Monday, March 09, 2009
I want to be a science writer
A doubting Thomas asks about how it is that random selection on billions of genes can replace intelligent design. A few thoughts follow, some culled from Mr. R. Dawkins, which it would be interesting to have corrected by someone who knows a bit more about it than I do.
Natural selection is a better name than random selection as the selection is not random at all. It is the mutations which might be random.
There might be a lot of genes, but there has been a lot of natural selection. There have been genes on earth for more than a billion years and in that time there have been many trillions of individuals and many trillions of mutations. Only a small proportion need to work and so to be selected.
Human aside: by way of comparison, we have been around for only a few hundred thousand years.
Technical aside: you get the sort of genes you get in people in all eukaryotes. Eukaryotes are all those things made up of one or more cells of the sort you used to learn about in GCE O level biology. Eukaryotes have been around for more than a billion years. Things which are not eukaryotes are things like bacteria, viruses and the strange (small) things that live in volcanic fountains at the bottom of oceans.
The genome can be thought of as being structured rather like a computer program, a tree or a hierarchy. And whole branches of the tree tend to get moved around as a whole. So once natural selection has come up with, say a liver, it sticks with it. So the liver is much the same in all animals that have one. This reduces the amount of work that natural selection has to do.
One might perhaps think of cars. Different makes of car can look very different, but they might all be largely made from the same set of parts. It is only the trim that is a bit different. Or meccano or lego.
Not only are the big parts the same in most large animal life, but the basic, molecular mechanisms of life, the basic building bricks, are the same in most life altogether. A large part of the genome is about specifying or enabling these basic mechanisms. This large part of the job of design only had to be done once. And it was done with small early life which involved a very long time and a very large number of individuals. Once this was cracked, big later life was easy.
A large proportion of mutations in the genome will fatally damage these basic molecular mechanisms and the individual concerned will not get very far. Nature does not waste much time, space or energy on failure. You only get to see the more sensible ideas.
A large proportion of the genome is waste paper. Doesn’t do anything at all. So genomes are not as big as they look.
Another thought from the computer analogy, is that nature can try things out in simple animals before trying them out in complicated ones. Get the liver right in a worm before trying it out in a lion. And given that, in very round terms, life evolves from the simple to the complicated, testing by evolution is efficient in this sense. Just the same as in an IT shop.
Natural selection is not that clever. It often results in bad designs, the mammalian eye being a good example. One might have thought that an intelligent designer would have done better.
Contrariwise, human selection is not that fast. It has taken a 100 years to bring the design of a car to the pitch that it is now at. So maybe faster than natural selection, but cars did not pop out of nowhere, any more than dinosaurs did.
There is plenty of evidence about the detail of evolution and the workings of natural selection. The evolution and spread of the colour of newts across North America or whatever. Evolution does work.
Not part of the argument here, but it might help to think of a genome being made up of genes, each one of which is a recipe for a particular protein, with a switch. The switch might be turned on by some other protein being present or turned off by yet another. Add a few bells and whistles, and you get a device which could be used to make anything; rather in the way that a Turing machine can replicate any sensible computer algorithm.
Natural selection is a better name than random selection as the selection is not random at all. It is the mutations which might be random.
There might be a lot of genes, but there has been a lot of natural selection. There have been genes on earth for more than a billion years and in that time there have been many trillions of individuals and many trillions of mutations. Only a small proportion need to work and so to be selected.
Human aside: by way of comparison, we have been around for only a few hundred thousand years.
Technical aside: you get the sort of genes you get in people in all eukaryotes. Eukaryotes are all those things made up of one or more cells of the sort you used to learn about in GCE O level biology. Eukaryotes have been around for more than a billion years. Things which are not eukaryotes are things like bacteria, viruses and the strange (small) things that live in volcanic fountains at the bottom of oceans.
The genome can be thought of as being structured rather like a computer program, a tree or a hierarchy. And whole branches of the tree tend to get moved around as a whole. So once natural selection has come up with, say a liver, it sticks with it. So the liver is much the same in all animals that have one. This reduces the amount of work that natural selection has to do.
One might perhaps think of cars. Different makes of car can look very different, but they might all be largely made from the same set of parts. It is only the trim that is a bit different. Or meccano or lego.
Not only are the big parts the same in most large animal life, but the basic, molecular mechanisms of life, the basic building bricks, are the same in most life altogether. A large part of the genome is about specifying or enabling these basic mechanisms. This large part of the job of design only had to be done once. And it was done with small early life which involved a very long time and a very large number of individuals. Once this was cracked, big later life was easy.
A large proportion of mutations in the genome will fatally damage these basic molecular mechanisms and the individual concerned will not get very far. Nature does not waste much time, space or energy on failure. You only get to see the more sensible ideas.
A large proportion of the genome is waste paper. Doesn’t do anything at all. So genomes are not as big as they look.
Another thought from the computer analogy, is that nature can try things out in simple animals before trying them out in complicated ones. Get the liver right in a worm before trying it out in a lion. And given that, in very round terms, life evolves from the simple to the complicated, testing by evolution is efficient in this sense. Just the same as in an IT shop.
Natural selection is not that clever. It often results in bad designs, the mammalian eye being a good example. One might have thought that an intelligent designer would have done better.
Contrariwise, human selection is not that fast. It has taken a 100 years to bring the design of a car to the pitch that it is now at. So maybe faster than natural selection, but cars did not pop out of nowhere, any more than dinosaurs did.
There is plenty of evidence about the detail of evolution and the workings of natural selection. The evolution and spread of the colour of newts across North America or whatever. Evolution does work.
Not part of the argument here, but it might help to think of a genome being made up of genes, each one of which is a recipe for a particular protein, with a switch. The switch might be turned on by some other protein being present or turned off by yet another. Add a few bells and whistles, and you get a device which could be used to make anything; rather in the way that a Turing machine can replicate any sensible computer algorithm.
# posted by Jim @ 11:50 AM 
