More preaching to the choir, but you don't even know how bad it truly is. The "point charge in a uniform field" bit actually isn't working all that well these days. ..when you try to represent water with a continuum dielectric the standard approximations that are fast enough to use in a motional simulation (generalized Born) have something like 50 percent error when applied to a protein. ..that 50 percent error really kills you when you're trying to propose a thesis. It's no wonder the drug industry is giving up on this stuff. We need something like a thousand times more computational power, about 5 nobel prizes worth of insight, maybe 20 years, then it might be truly viable.
Easy there! I share your caution about randomly introducing things into an ecosystem, but they were talking about POND SCUM here. . . (plankton). It will grow in a vat in their lab and that's about it. Now the 1957 pandemic flu strain that was mistakenly shipped to medical labs all over the world... that's something to be worried about.:-)
Just for context that seems to be lost, the scientists are not trying to prove evolution generically but rather one of the particular mechanisms of evolution. The mechanisms have always been under great debate among evolutionary biologists.
Red Queen is more specific than just vanilla Darwin, it proposes a competitive advantage for organisms that have sex (as oppose to cloning like bacteria). Wikipedia actually has a decently clear and correct entry on this (always nice to see) http://en.wikipedia.org/wiki/Red_Queen [wikipedia.org]
I just literally cut and paste my answer from an identical question in another thread. Amazing how many journalists forget to put any contextual background when they report on science these days!
Red Queen is more specific than just vanilla Darwin, it proposes a competitive advantage for organisms that have sex (as oppose to cloning like bacteria).
Wikipedia actually has a decently clear and correct entry on this (always nice to see)
http://en.wikipedia.org/wiki/Red_Queen
Just for context that seems to be lost here, the scientists are not trying to prove evolution generically but rather one of the particular mechanisms of evolution. The mechanisms have always been under great debate among evolutionary biologists.
If I recall correctly, alot of similar pond organisms have dormant eggs as a natural way of surviving winter/the dry season and they hatch in spring. I once saw a talk about someone who was taking sediment samples from the bottom of lakes in order to count these sort of eggs and judge how water pollution was affecting their population. Cool stuff!
If I recall correctly, ebola-zaire seemed to be less and less fatal with each serial infection. By the time it got to tertiary infections (i.e. infected by someone in contact with someone who contacted a patient from the original outbreak) very few people died or even showed major symptoms. Ebola's dayjob is in some unknown organism where it is less fatal and longer incubating. Hopefully it stays that way.
Heh. . but the research monkey would cost more than the grad student since you have to pay for its "retirement" up front, mildly ironic.
But I agree it does often worried me that cell culture cells have to be pretty f--ed up to grow nicely in dishes in the first place. . .
And the next can of worms is that some things work great in animal models and then totally fall flat in human testing (like that B-amyloid vaccine).
Best of my knowledge HIV is transmitted from mother to child during birth . . but not in-utero. I remember reading single carefully timed doses of AZT could reduce mother-child HIV transmission in third world countries by ~50 percent. ..in other words, with proper care the children don't have to be automatically infected.
I guess the obvious question is whether your school offers any bioinformatics-specific classes. If not, you should strongly consider seeing if you attend a similar class at another nearby university. Trying to piece together the CS and Bio on your own is not particularly efficient! My school here has an excellent computational molecular biology class you can check out the syllabus at
http://bio5495.wustl.edu/
there are some books mentioned recommended by the professor at
http://www.genetics.wustl.edu/bio5491/
you will doubtless be using some pretty heavy duty statistical analysis, check out a syllabus from my undergrad school on statistical genomics.
http://www.bscb.cornell.edu/Homepages/raaz/BTRY4-6 82/
Good luck!
There actually was a STAR TREK TNG episode where an alien data capsule stored flight-path information as RNA. I excuse the science advisors for choosing RNA over DNA, although having extensively handled both in the lab I can say DNA would be a much more sound choice. RNA tends to degrade rapidly if you look at it funny, whereas DNA is reassuringly robust and you can heat it up and shake it up and it won't break. You choose which one to put in a data capsule in an exploding escape pod being chased by aliens with bigs guns:-)
As a grad student in biophysics, I can say that there are alot of similar musings in the literature, as well as a few preliminary "proof of concept" experiments as other posters have alluded to.
You might want to check out:
http://www.nature.com/nsu/030421/030421-6.html
Science. 2002 Apr 19;296(5567):499-502
and also EMBO reports vol 4 No 1 2003 pg. 7-10
Certainly evolution has come up with a pretty robust (although for some of us not robust enough)
methods of error correction and data storage. However, the bottleneck for us taking advantage of this would be the ability to manipulate DNA - our ability to duplicate, cut, and splice DNA in the lab has been entirely dependent on isolating enzymes from organisms that happen to do the exact reaction that we want (the use of t. aquaticus polymerase in PCR is a good example). Biological systems use DNA to, well, express proteins in response to stimuli and most of these enzymes aren't of obvious use for computational endeavors.
IMHO biological computing won't be truely feasible until we understand enough about protein structure & function that we can design new enzymes from scratch to do whatever DNA-manipulations we want. Which opens a whole other can of worms, trying to simulate biochemical pathways is another interesting field of recent interest (e.g. enzymes are regulated by large interconnected networks of enzymes that modify each other). Some people at UCSD are trying to simulate whole cells using modified analog circuit simulation software (since analog circuits are coupled DiffEQ's just like biochemical networks). Pretty far out stuff.
