Re: 2nd law clarifications

On 4/01/2025 3:05 am, Ernest Major wrote:

On 03/01/2025 13:08, MarkE wrote:

>
Wrong. Near-neutral (i.e. mildly detrimental changes) by definition have a very low selection co-efficient and therefore typically will not be removed by selection.

 As an allele approaches a selection coefficient of zero the chance of fixation (and therefore also elimination) approaches 50%. The chance of a mildly detrimental change being removed (by the combination of drift and selection) remains greater than 0.5.
 If you want to consider genetic drift acting alone, you have to not only eliminate natural selection, but also mutation and other sources of variation. In this case a population will evolve to eliminate genetic variation and then enter stasis.
 

Population genetics is an interesting and difficult area (not necessarily disputing your statement above).
In pondering this issue, I wrote a program to simulate a population (~10,000) subject to sexual reproduction, with chromosomes, recombination etc.
I realised that it was in part a kind of reverse-engineering investigation of what distributions of selection coefficients would lead to fixation, extinction etc. E.g., empirically estimated distributions concentrate the majority of mutations just less than 0 (near-neutral deleterious), some as -1 (lethal), and a very small proportion just above 0 (near-neutral beneficial).
I'm still working on it. Tricky to be sure that assumptions and simulation are valid.