Also Available in:

Critic ignores reality of Genetic Entropy

The author of a landmark book on genomic decay responds to unsustainable criticisms.

by John Sanford

Published: 7 March 2013 (GMT+10)

I do not normally spend my time responding to bloggers, but several people have asked me to respond to Scott Buchanan’s polemic1 against my book Genetic Entropy. This article is a one-time clarification as I cannot afford the time to be drawn into the blog-o-sphere and its associated ‘death by a thousand emails’.

Scott’s lengthy essay is certainly not an objective review of my book; it is an ideological attack based upon a commitment to the standard Darwinian theory. He does not acknowledge any of the legitimate concerns I raise regarding the Darwinian process, not even the many points widely acknowledged by my fellow geneticists. Shouldn’t even ardent Darwinists honestly acknowledge the known problems with current Darwinian theory? I will only briefly touch on each of his arguments.

Scott gives his arguments in this order:

  1. First, he claims, that my book is all about deliberate deception, and I am fundamentally a liar (but perhaps I am otherwise a nice Christian man).
  2. He then spends three pages expressing how annoyed he is with the exact way in which I cite a Kimura reference—as I try to make clear the actual distribution of mutational effects.
  3. He argues that, while beneficial mutations are rare, they are not as rare as I suggest. Since beneficial mutations clearly happen, and since adaptation clearly happens, he imagines the Primary Axiom—that man is merely the product of random mutations plus natural selection—must then obviously be true.
  4. The Primary Axiom:

    Man is merely the product of random mutations plus natural selection.

  5. Scott cites a series of flawed ‘mutation accumulation’ experiments, which he thinks demonstrate extremely high rates of beneficial mutation.
  6. He points out that duplications can have real biological consequences.
  7. He also points out that we cannot generally see measurable degeneration in extended lab experiments. He argues that, if I were correct, then in just the last few thousand years all forms of life having short life cycles (bacteria, mice) should have gone extinct. Since we do not see obvious degeneration happening, the Primary Axiom must be true, according to him.
  8. Scott suggests that, since human life spans have increased in the last several centuries, this proves that man is not degenerating.
  9. He argues that Crow’s conclusion that the human race is presently degenerating at 1–2% per generation2 does not mean he stopped being faithful to the Primary Axiom.
  10. He also argues that synergistic epistasis3 really happens (at least to some extent), and cites those who feel this might aid in mutation elimination.
  11. Finally, he casually dismisses all the papers I cite in Appendix 1 of my book, where the leaders of the population genetics field acknowledge all of the basic problems with the Primary Axiom.

It seems to me that Scott makes his arguments in the wrong order, starting with the trivial, and at the end simply waving off the most crucial issues, so I wish to address his points in reverse order:

10. What other geneticists say:

Let me begin by going to the very end of my book (Appendix 1), where I quote key papers written by the leading experts within the field of population genetics. Scott refers to this as my “final shotgun-blast of misrepresentation to the gullible reader”. This seems grossly unfair, since I am simply quoting the leaders in the field where they acknowledge major aspects of my thesis. In my introduction to that section, I am careful NOT to imply that those scientists would agree with my personal viewpoint, but I point out that they all very clearly acknowledge the major problems which I outline in my book regarding the Primary Axiom.

An artistic representation of mutational meltdown

These experts in the field provide strong support for all the main points of my book …

  1. Is man presently degenerating genetically? It would seem so, according the papers by Muller, Neal, Kondrashov, Nachman/Crowell, Walker/Keightley, Crow, Lynch et al., Howell, Loewe and also myself. Scott suggests this is foolishness and dismisses the Crow paper (1–2% fitness decline per generation). But Kondrashov, an evolutionist who is an expert on this subject, has advised me that virtually all the human geneticists he knows agree that man is degenerating genetically. The most definitive findings were published in 2010 in the Proceedings of the National Academy of Science by Lynch.4 That paper indicates human fitness is declining at 3–5% per generation. I personally feel the average mutational effect on fitness is much more subtle than Lynch does—so I think the rate of human degeneration is much slower than he suggests—but we at least agree that fitness is going down, not up. Can Scott find any qualified geneticist who asserts man is NOT now degenerating genetically? There is really no debate on current human genetic degeneration—Scott is 100% wrong here, and is simply not well informed.

