DNA Mutation Rates and Evolution



Sean D. Pitman M.D.

©August, 2001

(Updated August 2008)







Mitochondrial DNA Mutation Rates


Mitochondria are “organelles” within living cells that are responsible for making the currency of energy called ATP (Adenosine Triphosphate), which all cells need to function.    Mitochondria carry their own separate DNA (mtDNA) that is independent of the nuclear DNA of the same cell.  Human mtDNA is composed of 37 genes totaling about 16,000 base pairs.  This mtDNA also mutates at a  much faster rate than nuclear DNA (nucDNA) does.  Human mtDNA has been completely mapped and all the coding regions are known (As well as the proteins or RNA for which they code).  Some of the mtDNA does not code for anything (thought to make these sections immune from “natural selection pressure”), and are known as the “control regions”. One particular region appears to mutate faster than any other region (1.8 times faster), because the variation among humans is greatest here.4 When the cell divides, each cell takes some of the mitochondria with it.  The mitochondria replicate themselves independently within the cell.  Beyond this, it has been generally assumed that mitochondria are always passed on from the mother to the offspring without being involved with genetic shuffling and recombining of mtDNA with the mtDNA of the father.  Recently, however, this notion has been challenged.  As it turns out, many cases of paternally derived mtDNA have been detected in modern families of humans as well as other species.   Consider the findings of an interesting study published by Schwartz and Vissing in the 2002 issue of the New England Journal of Medicine:


“Mammalian mitochondrial DNA (mtDNA) is thought to be strictly maternally inherited. Sperm mitochondria disappear in early embryogenesis by selective destruction, inactivation, or simple dilution by the vast surplus of oocyte mitochondria. . . The underlying mechanism responsible for the elimination of sperm mtDNA in normal embryos is not well understood. We speculate that the process in some cases may be defective, allowing sperm mitochondria to survive and giving those with a selective advantage the possibility of prevailing in certain tissues. . . Very small amounts of paternally inherited mtDNA have been detected by the polymerase chain reaction (PCR) in mice after several generations of interspecific backcrosses. Studies of such hybrids and of mouse oocytes microinjected with sperm support the hypothesis that sperm mitochondria are targeted for destruction by nuclear-encoded proteins. We report the case of a 28-year-old man with mitochondrial myopathy due to a novel 2-bp mtDNA deletion in the ND2 gene (also known as MTND2), which encodes a subunit of the enzyme complex I of the mitochondrial respiratory chain. We determined that the mtDNA harboring the mutation was paternal in origin and accounted for 90 percent of the patient’s muscle mtDNA.”47


So, what does such a finding mean with regard to mtDNA mutation rates and molecular clocks?  Well, consider the following comments by Morris and Lightowlers published in a 2000 edition of The Lancet:


Mitochondrial DNA (mtDNA) is generally assumed to be inherited exclusively from the mother. Several recent papers, however, have suggested that elements of mtDNA may sometimes be inherited from the father. This hypothesis is based on evidence that mtDNA may undergo recombination. If this does occur, maternal mtDNA in the egg must cross over with homologous sequences in a different DNA molecule; paternal mtDNA seems the most likely candidate. If mtDNA can recombine, irrespective of the mechanism, there are important implications for mtDNA evolution and for phylogenetic studies that use mtDNA. 48


Before this evidence of paternal inheritance was discovered it was assumed that mtDNA was strictly the result of maternal inheritance.  Based on this assumption, it was assumed that the mitochondrial offspring would get exact copies of the mitochondria that the mother had except if there was a mutational error.  This error rate in the non-coding portion of mitochondrial DNA has long been thought to occur once every 300 to 600 generations, or every 6,000 to 12,000 years for humans.

The Berkeley biochemists who developed the theory, Allan Wilson, Rebecca Cann, and Mark Stoneking, made several apparently reasonable assumptions. Since there were no DNA changes due to genetic recombination events (ie: with paternal DNA – now known to be a wrong assumption), they assumed that all changes in the mtDNA were the result of mutations over time and that these mutations occurred at a constant rate. On the basis of these assumptions, the researchers believed they had access to something like a “molecular clock.” Because mtDNA is thought to mutate faster than nuclear DNA (nucDNA), it was thought that the faster mutation rate of mtDNA would make for more accurate time keeping than nucDNA.

The original 1987 study involved mtDNA from 136 women from many parts of the world having various racial backgrounds. The analysis seemed to support the idea of a single ancestral mtDNA molecule from a woman living in sub-Saharan Africa about 200,000 years ago. Later, more detailed studies seemed to confirm this conclusion. Unfortunately though, there was a undetected bias in the computer program as well as with the researchers themselves. The researchers used a computer program designed to reveal a “maximum parsimony” phylogeny or the family tree with the least number of mutational changes.  This was based on the assumption that evolution would have taken the most direct and efficient path (which is not necessarily true, or even likely). Also, the computer program was biased by the order of data entry to favor the information entered first. This problem was recognized when the computer gave different results depending on the order that the data was entered. Now, after thousands of computer runs with the data entered randomly, it appears that the “African origin” for modern humans does not hold a statistical significance over other possibilities.26

The problems with these studies were so bad that Henry Gee, a member of the editorial staff for the journal, Nature, harshly described the studies as “garbage.” After considering the number of sequences involved (136 mtDNA sequences), Gee calculated that the total number of potentially correct parsimonious trees is somewhere in excess of one billion.25 Geneticist Alan Templeton (Washington University) suggests that low-level mixing among early human populations may have scrambled the DNA sequences sufficiently so that the question of the origin of modern humans and a date for “Eve” can never be settled by mtDNA.22 In a letter to Science, Mark Stoneking (one of the original researchers) acknowledged that the theory of an “African Eve” has been invalidated.23

