You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
1. Continuing our weekend journal club inspired by @pastramimachine and @3rdreviewer, let's talk about the most (only?) compelling argument for the Neutral Theory: the cost of selection.
2. Although Kimura's take on Neutral Theory was ostensibly motivated by the observation of a roughly constant rate of protein evolution, it was his use of Haldane's cost of selection argument that seems to give it its strongest grounding (at least when I was growing up).
3. Lewontin also highlights this thinking in talking about polymorphisms w/in pops. There would seem to be too much change to allow that much selection on every site in the genome. Each individual can only die once so a population can only sustain so much selection.
4. Haldane's cost of selection as used by Kimura tracks a substitution from birth to eventual fixation. As @pastramimachine and @3rdreviewer point out, this assumes that selection (s) is constant over that time period. This is a somewhat teological approach.
5. They do a nice job in highlighting this paper by Felsenstein (1971), who shows that if you imagine evolution more in terms of a continuous process to a changing environment, then the "cost" at each increment is small relative to the overall change. https://www.journals.uchicago.edu/doi/abs/10.1086/282698 …
6. So, can we maintain enough selection at enough sites for the Neutral Theory to not be prima facie true? Seems plausible, but that's what others are going to weigh in on.
7. Substitution load is one way of looking at how much selection can be maintained. Another way is a more quantitative genetic point of view based on total variation in fitness.
8. Crow invented the "opportunity for selection" to measure this, which is essentially the variance in relative fitness. Selection can not be greater than the total variance in relative fitness (normalized absolute fitness). http://www.jstor.org/stable/41449168?casa_token=zUPi1ZCbcw4AAAAA:CysM0GWbm4JCppk0MUpd2ugKJGY8MwvUj3mXvHYehVian34IkfQk3nW_4kTrUHvHsH8vyni7KX3GeNKk5ExNldiQKAHxjqK1ArI1hx5bBYu92z9WD-U&seq=1#page_scan_tab_contents …
9. The total amount of selection that can generated per generation therefore depends on the covariance between a trait and fitness, otherwise known as selection. Using standard notation, this would look like Var[w] = Sum[p alpha^2], with allele freq p and effect/selection alpha.
10. The sum is over all loci/alleles. Since we know that even very small selection coefficients can generate meaningful evolution given enough time (say s=0.0001), s^2 will be very small, and we can have a great many loci under selection without depleting the opportunity for sel
11. Interesting are the observation of @PetrovADmitri & Paul Schmidt and others of periodic very strong seasonal selection. Note that substitution is not usually the outcome here, so there can be selection without the total "cost".
12. Others have used dN/dS ratios to provide genome-wide estimates of total selection and found that fraction to be surprisingly high.https://academic.oup.com/mbe/article/26/9/2097/1196708 …
13. The rise of the neo-selectionist viewpoint, as exemplified in @pastramimachine and @3rdreviewer is a response to all of these observations. So if there is a true cost of selection, as envisioned by Kimura, then something doesn't mesh. Many would say that thing is the NT. Fin.
Also relevant: Ewens argued early on that (1) Kimura's calculations were off and selection didn't create as much load as he claimed and (2) neutral substitution creates substitution load too so NT didn't really solve the problem K claimed existed https://www.jstor.org/stable/2459776
Correct me if I am wrong. Ewens wasn't really talking about population density and its control. Selection, except sexual selection, should always affect population growth. If selection is hard, it's going to make populations extinct
I think Ewens' main point was that Kimura's assumptions were off, leading him to over-calculate how much load would be generated by observed substitution rates. I associate the density/soft selection argument with Wallace.
Wallace invented the terms hard and soft selection, but had nonsensical definitions. I preferred Christiansen's 1975 in https://www.jstor.org/stable/2459633 . My point is selection always affects pop growth, i.e. is hard
And if selection has real demographic consequences, then the cost of selection problem will apply to neo-selectionism as much as it did to early selectionists. There's a problem.
I'm way out of my depth here. It seems to me that logic of Ewens, Felsenstein, et al. should be equally applicable to neo-selectionism. What say you, @3rdreviewer @pastramimachine?
Yes, I'm out of depth too, but I think there's still a debate. I just checked out from library a long paper by Ewens 1977 summarizing his arguments. I need to read that and Maynard Smith and Felsenstein too. Weissman Barton paper K&H cite doesn't seem to deal w/ load
Compelling?!?!? I’m gonna enjoy this thread
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.