  2. Is there a theoretical problem associated with continuously growing genetic load due to subtle un-selectable deleterious mutations? Yes, according to Muller, Kondrashov, Loewe, and many others. Population geneticists all seem to acknowledge the fact that a large fraction of deleterious mutations are too subtle to be effectively selected away. The question is, what is that fraction? At what point does the fitness effect of a deleterious mutation become too small to be selected away? I have been studying this for about 7 years. Our numerical simulations indicate that for higher organisms, up to 90% of all deleterious mutations should be un-selectable (Chase W. Nelson and John C. Sanford, 2013, Computational Evolution Experiments Reveal a Net Loss of Genetic Information Despite Selection, Biological Information: New Perspectives pp. 338–368; https://doi.org/10.1142/9789814508728_0014). This manuscript was previously sent to Scott, but it seems he did not read it. Can Scott explain away this theoretical problem?

  3. What is Dr. Ohta’s view on genetic degeneration? Dr. Tomoko Ohta was a key student of Kimura, and published extensively with Kimura. Dr Ohta came to be known as the ‘Queen of Population Genetics’, and is now an honorary member of the American Academy of Arts and Sciences, and an associate of the National Academy of Sciences, USA. She is the world’s authority on the topic of near-neutral mutations. One of my co-authors went to Japan to spend several days discussing with her a manuscript in which we used numerical simulation to clearly demonstrate that near-neutral deleterious mutations generally escape selective removal and lead to continuous and linear accumulation of genetic damage. She acknowledged that our numerical simulations appeared to be valid, and that our conclusions appeared to be valid. This clearly reflects a profound evolutionary paradox (it is the same paradox Kondrashov addressed in his paper “why have we not died 100 times over?”5). When asked about synergistic epistasis, she immediately acknowledged that synergistic epistasis should make the problem worse, not better, just as I argue in my book. Using numerical simulations, we have confirmed that synergistic epistasis fails to slow mutation accumulation and accelerates genetic decline (Can Synergistic Epistasis Halt Mutation Accumulation? Results from Numerical Simulation; doi.org/10.1142/9789814508728_0013). I think Dr Ohta would like me to clarify that she is a faithful Darwinist and remains committed to the Primary Axiom, and that she is in fact hostile to the thesis of my book.

  4. The other quotes: I encourage Scott to read all the other quotes in the appendix. It is clear that the leading population geneticists have recognized major theoretical problems with the Primary Axiom for a long time. Why try to deny this?

9. Synergistic Epistasis:

Scott spends a lot of time selling synergistic epistasis. What he may not realize is that population geneticists almost universally understand that most mutations interact either additively or multiplicatively.

Suppose two mutations each reduce fitness 10%. When both mutations are present, then fitness might be reduced by 20% (additive interaction), or less than 20% (multiplicative interaction), or more than 20% (synergistic epistasis). To the extent synergistic epistasis happens, it obviously will accelerate degeneration! The only reason synergistic epistasis is even invoked in a generalized sense is when trying to argue that more genetic damage might somehow induce more effective selection (we can now clearly show that even if synergistic epistasis was normally true, it makes the degeneration problem worse). Population genetic theory was built on the understanding that genetic interactions are primarily additive or multiplicative. If synergistic epistasis was generally true (it is just a rare exception to the norm), most of the papers ever published in the field of population genetics would be invalid! Yet synergistic epistasis would have to be the norm before it would even be conceivable that it might enhance selection.

Naturally, all possible types of interactions occur in a complex genome, including instances of synergistic epistasis. Synergistic epistasis would be of little concern to population geneticists except for one thing; it is their last straw in the struggle to solve the degeneration problem. Careful consideration makes it clear that even if synergistic epistasis was the primary mode of interaction (rather than being a rare exception), it only makes the degeneration problem worse. I point this out in my book, and Ohta reaches the identical conclusion (see above). We have done extensive numerical simulations that show that this is true (https://doi.org/10.1142/9789814508728_0013).

8. Crow still believed in the Primary Axiom:

Scott spends a lot of time showing that even though Crow believed mankind is degenerating at 1–2% per generation, he remained faithful to the Primary Axiom. That is certainly true, and I made that clear in my book.

7. Human life span has recently been increasing:

It is obviously true that human longevity has increased in recent centuries, but that is not due to evolutionary advance. It is clearly due to improved diet, sanitation, and modern medicine. We have figured out how to keep people from dying in infancy and extended the life expectancy for those who catch many diseases associated with middle-age. Thus, the average has gone up. The maximum possible lifespan has not gone up. This is a simple concept.6

6. Genetic entropy is not obvious in lab experiments or in nature:

everystockphoto | svdmolenA scientist inspecting bacterial cultures growing in petri dishes

It is true that most lab experiments do not show clear degeneration. But Scott should realize that anything alive today must have been degenerating slowly enough to still be here, even in a young earth scenario. All three of the downward decay curves I show in my book indicate that degeneration slows dramatically as it becomes more advanced. If a species is alive today and has been around for thousands of years, the rate of degeneration must be very slow (too subtle to measure in most cases). Obviously, genetic degeneration is not going to be clearly visible in most lab experiments.