Another interesting aspect of the “molecular clock” theory is the way in which the mutation rate itself was determined. Contrary to what many might think, the mutation rate was not initially determined by any sort of direct analysis, but by supposed phylogenic evolutionary relationships between humans and chimps. In other words, the mutation rate was calculated based on the assumption that the theory in question was already true. This is a rather circular assumption and as such all results that are based on this assumption will be consistent with this assumption – like a self-fulfilling prophecy. Since the rate was calculated based on previous assumptions of evolutionary time, then the results will automatically “confirm” the previous assumptions.  If one truly wishes independent confirmation of a theory,  then one cannot calibrate the confirmation test by the theory, or any part of the theory, that is being tested. And yet, this is exactly what was done by scientists such as Sarich, one of the pioneers of the molecular-clock idea. Sarich began by calculating the mutation rates of various species “…whose divergence could be reliably dated from fossils.” He then applied that calibration to the chimpanzee-human split, dating that split at from five to seven million years ago. Using Sarich’s mutation calibrations, Wilson and Cann applied them to their mtDNA studies, comparing “…the ratio of mitochondrial DNA divergence among humans to that between humans and chimpanzees.”24 By this method, they calculated that the common ancestor of all modern humans, the “African Eve”, lived about 200,000 years ago.

Obviously then, these dates, calculated from the mtDNA analysis, must match the presupposed evolutionary time scale since the calculation is based on this presupposition. The circularity of this method is inconsistent with good scientific method and is worthless as far as independent predictive value is concerned. The “mitochondrial clock” theory was and is basically a theory within a theory in that it has no independent predictive power outside of the theory of evolution.  It is surprising then that scientists did not catch this inherent flaw earlier.  Interestingly enough though, this flaw in reasoning was not detected for many years and perhaps would have remained undetected for much longer if a more direct mutation-rate analysis had not been done.

Eventually, scientists, who study historical families and their genetic histories, started questioning the mutation rates that were based on evolutionary phylogenetic assumptions.  These scientists were “stunned” to find that the mutation rate was in fact much higher than previously thought.  In fact it was about 20 times higher at around one mutation every 25 to 40 generations (about 500 to 800 years for humans).  It seems that in this section of the control region, which has about 610 base pairs, humans typically differ from one another by about 18 mutations. 3 By simple mathematics, it follows that modern humans share a common ancestor some 300 generations back in time.  If one assumes a typical generation time of about 20 years, this places the date of the common ancestor at around 6,000 years before present.  But how could this be?!   Thomas Parsons seems just as mystified. Consider his following comments published April of 1997, in the journal Nature Genetics:


“The rate and pattern of sequence substitutions in the mitochondrial DNA (mtDNA) control region (CR) is of central importance to studies of human evolution and to forensic identity testing. Here, we report a direct measurement of the intergenerational substitution rate in the human CR. We compared DNA sequences of two CR hypervariable segments from close maternal relatives, from 134 independent mtDNA lineages spanning 327 generational events. Ten substitutions were observed, resulting in an empirical rate of 1/33 generations, or 2.5/site/Myr. This is roughly twenty-fold higher than estimates derived from phylogenetic analyses. This disparity cannot be accounted for simply by substitutions at mutational hot spots, suggesting additional factors that produce the discrepancy between very near-term and long-term apparent rates of sequence divergence. The data also indicate that extremely rapid segregation of CR sequence variants between generations is common in humans, with a very small mtDNA bottleneck. These results have implications for forensic applications and studies of human evolution.

The observed substitution rate reported here is very high compared to rates inferred from evolutionary studies. A wide range of CR substitution rates have been derived from phylogenetic studies, spanning roughly 0.025-0.26/site/Myr, including confidence intervals. A study yielding one of the faster estimates gave the substitution rate of the CR hypervariable regions as 0.118 +- 0.031/site/Myr. Assuming a generation time of 20 years, this corresponds to ~1/600 generations and an age for the mtDNA MRCA of 133,000 y.a. Thus, our observation of the substitution rate, 2.5/site/Myr, is roughly 20-fold higher than would be predicted from phylogenetic analyses. Using our empirical rate to calibrate the mtDNA molecular clock would result in an age of the mtDNA MRCA of only ~6,500 y.a., clearly incompatible with the known age of modern humans. Even acknowledging that the MRCA of mtDNA may be younger than the MRCA of modern humans, it remains implausible to explain the known geographic distribution of mtDNA sequence variation by human migration that occurred only in the last ~6,500 years.” 27


The calculation is done in the following way: Let us consider two randomly chosen human beings.   Assuming all human beings initially have identical mitochondrial DNA,  after 33 generations, two such random human families will probably differ by two mutations, since there will be two separate lines of inheritance and probably one mutation along each line. After 66 generations, two randomly chosen humans will differ by about four mutations. After 100 generations, they will differ by about six mutations. After 300 generations, they will differ by about 18 mutations, which is about the observed value.

These experiments are quite concerning to evolutionists who previously calculated that the “mitochondrial eve” (who’s mitochondria is thought to be the ancestor mitochondria to all living humans) lived about 100,000 to 200,000 years ago in Africa.1 The new calculations, based on the above experiments, would make her a relatively young ~6,500 years old.  Now, the previous notion that modern humans are up to 10,000 generations old has to be reevaluated or at least the mtDNA basis for that assumption has to be reevaluated – and it has been.2

More recent direct mtDNA mutation rate studies also seem to confirm the earlier findings by Parsons and others.  In an 2001 article published in the American Journal of Human Genetics, Evelyne Heyer et. al., presented their findings of the mtDNA mutation rate in deep-rooted French-Canadian pedigrees.


