h a l f b a k e r y"Look on my works, ye Mighty, and despair!"
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Please log in.
Before you can vote, you need to register.
Please log in or create an account.
|
I'm not sure if we have the computing power to do this yet, but we might with distributed networks. DNA is made of discrete blocks of information that should work well with a computing environment. All environmental inputs/stimuli would have to be synthetic as the simulated creature(nobody said we
had to start with humans) would not have access to the outside world. If we started with, say, a fruit fly, we could eventually modify it's 'instincts' to achieve useful goals, such as search functions. Care will need to be taken to avoid replication and exessive learning capabilities. Not only would this begin the path toward true AI, it would help us understand DNA and biological structures. Modifications/mutations would be easy to introduce and test in a purely simulated and controlled environment with variable time scales.
Wikipedia: Genetic algorithm
http://en.wikipedia...i/Genetic_algorithm
I think this is the programming technique you're trying to invent - but it's much, much simpler than a full biological simulation. [jutta, Nov 20 2008]
[link]
|
| |
I'm actually talking about creating a virtual organism. Starting with the mapped genetic code of something and simulating a cellular environment, probably with an fertilized egg kind of thing. Start it just before it starts splitting cells. |
|
| |
Not doable with current knowledge or technology. |
|
| |
First, no life starts from DNA. You need at least some machinery (quite a lot of machinery, actually) to make DNA work. DNA makes RNA makes protein, but it's protein that makes DNA make RNA make protein. It is, almost literally, a chicken and egg problem. |
|
| |
However, you could include basic machinery in your starting situation, in addition to the DNA. |
|
| |
Second problem: we are quite good at predicting the sequences of proteins encoded by DNA (though not quite as good as predicting how genes are spliced in eukaryotes - a whole other ball game). But we are completely hopeless at predicting how those proteins will fold. The best current algorithms are incredibly slow, give only approximate answers that are sometimes completely wrong, and rely quite heavily on known structures to act as models. And even if we know how a protein folds, we are very bad at predicting its function, except in general terms. Even given an angstrom-resolution structure for a restriction enzyme, for instance, we can't tell what its recognition sequence is. We're even worse at predicting protein-protein interactions, where all the action is. |
|
| |
Once we solve problems 1 and 2, there's a list of others waiting in the wings. |
|
| |