Regarding Scott’s argument about viruses and bacteria, such microbes should degenerate very slowly because mutation rate per genome is low, and selection is intense and continuous. Despite this, we have just published a paper showing that RNA viruses are clearly subject to genetic entropy.7 Another reason viruses (and bacteria) can persist in spite of genetic entropy is that they can be preserved in a dormant state for thousands of years. Therefore, even if most active strains continuously died out (say after a thousand years), new strains could be continuously reseeded into the environment from natural dormant reservoirs.

Regarding mice, our numerical simulations suggest organisms like mice should last longer than longer-lived mammals because they have lower mutation rates per generation and much more frequent cycles of selection. The first species to go extinct due to mutation accumulation should be large, long-lived organisms.8 Even if species are not actually degenerating, the question of what sustains them would remain. Careful analysis shows mutations/selection could not be that sustaining force.

5. Duplications have biological effects:

This is obviously true, but how is it relevant? Like the accidental duplications that happen in emails and student essays, duplications are almost universally deleterious. Very rarely, some are beneficial. A few rare beneficial duplications cannot offset the many accumulating deleterious duplications, let alone all the other accumulating mutations.9

4. Mutation accumulation experiments suggest extremely high rates of beneficial mutations:

Mutation accumulation experiments are a very poor way to understand deleterious mutation accumulation. Such experiments do not study actual mutations, they only study performance of strains (the supposed ‘mutations’ are only inferred). In the papers of this type I have examined, zero mutations are actually documented. All that is observed is differential performance of strains. Non-genetic causes, including epigenetic effects or gain/loss of viruses in some bacterial culture, etc., cannot be precluded. More to the point, since the overwhelming majority of mutations are very subtle and do not express a clear phenotype, almost all mutations will be invisible in these experiments, which only monitor gross differences in performance. Only high-impact mutations can be observed in such experiments, and these represent a biased sampling of the actual mutational spectrum. Furthermore, high-impact deleterious mutations will still always be selected away in such experiments, no matter how hard the experimenter tries to preclude natural selection. Therefore there will be a strong tendency to preferentially observe only high-impact beneficials. Since the crux of the genetic entropy argument involves the low-impact deleterious mutations (which will always be invisible in such experiments), these types of experiments have no relevance to this discussion. A final point: in these experiments, fitness is always narrowly defined (i.e., ability to grow on a given medium). For simple, one-dimensional traits like this, any genetic change affecting that trait has a reasonable chance of being beneficial (in a one-dimensional system, any change can only be either up or down, as opposed to improving a real-world complex network of traits where fitness is enormously multi-dimensional).

3. Just how rare are beneficial mutations?

Scott speaks as if I do not acknowledge there are beneficial mutations. I acknowledge them very openly in the book, but I also insist that beneficials must be very rare compared to deleterious mutations (as do nearly all geneticists). The critical question is “how rare?”

flickr.com/visulogik (CC by 2.5)A rusty old car

Genomes are the genetic specifications that allow life to exist. Specifications are obviously inherently SPECIFIC. This means that random changes in specifications will disrupt information with a very high degree of certainty. This has become especially clear ever since the publication of the ENCODE results, which show that very little of our genome is actually ‘junk DNA’.10 The ENCODE project also shows that most nucleotides play a role in multiple overlapping codes, making any beneficial mutations which are not deleterious at some level vanishingly rare (https://doi.org/10.1142/9789814508728_0006). Our own numerical simulations (in press) show that that unless beneficial mutations are extremely common, they are not sufficient to compensate for accumulating deleterious mutations. The bottom line is that selection removes only the worst deleterious mutations and amplifies only the best beneficial mutations. This means that the accumulating damage is largely invisible (like rust on a car), while adaptations tend to be highly visible (e.g., antibiotic resistance). This means that even if Scott presents us with 1000 examples of adaptation via beneficial point mutation, he has still failed to address the key issue—net gain versus net loss. Adaptation explains fine-tuning to an environment; it does not explain the astounding internal workings of life. It does not begin to explain the mystery of the genome.

