To determine if transferrin saturations in African Americans may refle
ct the presence of a gene that influences iron metabolism, we analyzed
the distribution of these values in 808 African Americans from the se
cond National Health and Nutrition Survey. We tested for a mixture of
three normal distributions consistent with population genetics for a m
ajor locus effect in which the proportion of normal homozygotes is p(2
); of heterozygotes, 2pq; of affected homozygotes, q(2); and in which
p+q = 1. Three subpopulations based on transferrin saturation were pre
sent (P < .0001) and the fit to a mixture of three normal distribution
s was good (P = .2). A proportion of .009 was included in a subpopulat
ion with a mean +/- standard deviation transferrin saturation of 63.4%
+/- 5.7% (postulated homozygotes for a gene that influences iron meta
bolism), while a proportion of .136 had an intermediate saturation of
38.0% +/- 5.7% (postulated heterozygotes) and .856 a saturation of 24.
6% +/- 5.7% (postulated normal homozygotes), These proportions were co
nsistent with population genetics because the sum of the square roots
of the proportions with the lowest mean transferrin saturation (P = .9
25) and the highest (q = 0.093) was approximately 1 (1.018). The resul
ts are consistent with the presence in African Americans of a common l
ocus that influences iron metabolism. (C) 1998 by The American Society
of Hematology.