Human genetics

Mystery of the mutagenic male

Old fathers are the source of more genetic mutations in their offspring than either young fathers or mothers of any age. But the apparently most plausible explanation for this effect might not hold.

Mutations are the raw material for evolution, and the cause of genetic disease. But how do they arise? A clue is the finding that fathers are the source of more mutations than mothers. The usual explanation is that the copying of DNA (replication) is error prone1,2,3, and that women's reproductive cells suffer fewer such errors than men's. The reason is that eggs are the product of relatively few rounds of replication, all of which are finished before birth, but sperm are the result of many more.

Pre-sperm cells go through a few replications before starting the process of becoming sperm. But to replenish the supply, some of the products of these replications go back to the pre-sperm cell bank. As the manufacture of sperm continues throughout a man's life, the number of replication events experienced by the DNA of cells in this cell bank goes up as men age. Consequently, not only should fathers be the source of most mutations but older dads would be expected to contribute more mutations than younger ones1,4. Writing in Proceedings of the National Academy of Sciences, however, Irene Tiemann-Boege and colleagues5 question whether this 'male age effect' can be accounted for by replication errors alone.

Tiemann-Boege et al. investigated a mutation that results in achondroplasia, the most common form of dwarfism. It had been previously established that fathers are the source of all the mutations responsible for this condition6, and that the probability of having an affected offspring increases exponentially as a function of the father's age4. If we assume an increasing error rate per replication as males age, this pattern is consistent with the replication hypothesis1.

Using the polymerase chain reaction to identify the mutation in sperm from males of different ages, Tiemann-Boege et al. asked whether they could detect an exponential increase in the frequency of sperm containing the mutation. They could not. Up to age 40, the men in their study all had about the same proportion of sperm bearing the mutation. Over age 55 there was also no increase (although the variance was large). Between ages 40 and 55 there was an increase, but overall there was no exponential male age effect. Further, the absolute increase in the mutation rate in young compared with old males was nowhere near that needed to explain the observed rates of achondroplasia: there was a tenfold difference in the rate of achondroplasia in the children of males aged 20 compared with those aged 50, but only a twofold difference in the frequency of the mutation in sperm (Fig. 1).

Figure 1: Sperm — not guilty?

The increase with paternal age of spontaneous cases of achondroplasia (dwarfism) in offspring is indicated in green. The results of Tiemann-Boege et al.5, showing the increase — or the comparative lack of increase — with age in the frequency of the achondroplasia mutation in sperm, are in red. On the face of it, the results run counter to the idea that replication errors are one of the main causes of mutation. (Adapted from ref. 5.)

What might explain the discrepancy? The answer is not known, but Tiemann-Boege et al. propose several alternatives. One is that there is some kind of selective effect. If sperm carrying the mutation are more likely to fertilize the egg, or if fertilized eggs carrying the mutation have a higher chance of success, then the low mutation counts in the sperm could translate to higher rates of affected offspring at birth. Why this might be age-dependent is unclear. Regardless of that, the idea that the exponential age effect owes more to replications, and a higher error rate per replication in older males, is not straightforwardly supported by the findings.

We should, however, hesitate in extrapolating from this one result, not least because the mutation in question may not be typical. For one thing, the DNA at the site of the mutation is modified by the addition of a methyl group, which affects mutational processes. Further, not only is the incidence of the disease-causing mutation remarkably high, but studies of other disease-associated genes indicate that the strength of the male age effect is highly variable4, as is the ratio of male-derived to female-derived mutations3. For example, although all mutations are male-derived in achondroplasia, only 13 of 23 mutations in the gene NF2 (which result in type 2 neurofibromatosis) are from fathers7. This variation is puzzling from the standpoint of the replication model. Perhaps, for all these genetic loci, there is variation in the ratio of mutated sperm to affected offspring but no difference in the underlying male bias to the mutation rate. It would be valuable to know whether, in other diseases in which all mutations are male-derived (Apert's syndrome, for example, which manifests itself as skull, hand and feet malformations), there is the same discrepancy between mutated sperm and affected offspring.

