PROBLEMS OF MEASUREMENT OF MUTATION RATES 



The human markers that might have been used for mutation studies 

 are much more numerous than those that have been used, so some kind 

 of selection has gone on. However, extreme rarity is only one of many 

 reasons not to estimate the mutation rate. I think that the rates re- 

 ported may be fairly representative. 



Auerbach: May I ask a question? Do you feel sure that, for those 

 which have been studied, only one locus is involved? That is a big 

 assumption. 



Atwood: No, that is never certain. I don't know whether there is any 

 point in more discussion of how badly mistaken we can be with respect 

 to locus plurality, phenocopies, penetrance, somatic mutation, or other 

 things that give spurious frequencies of affected individuals. These are 

 known to everyone, and they are, I think, really impossible to assess. 



As to the determination of relative fertility, there have been different 

 opinions about what controls are appropriate. An apparent mutation 

 rate has been revised by as much as a factor of 5, depending on which 

 extreme view was adopted (35) . 



So we have an unknown amount of uncertainty caused by things 

 other than gene frequency that influence the frequency of affected in- 

 dividuals, and possibly a factor of 5 due to uncertainty of the relative 

 fertility, and factors of from 100 to 1000 involved in the direction of 

 the mutation which in man must almost always be forward in the 

 sense that many sites within a gene contribute to the rate, rather than 

 just a single one. 



Neel: Aren't back mutations a rather special case? Surely in man, 

 and in Drosophila, unless specifically stated to the contrary, the muta- 

 tion rates are forward mutations. I don't see where this problem of the 

 difference between forward and back mutation has entered into the 

 discussion so far. 



Atwood: It comes in in this way, that because it is easy to look for 

 forward mutations in man and easy to look for back mutations in 

 microorganisms, you expect the reported rates to show a discrepancy. 



Lederberg: For your purposes, the ranges are not so very different. 

 There are recessive losses of function in the bacteria that have rates 

 anywhere from lO"*" to 10"^^. Streptomycin resistance happens to be in 

 that category; so there may very well be clusterings to higher fre- 

 quencies for losses and for returns of function, presumably based on the 

 implicit risk. I don't see how that is going to affect your argument 

 seriously. 



Neel: If you could exclude from the bacterial data the back mutation 

 rates, then what would the average be? 



Atwood: Only four such values are included. 