Where are the beneficial mutations in man? It is very well documented that there are thousands of deleterious Mendelian mutations accumulating in the human gene pool, even though there is strong selection against such mutations. Yet such easily recognized deleterious mutations are just the tip of the iceberg. The vast majority of deleterious mutations will not display any clear phenotype at all. There is a very high rate of visible birth defects, all of which appear deleterious. Again, this is just the tip of the iceberg. Why are no beneficial birth anomalies being seen? This is not just a matter of identifying positive changes. If there are so many beneficial mutations happening in the human population, selection should very effectively amplify them. They should be popping up virtually everywhere. They should be much more common than genetic pathologies. Where are they? European adult lactose tolerance appears to be due to a broken lactase promoter [see Can’t drink milk? You’re ‘normal’! Ed.]. African resistance to malaria is due to a broken hemoglobin protein [see Sickle-cell disease. Also, immunity of an estimated 20% of western Europeans to HIV infection is due to a broken chemokine receptor—see CCR5-delta32: a very beneficial mutation. Ed.] Beneficials happen, but generally they are loss-of-function mutations, and even then they are very rare!

Scott makes a big deal about Lenski’s long-term bacterial experiments, but these actually support my thesis. Although a very trivial adaptation happened (optimal growth on a given medium), his bacteria shrank in genome size (the functional genome decreased). Evidently the more rapid growth was largely accomplished through genetic degeneration. Many useful genes not essential in that artificial environment were apparently lost. When transferred to a natural environment, those highly degenerated bacteria would essentially be dead-on-arrival.

Scott seems to think that as long as beneficials happen (regardless of how rare they are) the Primary Axiom must be true. Likewise he thinks that, as long as he can show selective adaptations happen (no matter how trivial), this proves the Primary Axiom. I do not think he grasps just how difficult it is to build a genome apart from design. Nor does he seem to understand that a population can be undergoing genetic decline due to vast numbers of slightly deleterious mutations even while selection may be amplifying a handful of beneficial mutations. He seems to fail to realize that a species can undergo minor adaptive fine-tuning to its environment even while degenerating in many other ways. So let me try to make it even simpler. Picture a ten year old car. It is degenerating in all possible ways. Install new windshield wipers. Has the car stopped degenerating? There has certainly been an improvement, but not the type of improvement that can reverse the ubiquitous and systematic degeneration.

In collaboration with other scientists, we have advanced the field of population genetics by developing the ‘state of the art’ in terms of numerical simulation of the mutation/selection process.11 Using biologically realistic parameters, the program ‘Mendel’s Accountant’ consistently shows genetic decline even given very generous rates of beneficial mutation. This strongly validates my book. Mendel’s Accountant cannot tell us the true history of life, but what it can do is tell us what selection can and cannot realistically do in the present.

from Genetic Entropy pages 31 and 32Kimura mutations diagram

Kimura diagram with beneficial mutations added

2. Kimura’s Figure:

Scott makes a huge deal about my reference to a figure in Kimura’s work. He misrepresents me by arguing I misrepresented Kimura (I did not claim Kimura agrees with me). But this is a rabbit trail; the argument is not about Kimura. The crucial issue is about defining the correct distribution of mutation effects. For deleterious mutations, Kimura and most other population geneticists agree the distribution is essentially exponential. Figure 3c in my book (based upon Kimura) shows an exponential-type distribution of deleterious mutations, with most deleterious mutations being ‘nearly-neutral’ and hence un-selectable (effectively neutral). But, as I point out, Kimura’s picture is not complete, because degeneration is all about the ratio of good to bad mutations. Kimura does not show the beneficial distribution, which is essential to the question of net gain versus net loss! When I show the beneficial distribution (while Kimura did not do this, I suspect he would have drawn it much as I did), anyone can see the problem: the vast majority of beneficial mutations will be un-selectable (Figure 3d). Scott does not appear to contest my representation of the mutational effect distribution, which is the main issue here. Scott should easily be able to see that most mutations fall within the ‘no-selection zone’ and that almost all of them are deleterious. So even with strong selection, this entire zone can only undergo degeneration. Outside this zone, the substantially bad mutations will be selected away, and an occasional rare high-impact beneficial will be amplified (which can explain isolated events such as antibiotic resistance).