Their findings “Confirm[ed] earlier findings of much greater mutation rates in families than those based on phylogenetic comparisons. . .  For the HVI sequences, we obtained 220 generations or 6,600 years, and for the HVII sequences 275 generations or 8,250 years.  Although each of these values is associated with a large variance, they both point to ~7,000-8,000 years and, therefore, to the early Neolithic as the time of expansion [mostly northern European in origin] . . . Our overall CR mutation-rate estimate of 11.6 per site per million generations . . . is higher, but not significantly different, than the value of 6.3 reported in recent the recent pedigree study of comparable size . . . In another study (Soodyall et al. 1997), no mutations were detected in 108 transmissions. On the other hand, two substitutions were observed in 81 transmissions by Howell et al. (1996), and nine substitutions were observed in 327 transmissions by Parsons et al. (1997).  Combining all these data (1,729 transmissions) results in the mutation rate of 15.5 (Cl 10.3-22.1).  Taking into account only those from deep-rooting pedigrees (1,321 transmissions) (Soodyall et al. 1997; Sigurdardottir et al. 2000; the present study) leads to the value of 7.9.  The latter, by avoiding experimental problems with heteroplasmy, may provide a more realistic approximation of the overall mutation rate.”  44


Also, consider an even more recent paper published in a 2003 issue of the Annals of Human Genetics by B. Bonne-Tamir et al. where the authors presented their results of a their study of “Maternal and Paternal Lineages” from a small isolated Samaritan community.  In this paper they concluded:


“Compared with the results obtained by others on mtDNA mutation rates, our upper limit estimate of the mutation rate of 1/61 mutations per generation is in close agreement with those previously published.” 45 [compared with the rate determined by Parsons of 1/33 generations, a rate of 1/61 is no more than double]


One more interesting paper published in September 2000 in the Journal Scientist by Denver et al. is also quite interesting. These scientists reported their work with the mtDNA mutation rates of nematode worms and found that these worm’s molecular clocks actually run about “100 times faster than previously thought”  [emphasis added].46


“Extrapolating the results directly to humans is not possible, say the scientists. But their results do support recent controversial studies suggesting that the human molecular clock also runs 100 times faster than is usually thought.  This may mean that estimates of divergence between chimpanzees and humans, and the emergence of modern man, happened much more recently than currently believed, says the team.  ‘Our work appears to support human analyses, which have suggested a very high rate,’ says Kelley Thomas of the University of Missouri.  ‘This work is relevant to humans,’ says Doug Turnbill of the institute for Human Genetics and Newcastle University, UK. ‘If the human mutation rate is faster than thought, it would have a lot of impact in looking at human disease and forensics, as well as the evolutionary rate of humans.’ . . .

Mutation rates of mtDNA in humans are usually estimated by comparing sequences of DNA from people and other animals. ‘This is kind of analysis that was used to determine that the African origin of modern humans was about 200,000 years ago,’ says Thomas. ‘The problem with this approach is that you are looking at both the mutation rate and the effects of natural selection,’ he says.  The technique would also miss multiple mutations in the same stretch of mtDNA, says Paul Sharp of the Institute of Genetics at Nottingham University, UK.

More recent studies have looked at the mtDNA of people who are distantly related but share a female ancestor.  This approach has revealed higher mtDNA mutation rates.  But the results have not been accepted by many scientists [emphasis added].

Knowing the exact rate of mutation in humans is very important for forensic science and studies of genetic disease, stresses Turnbill.  Forensic identification often rests on comparing samples of DNA with samples from suspected relatives. Faster human molecular clocks could complicate established exact relationships, he says.” 46


Obviously then, these rates, based on more direct observations, are nowhere near those based on indirect evolutionary assumptions. This certainly does “complicate” things just a bit now doesn’t it?  Isn’t it strange though that many scientists are still loath to accept these results?  The bias in favor of both evolution as well as millions of years for assumed divergence times between creatures like apes and humans is so strong that changing the minds of those who hold such positions may be pretty much impossible.

There are many other potential problems for phylogenies that rely on mtDNA sequence analysis and mutation rates.  One problem is that mtDNA functions as a single genetic locus, much like a single gene does in nucDNA.  Studies that work off a single genetic locus are more likely to be affected by random genetic changes than are studies that include more than one locus (the more the better).  Therefore, single locus studies are less accurate in characterizing a population.  Beyond this, the new evidence for paternal mtDNA mixing is quite problematic.16

Also, as briefly discussed above, the use of control regions as a “molecular clock” may not be as valid as was previously hoped.  Some nucleotide regions mutate slowly, while others can mutate relatively rapidly.17 These mutational “hotspots” can mutate fairly rapidly even within a single lifetime and are intuitively rather common in the aged.18 Of course such “somatic” mutations arise in mitochondria of various bodily tissues and, unless they involve gametes, they are not passed on to the next generation.  However, they would still affect phylogenetic interpretations.  Scientists have tried to compensate for these problems, but the various methods have produced divergent results.19 Also, as discussed above, direct comparisons of modern sequences with historical sequences often yield very difference results from those estimated by indirect methods that are based on present day sequence differences.  For another example from a different species, direct comparisons of modern penguins with historically sequenced penguins have shown that their mtDNA mutation rates are 2 to 7 times faster than had previously been assumed through indirect methods.20 Certain of these problems have in fact led some scientists to stop using control-region sequences to reconstruct human phylogenies.21

Those scientist that continue to try and revise the molecular clock hypothesis have tried to slow down the clock by showing that some mtDNA regions mutate much more slowly than do other regions.  The problem here is that such regions are obviously affected by natural selection.  In other words, they are not functionally neutral with regard to selection pressures.

For example, real time experiments have shown that average mitochondrial genome mutation rates are around 6 x 10-8 mut/site/mitochondrial generation – in line with various estimates of average bacterial mutation rates (Compare with nDNA rate of 4.4 x 10-8 mut/site/human generation).  With an average generation time of 45 days, that’s about 5 x 10-6 mut/site/year and 5 mut/site/myr.

This is about twice as high as Parsons’ rate of 2.5/mut/site/myr and about 40 to 50 times higher than rates based on phylogenetic comparisons and evolutionary assumptions. And, this is the average rate of the entire mitochondrial genome of 16,000pb.  One might reasonably think that all aspects of the hypervariable regions (HVI & HVII) would have a higher than average rate of mutation if truly neutral with regard to functional selection pressures.  Given this, those “slowly mutating sites” with rates as slow as 0.065 mut/site/myr (Heyer et al, 2001) would seem to be maintained in a biased way by natural selection.

