Larry, I've finished reading the book. It should be required reading for every molecular biologist. Looking forward to major reviews in all the big journals and deafening silence from ENCODE people.
I do have one quibble, from page 94: "When the sequence similarity is high (greater than 30 percent or so), the region is said to be "conserved"..." This may be true for protein sequences, but for DNA sequences I doubt such a sequence would even be alignable. After all, under a Jukes-Cantor model 25% similarity is what you would get from random sequences, and other models would have higher levels of random similarity. I'd define "conserved" as "greater similarity than expected for neutrally evolving sequences in the same genomes". And a good approximation to "neutrally evolving" would be the average similarity over the whole genome. Thus the human-chimp comparison is on average 98.7% similar, while protein-coding exons average 99.5% similar, so we can say they're conserved.
Hi Professor Larry, for the genetic load argument, it's a very strong argument but the argument assumes all mutations that will occur in any functional DNA will be affected in the same way as the coding regions, is this assumption justified?
the [mutational load] argument assumes all mutations that will occur in any functional DNA will be affected in the same way as the coding regions Not really. It seems as if strongly deleterious mutations must impose a bigger mutational load than less deleterious ones. But Haldane (1939) and Muller (1950) showed that this isn't so. Strongly deleterious alleles are held to lower equilibrium frequency, and end up imposing the same load as less deleterious alleles. (As long as the selection coefficient against the deleterious alleles is less than the mutation rate to it).
6 comments :
Larry, I've finished reading the book. It should be required reading for every molecular biologist. Looking forward to major reviews in all the big journals and deafening silence from ENCODE people.
I do have one quibble, from page 94: "When the sequence similarity is high (greater than 30 percent or so), the region is said to be "conserved"..." This may be true for protein sequences, but for DNA sequences I doubt such a sequence would even be alignable. After all, under a Jukes-Cantor model 25% similarity is what you would get from random sequences, and other models would have higher levels of random similarity. I'd define "conserved" as "greater similarity than expected for neutrally evolving sequences in the same genomes". And a good approximation to "neutrally evolving" would be the average similarity over the whole genome. Thus the human-chimp comparison is on average 98.7% similar, while protein-coding exons average 99.5% similar, so we can say they're conserved.
@John
Good point! I was thinking of amino acid sequences but the context is clearly DNA similarity. My bad.
I'm not sure there will be deafening silence from ENCODE or Nature. I think I've been provocative enough to force a response. We'll see.
Hi Professor Larry, for the genetic load argument, it's a very strong argument but the argument assumes all mutations that will occur in any functional DNA will be affected in the same way as the coding regions, is this assumption justified?
Or can the functional DNA be uneven in the effect of mutations on it?
the [mutational load] argument assumes all mutations that will occur in any functional DNA will be affected in the same way as the coding regions
Not really. It seems as if strongly deleterious mutations must impose a bigger mutational load than less deleterious ones. But Haldane (1939) and Muller (1950) showed that this isn't so. Strongly deleterious alleles are held to lower equilibrium frequency, and end up imposing the same load as less deleterious alleles. (As long as the selection coefficient against the deleterious alleles is less than the mutation rate to it).
(As long as the selection coefficient against the deleterious alleles is less than the mutation rate to it).
Oops, should be "is greater than ..."
Post a Comment