1. Sanford is a liar:

Scott repeatedly asserts that my book is all about deliberate deception, and I am fundamentally a liar. He bases this upon two things: a) there were a few references he thinks highly relevant, which I failed to cite and which he says proves I have withheld and suppressed evidence; b) He argues I must surely know that beneficial mutations happen, that natural selection happens, and also that long term lab experiments do not show rapid degeneration. Therefore, I must be dishonestly pretending to be ignorant of these things in order to deceive the ignorant. He has not considered these possibilities: a) given the mountain of relevant literature, I might legitimately miss a few papers; b) I do not share his view on which papers are significant. He cites a great many papers which only speak of the obvious: beneficials do happen, selection does happen, adaptation does happen. Any high school student knows these things. My argument only begins AFTER acknowledging these obvious things.

Scott and I corresponded briefly before his posting, and I tried to explain to him why his criticisms were not correct. I did not find him to be a very good listener as I tried to explain how he was misrepresenting me. I then sent him a series of preprints (in press), which extensively and conclusively addressed all his objections. Upon reading his essay now, I can see he did not bother reading those preprints, which are very rigorously written scientific research papers. I also see from his current arguments, that he really did not give my book a fair read. If Scott has misrepresented both the book and myself, then which of us is lacking in integrity?

This book cost me a great deal. I basically laid down my reputation and my career in order to say what I believe to be the truth. I believe the real deception is clearly the Primary Axiom. I am still convinced I can persuade any impartial person that the Primary Axiom is indefensible (if they will listen). So why would I lie? I am a sincere orthodox Christian, I believe God will judge me in a very literal sense, and I consider lying is a very serious sin. I am distinguished in my field and I greatly value my integrity as an honest scientist. Yet my integrity as a Christian is much more important to me than my scientific standing. That is why I have been willing to defend what I believe to be true, even knowing that attacking this sacred cow (the Primary Axiom) would bring slander and scorn. Why would I write a book that would ruin a very good scientific reputation knowing it would make me a liar before God?

In our personal correspondence, Scott closed our conversation saying he intended to present me as being intentionally deceitful. My last word to him was that while I might be technically in error on certain points, my book reflects what I really believe to be true. Any technical errors in my book show that I am human, but there certainly was no deliberate deception in my book. In light of our previously open exchange, and since Scott professes to belong to Jesus, I do not think he should slander a Christian brother in this way, and I believe he should apologize and withdraw these personal attacks. In terms of the scientific issues, I would ask Scott to append this response to his blog attack.

I still welcome any fair-minded and balanced analysis of the scientific merits of my book and my subsequent studies.


John Sanford


  1. The original critical blog post can be found at letterstocreationists.wordpress.com/stan-4. Return to text.
  2. Crow, J., The high spontaneous mutation rate: Is it a health risk? Proceedings of the National Academy of Sciences 94(16):8380–8386,1997; pnas.org/content/94/16/8380.full. Return to text.
  3. Synergistic epistasis is a theoretical construct where the combined effect of mutations is greater together than the sum of their individual effects. This is obviously a good situation for beneficial mutations, but very bad for harmful mutations. See Doyle, S., The diminishing returns of beneficial mutations, Journal of Creation 25(3):8–10, 2011. Return to text.
  4. Lynch, M., Rate, molecular spectrum, and consequences of human mutation, Proceedings of the National Academy of Sciences 107(3):961–968, 2010.Return to text.
  5. Kondrashov, A., Contamination of the genome by very slightly deleterious mutations: why have we not died 100 times over? Journal of Theoretical Biology 175(4):583–594, 1995. Return to text.
  6. See Carter, R., Genetic entropy and human lifespans: If the human genome is degrading, shouldn’t lifespans be getting shorter? 15 July 2012. Return to text.
  7. Carter, R. and Sanford, J., A new look at an old virus: patterns of mutation accumulation in the human H1N1 influenza virus since 1918, Theoretical Biology and Medical Modelling 9(42):1–19, 2012. Return to text.
  8. See Carter, R., Genetic entropy and simple organisms: If genetic entropy is true, why do bacteria still exist? 25 October 2012. Return to text.
  9. See Liu, Y. and Moran, D., Do new functions arise by gene duplication? Journal of Creation 20(2):82–89, 2006. Return to text.
  10. See Williams, A., Astonishing DNA complexity uncovered, 20 June 2007, and Astonishing DNA complexity update, 3 July 2007. Return to text.
  11. See Sanford, J. and Nelson, C., The Next Step in Understanding Population Dynamics: Comprehensive Numerical Simulation, Chapter 7 (pages 117–135) in: Carmen Fusté, M. (Ed.), Studies in Population Genetics, InTech, 2012. Return to text.

Readers’ comments

Comments are automatically closed 14 days after publication.