Again, such non-neutral changes are not necessarily the reflection of elapsed time since a common ancestor so much as they are the reflection of the different functional needs of different creatures in various environments.



Nuclear DNA Mutation Rates

As with mitochondrial DNA mutation rates, the mutation rates of nuclear DNA have often been calculated based on evolutionary scenarios rather than on direct methods.   By such methods, the average mutation rate for eukaryotes in general is estimated to be about 2.2 x 10-9 mutations per base pair per year.29 With a 20 year average generation time for humans, this works out to be around 4.4 x 10-8 mutations per base pair per generation.  Since most estimates of the size of the diploid human genome run around 6.3 billion base pairs, this mutation rate would give the average child around 277 mutational differences from his or her parents.  This sounds like quite a high number and it is in fact on the high end of the spectrum when compared to studies looking more specifically at human mutation rates verses eukaryotic mutation rates in general.  A particular study by Nachman and Crowell estimated the average mutation rate specifically in humans by comparing control sequences in humans and chimpanzees.  Using these sequence comparisons, “The average mutation rate was estimated to be ~2.5 x 10-8 mutations per nucleotide site or 175 mutations per diploid genome per generation” [Based on a higher diploid genome estimate of 7 billion base pairs]. 30

These non-direct mutation rate estimates might actually seem reasonable given that they seem to match the error rates of DNA replication that occur between the formation of a zygote in one generation and the formation of a zygote in the next generation.  In the illustration31 below, notice that from fertilization to the formation of a woman’s first functional gamete, it takes about 23 mitotic divisions.  Men, on the other hand, contribute about twice as many germ line mutations as women do.33 At least part of the reason is that their stem cells keep dividing so that the older a man gets before having children more mitotic divisions occur.

Now, consider that each diploid fertilized zygote contains around 6 billion base pairs of DNA (~3 billion from each gamete/parent, using a conservative round number).32 From cell division to cell division, the error rate for DNA polymerase combined with other repair enzymes is about 1 mistake in 1 billion base pairs copied.42 At this rate, there are about 6 mistakes with each diploid cell replication event.  With a male/female average of 29 mitotic divisions before the production of the next generation, this works out to be about 175 mutations per generation.

Of course, this is right in line with the mutation rates that are based on evolutionary scenarios.   However, some estimates place the overall mutation rate as low as 1 mistake in 10 billion base pairs copied.43 At this rate, one would expect around 0.6 mistakes with each replication event and only around 17 mutations per person per generation.  So, perhaps something else is going on that also influences the nuclear DNA mutation rate?  As it turns out, replication errors are not the only sources of DNA mutations.  Damage to DNA can and does often occur spontaneously.  Genome stability is continually challenged by a diverse array of mutagenic forces that include errors during DNA replication, environmental factors such as UV radiation, and endogenous mutagens such as oxygen free radicals generated during oxidative metabolism.  This damage must also be detected and repaired on a constant basis.  Of course, this repair isn’t perfect and therefore likely contributes significantly to the actual mutation rate far over that estimated by the indirect methods discussed above.

In fact, the actual observed mutation rate is likely to be quite a bit higher than 1 x 10-8 per generation – – at least 10 fold higher and by some estimates (see Link).  Such high mutation rates are based on actual observed rates of functional mutations in the human genome and directly observed mutation rates in pseudogenes – such as those found in C. elegans (see further discussion below).

Again, consider that the rate of 1 x 10-9 per year that is referenced above is an indirect estimate based on evolutionary assumptions of the time since the MRCA between two species.  This particular commonly-referenced rate is based on the supposed time since the MRCA between the two species and the comparison of sequences which are though to be functionally neutral.


“Comparisons of pseudogenes and of synonymous sites between humans and chimpanzees have suggested mutation rates on the order of 10-8 per site per generation [ or about 10-9 per site per year] (e.g., KONDRASHOV and CROW 1993 Down; DRAKE et al. 1998 Down).” (see Link)


So, you see, mutation rate estimates based on this sort of evolutionary assumption produce a self-fulfilling prophecy when it comes to estimating the time of the MRCA between humans and apes based on mutation rate analysis.  On the other hand, more direct methods of detecting the nuclear mutation rates in animals suggest that the actual rate is likely to be about ten times higher than estimates based on indirect methods and evolutionary assumptions. Consider the following excerpt from Denver et. al. published in Nature in 2004:


“Alternative approaches in mammals, relying on phylogenetic comparisons of pseudogene loci and fourfold degenerate codon positions, suffer from uncertainties in the actual number of generations separating the compared species and the inability to exclude biases associated with natural selection. Here we provide a direct and unbiased estimate of the nuclear mutation rate and its molecular spectrum with a set of C. elegans mutation-accumulation lines that reveal a mutation rate about tenfold higher than previous indirect estimates and an excess of insertions over deletions.” (see Link)


Such a high mutation rate, a rate of at least 2000 per person per generation, might be a more of a problem than it is for humans if it were not for the fact that much of the human genome is thought by most scientists to have no significant functional role and can therefore sustain mutations without significant detrimental effects on the overall function of the organism.  The amount of this non-functional DNA has been estimated by calculating the coding portion of DNA and subtracting this from the total genomic real estate.  It seems as though the average coding portion of a human gene is around 1,350 base pairs in size.  Of course, this gene would code for a protein averaging 450 amino acids.38 Now, multiplying this number by the total number of genes should give a reasonable estimate of the coding genetic real estate.

