they are a bona fide student so long as they pursue a full course of study at an institution of higher education in a field of science, technology, engineering or mathematics. To address the fact that workers from some countries face unreasonably long backlogs that have no responsiveness to America’s economic needs, this proposal eliminates the per-country employment immigration caps. "
http://www.politico.com/static/PPM154_proposal.html
Permalink to Comment2. John on April 30, 2010 8:15 AM writes...
Derek, are these experimental results or just a virtual screening effort?
Permalink to Comment3. Me on April 30, 2010 9:47 AM writes...
Nice post! Good reporting on different approaches to solve a similar problem.
Permalink to Comment4. anon on April 30, 2010 10:13 AM writes...
These approaches are in an unfortunate middle ground between severely compromised physics and a disregard for how proteins actually fold in a cell. So implications in drug development are low on at least two very relevant scales.
The enzyme engineering stuff is very cool, it's a pity they can't just sell the idea on those terms.
Permalink to Comment5. okemist on April 30, 2010 10:46 AM writes...
Isn't this where you medchemists make a peptomimetic that us synthetic chemists can make easily and we make billions?
Permalink to Comment6. Andrew Kitchin on April 30, 2010 12:12 PM writes...
It's worth mentioning that a program called foldit (URL: http://fold.it) that has been used for protein folding problems for a while as well.
Permalink to Comment7. Ethan on April 30, 2010 12:20 PM writes...
Andrew, just a FYI, foldit is also a product of the Baker lab: http://depts.washington.edu/bakerpg/
Permalink to Comment8. Slurpy on April 30, 2010 7:09 PM writes...
So what's the difference between Rosetta@Home and Folding@Home?
I've been running F@H for about six years; is there some secret that should make me feel guilty for running F@H instead of Rosetta?
Permalink to Comment9. Nat on May 1, 2010 5:21 PM writes...
So what's the difference between Rosetta@Home and Folding@Home?
Rosetta is a protein structure prediction and design program; it is not intended to generate physically plausible simulations, but instead uses various heuristics (fragment library from the PDB, for instance) to attempt to generate a physically plausible final model of the protein of interest. In other words, it isn't the process that's important, it's the end result.
Folding@Home is a simulation program, I think it uses the molecular dynamics program GROMACS now. It is explicitly trying to simulate the process of protein folding, typically using small model proteins. The reason distributed computing works for this is the stochastic nature of these simulations - the majority are dead ends, but a tiny fraction will produce the desired results.
I've been running F@H for about six years; is there some secret that should make me feel guilty for running F@H instead of Rosetta?
Not really; it depends whose science you think is more interesting and worthwhile. As a crystallographer, I also care more about end results, but very little about processes, and I like Rosetta because it has the potential to simplify solving protein structures. Protein folding, on the other hand, puts me to sleep. But that's just my bias; F@H is an elegant solution to a very difficult problem, and the results look convincing. Neither program, of course, could ever be a complete substitute for good old-fashioned bench work, but both are good at generating hypotheses for experimentalists to test - which should be the ultimate goal of any theoretical study.
Permalink to Comment