The genetic analysis of familial traits

Summary

• 1. If familial trait is determined by a single recessive gene

  $$\frac{R}{{\Sigma \frac{{s. n_s }}{{1 - q^s }}}} = p$$

  (1)

  wherep is the probability that the offspring of two heterozygous parents will be recessive,R is the total number of recessives in observed families of at least one recessive,n _s is the number of observed s-membered fraternities containing at least one recessive, and (p + q) = 1.

• 2. The standard deviation ofp determined in this way is

  $$\frac{{\sqrt {\Sigma n_s k_s } }}{{\Sigma n_s c_s }}$$

  (2)

  .

• 3. Tables ofc _s andk _s are given fors = 1 to 20,q being 3/4.

• 4. In general the proportion of recessives in recorded families showing either autosomal or sex-linked familial traits is significantly higher than 1/4. This may be due to the fact that cases showing a high familial incidence are recorded in medical journals more frequently than cases showing a low familial incidence.

• 5. If this is so, little progress will be made in genetical analysis of pathological traits by compilation of recorded data. This conclusion is supported by the fact that two cases showing comparatively insignificant discrepancies were based on unweighted observations of individual investigators, and the only case of completely satisfactory agreement is a disorder which is not rare.