However, there is some argument as to the total number of genes in the human genome.  For many years it was thought that humans had between 70,000 to 140,000 genes.  However, scientists working on the human genome project made a surprising discovery.  When they finished the project in February of 2001, they estimated that the actual gene count was somewhere between 30,000 to 40,000 genes.39 But a year later, in February of 2002, at the annual meeting of the American Association for the Advancement of Science (publisher of Science), one of the presenters, Victor Velculescu, suggested that the real number of genes in the human genome may actually be closer to 70,000 genes after all.   He and his colleagues, at Johns Hopkins University in Baltimore, Maryland, have gone back to the lab to look for genes that the computer programs may have missed. Their technique, called serial analysis of gene expression (SAGE), works by tracking RNA molecules back to their DNA sources. After isolating RNA from various human tissues, the researchers copy it into DNA, from which they cut out a kind of genetic bar code of 10 to 20 base pairs. Velculescu proposes that the vast majority of these tags are unique to a single gene. If so, the tags can then be compared to the human genome to find out if they match up with genes discovered by the computer algorithms. Velculescu stated that only about half of the tags he used match the genes identified earlier in the genome project.  Therefore, he suggests that the human inventory of genes had been underestimated by about half.

The reason for the disparity may be that the standard computer programs were largely developed for the genomes of simple (prokaryotic) organisms, not for the more complex sequences found in the genomes of humans and other eukaryotes. “We’re still not very good at predicting genes in eukaryotes,” said Claire Fraser of The Institute for Genomic Research in Rockville, Maryland. “It is entirely possible that there could be more than 32,000 genes, and SAGE is an important approach to finding them. You absolutely have to go back into the lab and get away from the computer terminal.” 40

So, it seems as though there is still some question as to exactly how many genes the human genome contains. This is especially true now that non-protein-coding portions of DNA, like many so-called pseudogenes and miro-RNAs, are now being found to have significant functionality.   But, for the sake of argument, lets go with a lower estimate of ~40,000 genes.  With each gene averaging 1,350 base pairs in size, only around 108 million base pairs out of 6 billion base pairs (diploid) would code for anything.  This is only around 1.8% of the total genome.  Much of the rest of the human genome (At least 50%) is thought to be composed of a large amount of “repetitive DNA” that is made up of similar sequences occurring over and over.33,38 At least some of the other 48% of the genome is thought to provide structural integrity as well as regulating the production of the coding sequences of DNA as far as when, where, and how much protein to make.  However, exactly how much of the non-protein-coding genome is functional is not clearly understood but may be quite high (i.e., well over 50% with at least some functionality; see Link).  In any case, for the purposes of this discussion a rough figure of 2% will be used as the amount of functional DNA in the human genome.



The Detrimental Mutation Rate

and the Genetic Deterioration of Mankind

Since mutations are the only possible source of novel genomic function in the evolution of living things, we should consider a few facts about these mutations.  Mutations are thought to be purely random events causes by errors of replication and maintenance over time. They occur anywhere in the entire genome in a fairly random fashion with each generation.  Given this information, lets consider how these mutations would build up and what effect, if any, they would have on a human lineage.

Some researchers suggests a detrimental mutation rate (Ud) of 1 to 3 per person per generation with at least some scientists (Nachmann and Crowell, 2000) favoring at least 3 or more.30 Notice that these detrimental mutation rates are based on overall DNA mutation rate estimates that are indirectly determined based on assumed evolutionary relationships. The actual mutation rates, as noted above, are likely to be much higher.  In any case, even given these assumptions, since detrimental mutations outnumber beneficial mutations by at least 1,000 to 1, it seems like the build up of detrimental mutations in a population might lead toward extinction. 34,36

Nachmann and Crowell detail the perplexing situation at hand in the following conclusion from their fairly recent paper on human mutation rates:


The high deleterious mutation rate in humans presents a paradox.  If mutations interact multiplicatively, the genetic load associated with such a high U [detrimental mutation rate] would be intolerable in species with a low rate of reproduction [like humans and apes etc.] . . .

The reduction in fitness (i.e., the genetic load) due to deleterious mutations with multiplicative effects is given by 1 – e -U (Kimura and Moruyama 1966).  For U = 3, the average fitness is reduced to 0.05, or put differently, each female would need to produce 40 offspring for 2 to survive and maintain the population at constant size.  This assumes that all mortality is due to selection and so the actual number of offspring required to maintain a constant population size is probably higher.

The problem can be mitigated somewhat by soft selection or by selection early in development (e.g., in utero).  However, many mutations are unconditionally deleterious and it is improbable that the reproductive potential on average for human females can approach 40 zygotes.  This problem can be overcome if most deleterious mutations exhibit synergistic epistasis; this is, if each additional mutation leads to a larger decrease in relative fitness.  In the extreme, this gives rise to truncation selection in which all individuals carrying more than a threshold number of mutations are eliminated from the population.  While extreme truncation selection seems unrealistic [the death of all those with a detrimental mutational balance], the results presented here indicate that some form of positive epistasis among deleterious mutations is likely.30



Nachmann and Crowell find the situation a very puzzling one.  How does one get rid of all the bad mutations faster than they are produced?  Does their hypothesis of “positive epistasis” adequately explain how detrimental mutations can be cleared faster than they are added to a population?  If the functional effects of mutations were increased in a multiplicative instead of additive fashion, would fewer individuals die than before?  As noted above, even if every detrimental mutation caused the death of its owner, the reproductive burden of the survivors would not diminish, but would remain the same.

For example, lets say that all those with at least three detrimental mutations die before reproducing.  The population average would soon hover just above 3 deleterious mutation rates.  Over 95% of each subsequent generation would have 3 or more deleterious mutations as compared with the original “neutral” population.  The death rate would increase dramatically.  In order to keep up, the reproductive rates of those surviving individuals would have to increase in proportion to the increased death rate.  The same thing would eventually happen if the death line were drawn at 100, 500, 1000, 10000 or more deleterious mutations.  The only difference would be the length of time it would take a given population to build up a lethal number of deleterious mutations in its gene pool beginning at a relatively “neutral” starting point.  The population might survive fairly well for many generations without having to resort to huge increases in the reproduction rate.  However, without getting rid of the accumulating deleterious mutations, the population would eventually find itself experiencing an exponential rise in its death rate as its average population crossed the line of lethal mutations.

