This blog also offers integration with Flickr via the "Blogger (New)" Blog type.
This will allow you to make cool posts like this one using Flickr, and include your photographs.
Also, just to show off my mad photography skills and our handsome patriarch.
Monday, January 29, 2007
Friday, January 26, 2007
Good Apples vs Bad Apples
OK, I'm in. Cool having your own blog. Of course, to be really cool, you have to talk about cool things, and say cool stuff. What could that be?
I have a question for Art, or Greg, or anybody else interested. It is regarding evolution, genes, and mutations and such. I'm not sure how to put it in words, but I'll try. Have you ever seen any work where people try to figure out the mathematical limits of evolution? I'm trying to answer the question of just how complex a genetic code could possibly be supported by evolution. Seems to me that there must be a mathematical limit, and I've never seen any work where people try to quantify that limit.
Evolution theory tells us that natural selection should favor "good" genetic mutations, while selectively eliminating bad ones, and the resulting positive selection pressure should push the gene pool toward the optimal solution. But it is not individual genes, but rather individual organisms which either procreate or not. So the good mutations and the bad mutations are all lumped together, and it’s all or nothing. This makes the selection process a very low resolution process. The bigger the genetic code, the lower the genetic resolution, and the harder for evolution to select good mutations while eliminating bad ones. If the genetic code of the individual increases in length, the resolution of the selection process decreases. Beyond a certain point, a certain level of genetic complexity, this limitation would prevent natural selection from removing bad mutations faster than they are introduced.
Imagine, if you will, that you are an apple farmer. Every year you harvest your apples, and select the best ones for your seed stock. By keeping good apples, and throwing away bad apples, you gradually remove bad genes, and your genetic strain improves. Of course, mutations are constantly occurring, and 99 out of 100 mutations are bad. But that is fine, because you individually select those bad apples and throw them out, and keep that wonderful 1 out of 100 good mutation. Soon you have the best apples in the land. Fine and dandy.
Now, imagine that your Ma and Pa Apple Farm grows to Wal-Mart proportions. Now, instead of selecting apples one at a time, you now have to select them one box at a time. Or, worse yet, one truckload at a time. You watch the trucks of apples drive by, and selectively choose which ones you will use as seeds for next years crops. The first truck has 10 bad apples in it. The next one has 12. The next, 20. Clearly, you pick the trucks with the fewest bad apples, and reject the trucks with the most. But since your resolution has decreased, you are much less effective at eliminating defects, and/or promoting improvements. The end result is that you are removing mutations from the gene pool at a slower rate. Meanwhile the mutation rate has not changed. Apples still mutate at the same rate as they always have. If this new mutation removal rate is less than the mutation injection rate, mutations will start to accumulate, and your apples will actually devolve rather than evolve.
So, with a given mutation rate, and a given probability of good vs. bad mutations, there is a mathematical limit to how big your truck size can be and still maintain your ability to remove bad mutations faster than they are introduced. There is a certain truck size at which the two rates are at equilibrium; beyond that, you actually go backward.
If the apple mutation rate were one mutation per 100 apples, and 9 out 10 mutations were "bad" mutations, then for each 1,000 apples, there would be 10 mutations, 9 bad and 1 good. If each box holds 1 apple, you could easily the sort the boxes by "goodness", keep the top half, and throw away the bottom half, and this would get rid off all the bad mutations, while keeping the good mutation. If each box holds 10 apples, you could probably do the same, but there is a fair chance that the good mutation may end up in a box with a bad mutation, and then you are either throwing away a good mutation, or keeping a bad mutation. As the box size grows, things get worse. If each box contains 100 apples, you now have 9 bad apples and 1 good apple spread randomly among ten boxes. With a truck size of 1000 apples, each truck averages 99 bad apples and one good apple. Now even the best trucks are full of defects, and you are removing them much slower that they are being introduced.
Consider, then, the human. I don't know how big the human genetic code is, but it's astronomical. Our "truck size" is mind-boggling. Let’s pretend, just for fun, that the mutation rate (the accumulation of all mutations of your genes from birth to procreation, due to radiation, free radicals, carcinogens etc.) is one in a billion. If the human genetic code were one billion genes long, then this would give an average of one mutation per person. If, however, the human genetic code were one trillion genes long, then this would give an average of 1000 mutations per generation. There is no way evolution could keep up in such a scenario. Even with perfect selection pressure (where only the most fit reproduce, and the less fit never "get lucky") the best you could hope for is to slow down the accumulation of defects.
I can't imagine that I'm the only one who has wondered about this. So, has anybody come across this question before? What is the answer provided by modern evolutionary theory? Surely, somebody has crunched the numbers. But I can't find it searching Google.
Art, you're the biology professor, help me out here. Please, save me from trying to figure this out myself.
I have a question for Art, or Greg, or anybody else interested. It is regarding evolution, genes, and mutations and such. I'm not sure how to put it in words, but I'll try. Have you ever seen any work where people try to figure out the mathematical limits of evolution? I'm trying to answer the question of just how complex a genetic code could possibly be supported by evolution. Seems to me that there must be a mathematical limit, and I've never seen any work where people try to quantify that limit.