Slashdot Mirror
More preaching to the choir, but you don't even know how bad it truly is. The "point charge in a uniform field" bit actually isn't working all that well these days. . .when you try to represent water with a continuum dielectric the standard approximations that are fast enough to use in a motional simulation (generalized Born) have something like 50 percent error when applied to a protein. . .that 50 percent error really kills you when you're trying to propose a thesis. It's no wonder the drug industry is giving up on this stuff. We need something like a thousand times more computational power, about 5 nobel prizes worth of insight, maybe 20 years, then it might be truly viable.
Easy there! I share your caution about randomly introducing things into an ecosystem, but they were talking about POND SCUM here. . . (plankton). It will grow in a vat in their lab and that's about it. Now the 1957 pandemic flu strain that was mistakenly shipped to medical labs all over the world. .. that's something to be worried about. :-)
Just for context that seems to be lost, the scientists are not trying to prove evolution generically but rather one of the particular mechanisms of evolution. The mechanisms have always been under great debate among evolutionary biologists. Red Queen is more specific than just vanilla Darwin, it proposes a competitive advantage for organisms that have sex (as oppose to cloning like bacteria). Wikipedia actually has a decently clear and correct entry on this (always nice to see) http://en.wikipedia.org/wiki/Red_Queen [wikipedia.org] I just literally cut and paste my answer from an identical question in another thread. Amazing how many journalists forget to put any contextual background when they report on science these days!
Red Queen is more specific than just vanilla Darwin, it proposes a competitive advantage for organisms that have sex (as oppose to cloning like bacteria). Wikipedia actually has a decently clear and correct entry on this (always nice to see) http://en.wikipedia.org/wiki/Red_Queen Just for context that seems to be lost here, the scientists are not trying to prove evolution generically but rather one of the particular mechanisms of evolution. The mechanisms have always been under great debate among evolutionary biologists.
If I recall correctly, alot of similar pond organisms have dormant eggs as a natural way of surviving winter/the dry season and they hatch in spring. I once saw a talk about someone who was taking sediment samples from the bottom of lakes in order to count these sort of eggs and judge how water pollution was affecting their population. Cool stuff!
If I recall correctly, ebola-zaire seemed to be less and less fatal with each serial infection. By the time it got to tertiary infections (i.e. infected by someone in contact with someone who contacted a patient from the original outbreak) very few people died or even showed major symptoms. Ebola's dayjob is in some unknown organism where it is less fatal and longer incubating. Hopefully it stays that way.
Heh. . but the research monkey would cost more than the grad student since you have to pay for its "retirement" up front, mildly ironic. But I agree it does often worried me that cell culture cells have to be pretty f--ed up to grow nicely in dishes in the first place. . . And the next can of worms is that some things work great in animal models and then totally fall flat in human testing (like that B-amyloid vaccine).
Best of my knowledge HIV is transmitted from mother to child during birth . . but not in-utero. I remember reading single carefully timed doses of AZT could reduce mother-child HIV transmission in third world countries by ~50 percent. . .in other words, with proper care the children don't have to be automatically infected.
http://bio5495.wustl.edu/6 82/
there are some books mentioned recommended by the professor at http://www.genetics.wustl.edu/bio5491/
you will doubtless be using some pretty heavy duty statistical analysis, check out a syllabus from my undergrad school on statistical genomics.
http://www.bscb.cornell.edu/Homepages/raaz/BTRY4-
Good luck!
There actually was a STAR TREK TNG episode where an alien data capsule stored flight-path information as RNA. I excuse the science advisors for choosing RNA over DNA, although having extensively handled both in the lab I can say DNA would be a much more sound choice. RNA tends to degrade rapidly if you look at it funny, whereas DNA is reassuringly robust and you can heat it up and shake it up and it won't break. You choose which one to put in a data capsule in an exploding escape pod being chased by aliens with bigs guns :-)
You might want to check out:
http://www.nature.com/nsu/030421/030421-6.html
Science. 2002 Apr 19;296(5567):499-502
and also EMBO reports vol 4 No 1 2003 pg. 7-10
Certainly evolution has come up with a pretty robust (although for some of us not robust enough) methods of error correction and data storage. However, the bottleneck for us taking advantage of this would be the ability to manipulate DNA - our ability to duplicate, cut, and splice DNA in the lab has been entirely dependent on isolating enzymes from organisms that happen to do the exact reaction that we want (the use of t. aquaticus polymerase in PCR is a good example). Biological systems use DNA to, well, express proteins in response to stimuli and most of these enzymes aren't of obvious use for computational endeavors.
IMHO biological computing won't be truely feasible until we understand enough about protein structure & function that we can design new enzymes from scratch to do whatever DNA-manipulations we want. Which opens a whole other can of worms, trying to simulate biochemical pathways is another interesting field of recent interest (e.g. enzymes are regulated by large interconnected networks of enzymes that modify each other). Some people at UCSD are trying to simulate whole cells using modified analog circuit simulation software (since analog circuits are coupled DiffEQ's just like biochemical networks). Pretty far out stuff.