Since the theory of positive epistasis does not seem to help the situation much, some other process must be found to explain how to preferentially get rid of detrimental mutations from a population.  Consider an excerpt from a fairly recent Scientific American article entitled, “Mutations Galore”:


According to standard population genetics theory, the figure of three harmful mutations per person per generation implies that three people would have to die prematurely in each generation (or fail to reproduce) for each person who reproduced in order to eliminate the now absent deleterious mutations [75% death rate].  Humans do not reproduce fast enough to support such a huge death toll.  As James F. Crow of the University of Wisconsin asked rhetorically, in a commentary in Nature on Eyre-Walker and Keightley’s analysis: “Why aren’t we extinct?”

Crow’s answer is that sex, which shuffles genes around, allows detrimental mutations to be eliminated in bunches.  The new findings thus support the idea that sex evolved because individuals who (thanks to sex) inherited several bad mutations rid the gene pool of all of them at once, by failing to survive or reproduce.

Yet natural selection has weakened in human populations with the advent of modern medicine, Crow notes.  So he theorizes that harmful mutations may now be starting to accumulate at an even higher rate, with possibly worrisome consequences for health.  Keightley is skeptical:  he thinks that many mildly deleterious mutations have already become widespread in human populations through random events in evolution and that various adaptations, notably intelligence, have more than compensated.  “I doubt that we’ll have to pay a penalty as Crow seems to think,” he remarks. “We’ve managed perfectly well up until now.” 37


Well, the answer might be found in a combination of processes where both sexual replication and natural selection play a role to keep a slowly reproducing population from going extinct.  For example consider the following chart showing how deleterious mutations build up in a population that reproduces via asexual means: 49

Notice how the most fit “Progenitor Class” (P) loss numbers in each generation while the numbers of those that have greater numbers of deleterious mutations build up more and more.  In this article Rice notes that in asexual populations the only way to really overcome this buildup of detrimental mutations is to increase the reproductive rate substantially.  But, what about beneficial mutations?  Rice comments, “Rare reverse and compensatory mutations can move deleterious mutations, via genetic hitchhiking, against the flow of genetic polarization. But this is a minor influence, analogous to water turbulence that occasionally transports a pebble a short distance upstream.” 49 So, how do sexually reproducing populations overcome this problem?

When it comes to sexually reproducing populations, the ability for genetic recombination during the formation of gametes makes it possible to concentrate both good and bad mutations.  For example, lets say we have two individuals, each with 2 detrimental mutations. Given sexual recombination between these two individuals, there is a decent chance that some of their offspring (1 chance in 32) will not have any inherited detrimental mutations.   But what happens when the rate of additional detrimental mutations is quite high – higher than 3?

To look into this just a bit more, consider another example of a steady state population of 5,000 individuals each starting out with 7 detrimental mutations and an average detrimental mutation rate of 3 per individual per generation.  Given a reproductive rate of 4 offspring per each one of the 2,500 couples (10,000 offspring), in one generation, how many offspring will have the same or fewer detrimental mutations than the parent generation?


Inherited After Ud = 3
7 901
6 631
5 378
4 189
3 76
2 23
1 5
0 0.45
< or = 7 2202


This Poisson approximation shows that out of 10,000 offspring, only 2,202 of them would have the same or less than the original number of detrimental mutations of the parent population.  This leaves 7,798 with more detrimental mutations than the parent population.51 Of course, in order to maintain a steady state population of 5,000, natural selection must cull out 5,000 of these 10,000 offspring before they are able to reproduce.  Given a preference, those with more detrimental mutations will be less fit by a certain degree and will be removed from the population before those that are more fit (less detrimental mutations).  Given strong selection pressure, the second generation might be made up of ~2,200 more fit individuals and only ~2,800 less fit individuals with the overall average showing a decline as compared with the original parent generation.  If selection pressure is strong, so that the majority of those with more than 7 detrimental mutations are removed from the population, the next generation will only have about 1,100 mating couples as compared to 2,500 in the original generation.  With a reproductive rate of 4 per couple, only 4,400 offspring will be produced as compared to 10,000 originally.  In order to keep up with this loss, the reproductive rate must be increased or the population will head toward extinction.  In fact, given a detrimental mutation rate of Ud = 3 in a sexually reproducing population, the average number of offspring needed to keep up would be around 20 per breeding couple (2eUd/2).  While this is about half that required for an asexual population (2eUd), it is still quite significant.

In this light, consider that more recent estimates suggest that the deleterious mutation rate is even higher. “Extrapolations from studies of humans and Drosophila (Mukai, 1979; Kondroshov, 1988; Crow, 1993) suggest that Ud > 5 is feasible.” 49 However, the number of required offspring needed to compensate for a detrimental mutation rate of Ud = 5 would soar to 148 per female per generation!  And, this is not the worst of it.  Recent genetic studies have shown that much of what was once thought of as “junk DNA” is actually functional ( Link ).  In fact, these recent studies suggest that the total amount of functional DNA in the human genome is not actually 2-3% as previously thought, but is upwards of 85-90% ( Link ).  Consider also that what were once thought to be neutral mutations are now being discovered to be functional mutations governed by natural selection.  In a 2007 paper published in the Indian Journal of Human Genetics, author Clyde Winters claims to have made a very interesting discovery.


It is often assumed that selection plays a limited role in the mtDNA control region. . .  However, there is a selective constraint on mutation frequencies of an mtDNA site. Some of the East African transitions . . . are the most rapidly occurring nucleotide substitutions in the human mitochondrial genome. These transitions are often referred too as “hotspots.” These hot spots of mutational activity suggest that positive selection influences mutation rates and not neutral selection which, theoretically, would manifest parallel mutations.53



Of course, this is not the only region in the human genome that was once thought to be limited to neutral mutations alone. Much of the genome is now known to be subject to differential selection.

So what.  What does this matter?  It matters to this particular problem because the actual detrimental mutation rate would be a significantly greater percentage of the total number of mutations experienced by the genome in each generation.  As noted above, the total number of mutations per offspring per generation is at least 175.  If the functional genome percentage was actually 50% (instead of just 2%), the likely detrimental mutation rate (Ud) would be well over 30 instead of the usual estimates of ~3 noted above.  This would increase the reproductive rate needed to avoid genomic decay from ~20 offspring per woman per generation to well over 10 trillion offspring per woman per generation – obviously an impossible hurdle to overcome.

In short, the best available evidence overwhelmingly supports the theory that the human genome is in decay.  The various forms of “positive epistasis” (see illustration by Rice below) 49 do not solve this problem.



Y-Chromosome Headed for Extinction?

Also, what about the Y-chromosome in males?  The Y-chromosome does not undergo significant sexual recombination. Are the males of slowly reproducing species, like humans, therefore headed for extinction at an even faster rate than females?


“The absence of recombination with a homologous partner means that it [The Y-chromosome] can never be repaired by recombination. This has led to suggestions that the Y is destined for extinction. “It will eventually dwindle to nothing. According to this model, its role in sex determination will eventually be taken on by genes elsewhere in the genome.”  50


The author of the above quoted article goes onto point out that several species, like the Armenian mole vole, are able to reproduce without the Y chromosome.  While this might explain where humans are headed, it doesn’t seem quite clear as to just how the Y-chromosome could have evolved over millions of years of time given its relative inability to combat high detrimental mutation rates.



