Polymorphic sites adjacent to known genes can be used to examine the segregation of a disease relative to that gene in families, or to map the gene of interest relative to other loci. The polymorphic region 5' to the human insulin gene (5' FP) permits such analysis, but the three size classes previously identified are insufficient for many studies. More alleles are identified with restriction enzymes that generate small fragments (Pvu II). Nonetheless, sufficient polymorphism for informative family analyses is often not present.
To facilitate such analyses, we searched for other polymorphisms in over 20 KB of DNA at the insulin locus in Pima Indians, American Blacks, and Caucasians. The previously described allelic variant at a Pst I site in the 3'-untranslated portion of the gene was not polymorphic in any race. An upstream Hinc II site (−62 BP) was present in only 48% of Black alleles, but was not polymorphic in Pima Indians or Caucasians. New polymorphisms were found at a Taq I site (−11,000 BP) and a Rsa I site (−13,000 BP). The Taq I site was present in 89% of Black alleles, 87% of Pima Indian alleles, and 84% of Caucasian alleles. In contrast, the Rsa I site was present in 60% of Black and Caucasian alleles, but in only 47% of Pima Indian alleles. The Hinc II, Rsa I, and Taq I sites show no obvious linkage with each other or the 5' FP.
A fourth polymorphism, previously identified with Sac I, was found to be the creation of a new Sac I site at +2500 BP in 10% of Black alleles. This site was present only in linkage with the presence of a Hinc II site and certain sizes of the 5' FP, and only in Blacks. The usefulness of the new polymorphisms in the 5' FP (Pvu II digest) results in an increase in the number of matings that would be informative for linkage of this locus with a disease from 70% to 93%. Thus, the addition of these markers to the 5' FP permits informative linkage analysis in almost all families.