Given the possibility that selection can affect estimates of the rate of occurrence of disease-associated mutations, a better approach to quantifying mutation rates may be to examine 'neutral' mutations — those that have no detrimental effects. This can be done by comparing the same DNA sequence in different species and estimating the rate of neutral evolution. If, in a stretch of DNA, natural selection does not affect the fate of the mutations, then the rate of evolution of the sequence is equal to the mutation rate. From the replication-errors model, it would be expected that the X chromosome should evolve more slowly than the Y chromosome: the Y is always in males whereas, in evolutionary terms, the X spends two-thirds of its time in females. The other chromosomes (the autosomes) should have an intermediate rate.

Differences in the rate of neutral evolution of the three classes of chromosome can then be used to estimate the extent of the sex bias in the mutation rate8. If nearly all mutations are replication-dependent, these estimates should accord with anatomically derived estimates of the differences in the number of replication events that precede egg and sperm production. Some calculations have conformed with these expectations, with estimates9 that in humans the ratio of the mutation rate in males to that in females (α) is around 5. Comparisons in rodents (α = 2), cats (α = 4 ), birds (α = 4) and fruitflies (α = 1) also seem to tie in roughly with expectations from anatomy2. In rodents, however, the ratio depends on which chromosomes are compared: a comparison of X with Y results in the figure consistent with expectations, but comparing X with autosomes suggests a much higher figure10.

Although this latter discrepancy is not apparent from human data, analyses of primate and rodent sequences have revealed a further curiosity: not only does mutation rate vary along chromosomes11,12,13, but also different autosomes have remarkably different rates of evolution12,14. As all autosomes undergo the same number of replications, perhaps this will be the next clue to understanding the causes of mutation.


  1. 1

    Crow, J. F. Nature Rev. Genet. 1, 40–47 (2000).

    CAS  Article  Google Scholar 

  2. 2

    Hurst, L. D. & Ellegren, H. Trends Genet. 14, 446–452 (1998).

    CAS  Article  Google Scholar 

  3. 3

    Chang, B. H. J., Shimmin, L. C., Shyue, S. K., Hewett-Emmett, D. & Li, W.-H. Proc. Natl Acad. Sci. USA 91, 827–831 (1994).

    ADS  CAS  Article  Google Scholar 

  4. 4

    Risch, N., Reich, E. W., Wishnick, M. M. & McCarthy, J. G. Am. J. Hum. Genet. 41, 218–248 (1987).

    CAS  PubMed  PubMed Central  Google Scholar 

  5. 5

    Tiemann-Boege, I. et al. Proc. Natl Acad. Sci. USA 99, 14952–14957 (2002).

    ADS  CAS  Article  Google Scholar 

  6. 6

    Wilkin, D. J. et al. Am. J. Hum. Genet. 63, 711–716 (1998).

    ADS  CAS  Article  Google Scholar 

  7. 7

    Kluwe, L. et al. Neurogenetics 3, 17–24 (2000).

    CAS  Article  Google Scholar 

  8. 8

    Miyata, T., Hayashida, H., Kuma, K., Mitsuyasu, K. & Yasunaga, T. Cold Spring Harb. Symp. Quant. Biol. 52, 863–867 (1987).

    CAS  Article  Google Scholar 

  9. 9

    Makova, K. D. & Li, W. H. Nature 416, 624–626 (2002).

    ADS  CAS  Article  Google Scholar 

  10. 10

    Smith, N. G. C. & Hurst, L. D. Genetics 152, 661–673 (1999).

    CAS  PubMed  PubMed Central  Google Scholar 

  11. 11

    Matassi, G., Sharp, P. M. & Gautier, C. Curr. Biol. 9, 786–791 (1999).

    CAS  Article  Google Scholar 

  12. 12

    Lercher, M. J., Williams, E. J. B. & Hurst, L. D. Mol. Biol. Evol. 18, 2032–2039 (2001).

    CAS  Article  Google Scholar 

  13. 13

    Smith, N. G. C., Webster, M. T. & Ellegren, H. Genome Res. 12, 1350–1356 (2002).

    CAS  Article  Google Scholar 

  14. 14

    Ebersberger, I., Metzler, D., Schwarz, C. & Paabo, S. Am. J. Hum. Genet. 70, 1490–1497 (2002).

    CAS  Article  Google Scholar 

Download references

Author information



Rights and permissions

Reprints and Permissions

About this article

Cite this article

Hurst, L., Ellegren, H. Mystery of the mutagenic male. Nature 420, 365–366 (2002).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Quick links

Sign up for the Nature Briefing newsletter for a daily update on COVID-19 science.
Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing