Nelson and Sanford (2013) argued that a net loss of “genetic information” occurs during evolution despite natural selection. They tried to show that a software named Avida operates under faulty premises:
Recently, […] experiments using the digital genetics software Avida have suggested that purifying selection can be extremely effective and that novel genetic information can arise via selection for high-impact beneficial mutations. […] In this study, we investigate why Avida and Mendel’s Accountant yield seemingly contradictory results. We find that most discrepancies are due to differences in default settings. Mendel’s default settings implement values plausible for modeling the human species, while Avida’s default settings have virtually no parallel in biological systems. Additionally, Avida introduces several un-biological mechanisms both for facilitating the development of novel genetic information and for preventing its loss.
I am trying to figure out where they went wrong. As a side note, Basener and Sanford recently published a critique of Fisher’s fundamental theorem of natural selection in the Journal of Mathematical Biology.
From their concluding remarks:
After we re-formulated Fisher’s model, allowing for dynamical analysis and permitting the incorporation of newly arising mutations, we subsequently did a series of dynamical simulations involving large but finite populations. We tested the following variables over time: (a) populations without new mutations; (b) populations with mutations that have a symmetrical distribution of fitness effects; and (c) populations with mutations that have a more realistic distribution of mutational effects (with most mutations being deleterious). Our simulations show that; (a) apart from new mutations, the population rapidly moves toward stasis; (b) with symmetrical mutations, the population undergoes rapid and continuous fitness increase; and (c) with a more realistic distribution of mutations the population often undergoes perpetual fitness decline.
Has Genetic Entropy finally gained acceptance in the scientific community?
The first clue that there is something very, very fishy about this paper is the number of quotations it contains.
The second clue is the number of straw-man fallacies. The biggest howler is the idea that fitness should necessarily increase. How can this possibly be, when most of the species that have ever existed have become extinct?
The main problem is they tried to use science to support their Biblical YEC beliefs. Sanford’s “genetic entropy” idea is based on the unsupported and frankly ridiculous claim the human genome was created “perfect” only 6000 years ago and has been degrading since “The Fall”.
Sanford’s “Mendel’s Accountant” program was a huge joke. Sanford wrote it so no matter what starting parameters one used the population being modeled quickly crashed. One of the funnier things I saw was when people used it to model the human population starting with the supposed 8 survivors from Noah’s Ark and humans all died out within a few centuries. Oops!
I don’t understand PS. This thread seems not to be in the list of threads, but then I don’t understand those, especially all the ones that get “bumped”. Anyway:
The fundamental observation that the Genetic Entropy crowd point to is that measured human mutation rates are very high. The most recent ones I have seen are 60 mutations per diploid newborn. Andrey Kondrashov has been pointing to these high rates with alarm for some time. Some of these land in junk DNA, of which there is quite a lot (sorry, creationists, it’s true in spite of the ENCODE foolishness). There are classical mutational load calculations that predict that the number of deaths or sterility caused by each mutation is about 1. So if more than 1/60 of all DNA is nonneutral, there should be a heavy mutational load, more than 1 genetic death per individual per generation. Possible explanations (1) there are fewer nonneutral sites than we think, (2) the mutation rate is historically higher in industrial civilizations than it was in the past, and we have not yet reached the equilibrium that would impose that heavy mutational load. Possible consequences that are simply silly: (1) therefore we can’t have been around more than 6,023 years, and (2) The End Is Nigh (sorry not to be Peaceful). As for “accepting Genetic Entropy”, no, if it means that we know that all species are having this problem, and no, if it means that we know that this is a normal long-term problem. Neither of those is established. It seems to be true that there is an alarmingly high and rising rate of fertility problems in industrial societies, and gynecologists are making as much noise about this as they can. But it is not established that these are due to genetic mutations (as opposed to environmental chemicals).
I will be happy to explain the logic of the Haldane (1939) - Muller (1950) mutational load calculation, if anyone needs to understand that.
I was addressing the issue of Genetic Entropy in general. As for Basener and Sanford’s paper, there have been some threads on it at TSZ. There will be more on this – suffice it to say that their calculations have little to do with Genetic Entropy.
Does that really matter for the point being made? There is an immense amount of purifying selection happening in utero, so that largely solves the paradox. Right?
The question is not whether individual mutations come to equilibrium frequencies, but whether the resulting reduction in mean fitness of the population is too great for it to bear.
Let’s try to distinguish the latter from the latter, or the former from the former, or whatever. When we measure the number of offspring per parent, we could either count all those conceptions, or just the ones that are born. We then, correspondingly, either include those chromosomal aberrations, or not. The simplest way is to not include them, but to count them as reducing the reproductive excess. The issue that Alex (not, as I mistakenly said, Andrey) Kondrashov has raised is whether the mutational load is becoming too great for the population to be avoid declining to extinction.
The difficulty is that you cannot tell anything about the absolute fitness of the population by counting how many deleterious mutations they carry on average. Partly because much selection is soft and affects only relative fitness, not absolute fitness, but also because at most the genetic load tells you how much less fit the population is than some ideal population with zero deleterious mutations, a population that has never existed.
From the Journal of Mathematical Biology’s submission guidelines…
Suggesting / excluding reviewers
Authors are welcome to suggest suitable reviewers and/or request the exclusion of certain individuals when they submit their manuscripts. When suggesting reviewers, authors should make sure they are totally independent and not connected to the work in any way. It is strongly recommended to suggest a mix of reviewers from different countries and different institutions. When suggesting reviewers, the Corresponding Author must provide an institutional email address for each suggested reviewer, or, if this is not possible to include other means of verifying the identity such as a link to a personal homepage, a link to the publication record or a researcher or author ID in the submission letter. Please note that the Journal may not use the suggestions, but suggestions are appreciated and may help facilitate the peer review process.
For those of you who publish, is it typical to choose your own reviewers? It strikes me as a rather loose peer review that can be gamed.
Well, I suppose it is safe to assume that Nelson and Sanford did not include yourself or prior posters John, Joe, Roy, or Steve, in their nominated reviewer list.
This is false. To see this, you just have to have a look at the concluding remarks made by Sanford and Basener in their paper on Fisher’s fundamental theorem of natural selection. By doing so, you will see that when using the same assumptions that Fisher made in his theorem (mutations have a symmetrical distribution of fitness effects), Mendel’s Accountant give the same results, ie, a net gain in fitness. This alone invalidates your claim that Sanford wrote Mendel’s Accountant so that « no matter what starting parameters one used the population being modeled quickly crashed ». And this alone supports the validity of MA for modeling genetic changes over time.
That doesn’t seem to follow. Just because MA will give a particular and expected result if it models a particular scenario, doesn’t mean it is correctly modeling other types of population genetic scenarios.
