One thing that often used to happen, perhaps not a lot any more, is that individuals will say that we don’t need to worry about degrees of selection because all selection can be reduced to selection performing directly on genes. All I can tell this is GAHHHH! I think a complete lot of people know that you cannot think of selection as acting on genes, but a lot of people also can’t articulate why it doesn’t work. So, if anybody asks you, the easy answer is that reductionism doesn’t work because of interactions. At the individual level this will be gene interactions of dominance and epistasis primarily.
In a fully additive system there would be no problem, and this May be the problem. Our intuition about genetics was developed using simple additive models. Within an additive system, knowing at what level selection was acting would be nice information, however the fitness of the phenotype can continually be algebraically reduced to fitness effects on individual loci.
- Make sure you are eating enough (see Concern 1)
- Do your quest before you go
- Has the longest battery pack life, enduring 10 days on a single charge
- 8 oz cut grilled chicken
- 2 cups summer squash, shredded
In other words, in additive systems, how the genes are packed really doesn’t have an effect on the effect of genes on the phenotype. I take advantage of Z to stress that we are not talking about fitness. The effect of the allele on the phenotype is not affected by the packaging. The average impact in the haploid system differs than in the diploid system now, . In other words, if we add the simplest possible form of nonadditivity the packaging does matter.
Trust me it gets worse. I am way to lazy to put on furniture for average results in epistatic systems, but I before have talked about this. It turns out that the variance in local average effects is a measure of the way the average ramifications of alleles are to genetic background. The key point is that the variance in local average results is zero in additive systems but non-zero whenever there is any kind of interactions.
This means that the educability of fitness effects to genes is a reasonable exercise in an additive system, but is not meaningful in epistatically interacting systems simply. To observe how bad this is, consider long-term directional selection in a functional system with AXA epistasis. With regards to the starting gene frequencies the common effect of an allele can actually reverse signs. For what it is worth, the dashed lines are the local, average results for an additive system, and the solid lines will be the local average results for AXA epistasis. This shows the comparison between additive systems and epistatic systems.
For the additive system, if you were to evaluate the fitness results in generation no they would provide a pretty good estimation of the fitness by the end (in this deterministic system a precise estimate). Alternatively, for the epistatic system, estimations of allelic results manufactured in eras zero rapidly become worthless, and by enough time fixation are reached they are exactly incorrect. In a single sense, Williams is correct absolutely. At any given instant it’s possible certainly, in principle, to do a least-squares regression analysis and assign fitness effects to individual loci.
However within a epistatically interacting system those fitness projects are ONLY good for the moment, or perhaps the generation, where the assignment is done. Those effects will change as gene frequencies change, and not just gene frequencies at the locus under study, but gene frequencies at any other loci as well.