Croonian Lecture - The formal genetics of man

Abstract

Man has obvious disadvantages as an object of genetical study. The advantages are that very large populations are available, and that many serological differences and congenital abnormalities have been intensively investigated. Some characters are found to obey Mendel's laws with great exactitude. In others the deviations are such as to suggest the existence of a considerable selective mortality, perhaps prenatal. In yet other cases the observations are biased because we only know that we are investigating the progeny of two heterozygotes when the family includes at least one recessive. Statistical methods which eliminate this bias are described. Still more complex methods are needed for the detection and estimation of linkage. Several such cases have been detected with greater or less certainty, and the frequency of recombination between the loci of the genes for colour-blindness and haemophilia is now estimated at 10 $\pm $ 4%. If the theory of partial sex-linkage be accepted, it is possible to make a provisional map of a segment of the human sex chromosome. When a gene is sublethal, as are those for haemophilia and achondroplasic dwarfism, its elimination by natural selection is in approximate equilibrium with its appearance by mutation, and the frequency of the latter process can be estimated. The mutation rates at five human gene loci lie between 4 $\times $ 10$^{-5}$ and 4 $\times $ 10$^{-6}$ per locus per generation. These are the only estimates available for vertebrates. The rates per generation are rather higher than those in Drosophila, but those per day are so small that much, or even all, human mutation may be due to natural radiations and particles of high energy.