Evolution theory tells us that natural selection should favor "good" genetic mutations, while selectively eliminating bad ones, and the resulting positive selection pressure should push the gene pool toward the optimal solution. But it is not individual genes, but rather individual organisms which either procreate or not. So the good mutations and the bad mutations are all lumped together, and it’s all or nothing. This makes the selection process a very low resolution process. The bigger the genetic code, the lower the genetic resolution, and the harder for evolution to select good mutations while eliminating bad ones. If the genetic code of the individual increases in length, the resolution of the selection process decreases. Beyond a certain point, a certain level of genetic complexity, this limitation would prevent natural selection from removing bad mutations faster than they are introduced.
Imagine, if you will, that you are an apple farmer. Every year you harvest your apples, and select the best ones for your seed stock. By keeping good apples, and throwing away bad apples, you gradually remove bad genes, and your genetic strain improves. Of course, mutations are constantly occurring, and 99 out of 100 mutations are bad. But that is fine, because you individually select those bad apples and throw them out, and keep that wonderful 1 out of 100 good mutation. Soon you have the best apples in the land. Fine and dandy.
Now, imagine that your Ma and Pa Apple Farm grows to Wal-Mart proportions. Now, instead of selecting apples one at a time, you now have to select them one box at a time. Or, worse yet, one truckload at a time. You watch the trucks of apples drive by, and selectively choose which ones you will use as seeds for next years crops. The first truck has 10 bad apples in it. The next one has 12. The next, 20. Clearly, you pick the trucks with the fewest bad apples, and reject the trucks with the most. But since your resolution has decreased, you are much less effective at eliminating defects, and/or promoting improvements. The end result is that you are removing mutations from the gene pool at a slower rate. Meanwhile the mutation rate has not changed. Apples still mutate at the same rate as they always have. If this new mutation removal rate is less than the mutation injection rate, mutations will start to accumulate, and your apples will actually devolve rather than evolve.
So, with a given mutation rate, and a given probability of good vs. bad mutations, there is a mathematical limit to how big your truck size can be and still maintain your ability to remove bad mutations faster than they are introduced. There is a certain truck size at which the two rates are at equilibrium; beyond that, you actually go backward.
If the apple mutation rate were one mutation per 100 apples, and 9 out 10 mutations were "bad" mutations, then for each 1,000 apples, there would be 10 mutations, 9 bad and 1 good. If each box holds 1 apple, you could easily the sort the boxes by "goodness", keep the top half, and throw away the bottom half, and this would get rid off all the bad mutations, while keeping the good mutation. If each box holds 10 apples, you could probably do the same, but there is a fair chance that the good mutation may end up in a box with a bad mutation, and then you are either throwing away a good mutation, or keeping a bad mutation. As the box size grows, things get worse. If each box contains 100 apples, you now have 9 bad apples and 1 good apple spread randomly among ten boxes. With a truck size of 1000 apples, each truck averages 99 bad apples and one good apple. Now even the best trucks are full of defects, and you are removing them much slower that they are being introduced.
Consider, then, the human. I don't know how big the human genetic code is, but it's astronomical. Our "truck size" is mind-boggling. Let’s pretend, just for fun, that the mutation rate (the accumulation of all mutations of your genes from birth to procreation, due to radiation, free radicals, carcinogens etc.) is one in a billion. If the human genetic code were one billion genes long, then this would give an average of one mutation per person. If, however, the human genetic code were one trillion genes long, then this would give an average of 1000 mutations per generation. There is no way evolution could keep up in such a scenario. Even with perfect selection pressure (where only the most fit reproduce, and the less fit never "get lucky") the best you could hope for is to slow down the accumulation of defects.
I can't imagine that I'm the only one who has wondered about this. So, has anybody come across this question before? What is the answer provided by modern evolutionary theory? Surely, somebody has crunched the numbers. But I can't find it searching Google.
Art, you're the biology professor, help me out here. Please, save me from trying to figure this out myself.
Me too
I think this is pretty cool - good idea Corey. I'd like to see this as a way we keep intouch with each other and let each other know what is going on in our lives ( am I asking too much?) Or, would this be a place for Trese's to write their very entertaining and spewing thoughts?
By the way, how are the triplets doing? Can we see any new pictures soon?
Dee
By the way, how are the triplets doing? Can we see any new pictures soon?
Dee
Thursday, January 25, 2007
Open Invitation to All "Trese Family" Members
This blog is open to anyone who is a Trese, considers themselves a Trese or in applying for membership to the Trese family.
Please contact 'cory.trese@gmail.com' for an invitation!
Please contact 'cory.trese@gmail.com' for an invitation!
The Draw of Social Internet Applications
why are people so inexorably drawn to social applications on the internet.
its like dad once said, people just love to watch, emulate, and be in the light so others will watch them. we watch reality TV so we can know what other humans are doing and what is accepted. the internet is a bigger reality TV show?
cool blog.
its like dad once said, people just love to watch, emulate, and be in the light so others will watch them. we watch reality TV so we can know what other humans are doing and what is accepted. the internet is a bigger reality TV show?
cool blog.
Subscribe to:
Posts (Atom)