  1. Gibbons, A. Calibrating the Mitochondrial Clock, Science 279, Volume 279, Number 5347 Issue of 2 Jan 1998, pp. 28 – 29
  2. Collins, F., M. Guyer, and A. Chakravarti, Variations on a Theme: Human DNA Sequence Variation, Science 278:1580-1581, 28 November 1997, page 1581.
  3. Genetics vol. 15, April 1997, pp. 363-367.
  4. S. Horai, K. Hayasaka, R. Kondo, K. Tsugane, and N. Takahata, Recent African origin of modern humans revealed by complete sequences of hominoid mitochondrial DNAs, Proc. Natl. Acad. Sci. USA 1995 Jan 17;92(2):532-536.
  5. Dorit, R.L., Akashi, H. and Gilbert, W. 1995. Absence of polymorphism at the ZFY locus on the human Y chromosome, Science 268 (26 May 1995):1183-1185.
  6. L. Simon Whitfield, John E. Sulston, and Peter N. Goodfellow, Sequence Variation of the Human Y Chromosome, Nature 378 (1995), pp. 379-380.
  7. V. Morell. The Origin of Dogs: Running With the Wolves, Science 1997 June 13; 276 (5319):1647 (in Research News).
  8. C. Vila, P. Savolainen, J. E. Maldonado, I. R. Amorim, J. E. Rice, R. L. Honeycutt, K. A. Crandall, J. Lundeberg, and R. K. Wayne, Multiple and Ancient Origins of the Domestic Dog, Science, June 13, 1997, vol. 276, no. 5319, pp. 1687-1689 (in Reports).
  9. Multiple independent transpositions of mitochondrial DNA control region sequences to the nucleus, PNAS 1996 93: pp. 15239-15243.
  10. J. Klicka and R. M. Zink. The Importance of Recent Ice Ages in Speciation: A Failed Paradigm, Science 1997 September 12; 277 (5332): p. 1666 (in Reports).
  11. Spetner, Not by Chance, Judaica Press, Brooklyn, New York, 1997, page 92.
  12. Moreel, V., Bacteria Diversify Through Warfare, Science, Volume 278, October 24, 1997, page 575.
  13. Kondrashev, A.S., 1988, Deleterious mutations and the origin of sexual reproduction, Nature vol. 336 Dec. 1 pp. 435-440.
  14. Ninth International Conference on Microbial Genomes, October 28th-November 1st, 2001. Gatlinburg, TN ( http://cgb.utmem.edu/meeting_reports/redwards_11_06_01.htm )
  15. http://genetics.hannam.ac.kr/lecture/Mgen02/Mutation%20Rates.htm
  16. Williams, Sloan R., Napoleon A. Chagnon, and Richard S. Spielman (2002) “Nuclear and mitochondrial genetic variation in the Yanomamö: A test case for ancient DNA studies of prehistoric populations.” American Journal of Physical Anthropology 117: 246-259.
  17. Stoneking, Mark (2000) “Hypervariable sites in the mtDNA control region are mutational hotspots.” American Journal of Human Genetics 67: 1029-1032.
  18. Nekhaeva, E., N.D. Bodyak, Y. Kraytsberg, S.B. McGrath, N.J. Van Orsouw, A. Pluzhnikov, J.Y. Wei, J. Vijg, and K. Khrapko (2002) “Clonally expanded mtDNA point mutations are abundant in individual cells of human tissues.” Proceedings of the National Academy of Sciences 99: 5521-5526.
  19. Heyer, Evelyne, Ewa Zietkiewicz, Andrzej Rochowski, Vania Yotova, Jack Puymirat, and Damian Labuda (2001) “Phylogenetic and familial estimates of mitochondrial substitution rates: Study of control region mutations in deep-rooting pedigrees.” American Journal of Human Genetics 69: 1113-1126.
  20. Lambert, D.M., P.A. Ritchie, C.D. Millar, B. Holland, A.J. Drummond, and C. Baroni (2002) “Rates of evolution in ancient DNA from Adélie penguins.” Science 295: 2270-2273.
  21. Ingman, Max, Henrik Kaessmann, Svante Pääbo, and Ulf Gyllensten (2000), Mitochondrial genome variation and the origin of modern humans, Nature 408: 708-713.
  22. Barinaga, Marcia , African Eve’ Backers Beat a Retreat, Science, 255 (7 February 1992): 687.
  23. Hedges, S. Blair , Sudhir Kumar, Koichiro Tamura, and Mark Stoneking, Human Origins and Analysis of Mitochondrial DNA Sequences, Science, 255 (7 February 1992): 737-739.
  24. Wilson, Allan C.,   Cann, Rebecca L., The Recent African Genesis of Humans, Scientific American, April 1992.
  25. Gee, Henry, Statistical Cloud over African Eden, Nature, 355 (13 February 1992): 583.
  26. Lubenow, Marvin, The Apple Computer Bites the African Eve, Impact No. 229, Institute for Creation Research, July 1992 (http://www.icr.org/pubs/imp/imp-229.htm )
  27. Parsons, Thomas J. A high observed substitution rate in the human mitochondrial DNA control region, Nature Genetics vol. 15, April 1997, pp. 363-367
  28. Coghlan, Andy, Proceedings of the National Academy of Sciences (DOI: 10.1073/pnas.172510699) (http://www.newscientist.com/news/news.jsp?id=ns99992833)
  29. Sudhir Kumar, Sankar Subramanian, Mutation Rates in Mammalian Genomes, PNAS, January 22, 2002, Vol. 99, No. 2, p. 803-808. ( www.pnas.org/cgi/doi/10.1073/pnas.022629899 )
  30. Nuchman, Michael W., Crowell, Susan L., Estimate of the Mutation Rate per Nucleotide in Humans, Genetics, September 2000, 156: 297-304 ( http://www.genetics.org/cgi/content/full/156/1/297? )
  31. http://www.nature.com/cgi-taf/DynaPage.taf?file=/nrg/journal/v1/n1/full/nrg1000_040a_fs.html
  32. http://www.bgsu.edu/departments/chem/midden/chem308/slides/DNARBW.pdf
  33. http://www.ornl.gov/hgmis/project/info.html
  34. http://www.cs.unc.edu/~plaisted/ce/genetics.html
  35. Ben Shouse, American Association for the Advancement of Science Annual Meeting, Human Gene Count on the Rise, Science Feb 22 2002: 1457. ( http://www.cs.unc.edu/~plaisted/ce/genome3.html )
  36. http://socrates.barry.edu/snhs-jmontague/courses/BIO%20440%20-%20Evolution/440%20ppt%20lectures/440%20web%20lec%2005%202002.ppt
  37. Beardsley, Tim , Mutations Galore, Scientific American, Apr99, Vol. 280 Issue 4, p32, 2p
  38. http://users.rcn.com/jkimball.ma.ultranet/BiologyPages/H/HGP.html
  39. Lemonick, M. Gene Mapper, Time, Vol. 156, No. 26, pp110, 2001.
  40. Shouse, Ben. American Association for the Advancement of Science Annual Meeting:  Human Gene Count on the Rise, Science, 22 Feb. 2002: 1457. (http://www.cs.unc.edu/~plaisted/ce/genome3.html)
  41. www.naturalselection.0catch.com/Files/GalatosidaseEvolution.html
  42. http://info.bio.cmu.edu/courses/03231/LecF02/Lec18/lec18.html
  43. http://www.blc.arizona.edu/marty/411/Modules/mod6.html
  44. Evelyn Heyer, Ewa Zietkiewicz, Andrezej Rochowski, Vania Yotova, Jack Puymirat, and Damian Labuda, Phylogenetic and Familial Estimates of Mitochondrial Substitution Rates: Study of Control Region Mutations in Deep-Rooting Pedigrees. Am. J. Hum. Genet., 69:1113-1126. 2001 ( http://www.journals.uchicago.edu/AJHG/journal/issues/v69n5/013122/013122.text.html )
  45. B. Bonn-Tamir, M. Korostishevsky, A. J. Redd, Y. Pel-Or, M. E. Kaplan and M. F. Hammer, Maternal and Paternal Lineages of the Samaritan Isolate: Mutation Rates and Time to Most Recent Common Male Ancestor, Annals of Human Genetics, Volume 67 Issue 2 Page 153  – March 2003 ( http://www.blackwell-synergy.com/links/doi/10.1046/j.1469-1809.2003.00024.x/full/ )
  46. Denver DR, Morris K, Lynch M, Vassilieva LL, Thomas WK. High direct estimate of the mutation rate in the mitochondrial genome of Caenorhabditis elegans. Science. 2000 Sep 29;289(5488):2342-4. ( http://www.sciencemag.org/cgi/content/full/289/5488/2342?ck=nck )
    Also reported by: Emma Young, Running Slow, New Scientist, September 28, 2000. ( http://www.newscientist.com/news/news.jsp?id=ns226930 )
  47. Schwartz, Marianne and John Vissing (2002), Paternal Inheritance of Mitochondrial DNA, New England Journal of Medicine, 347:576-580, August 22. ( http://0-content.nejm.org.catalog.llu.edu/cgi/content/full/347/8/576 )
  48. Morris, Andrew A. M., and Robert N. Lightowlers (2000), Can Paternal mtDNA be Inherited?, The Lancet, 355:1290-1291, April 15.  ( Full Text )
  49. William R. Rice, Requisite mutational load, pathway epistasis, and deterministic mutation accumulation in sexual versus asexual populations, Genetica 102/103: 71-81, 1998. 71 (Full Text)
  50. IJ, Sex and Death, The Human Genome – In The Genome, September, 2003 (Full Text)
  51. Poisson Distribution Calculator (Link)
  52. Binomial Calculator (Link)
  53. Clyde Winters (2007), “Can parallel mutation and neutral genome selection explain Eastern African M1 consensus HVS-I motifs in Indian M haplogroups,” Indian Journal of Human Genetics, 13(3):93-96. (Link)



Leave a Reply