Restriction sites containing CpG show a higher frequency of polymorphism in human DNA.

Unique loci in the human genome were examined with restriction enzymes in order to detect restriction fragment length polymorphisms (RFLPs). Of 31 arbitrary loci, nine were detectably polymorphic, reflecting ten polymorphic restriction sites. Nine of the ten polymorphic sites were revealed with two restriction enzymes, Msp I and Taq I, whose recognition sequences have in common the dimer sequence CpG. The cytosines in the CpG sequence are known to be frequently methylated in mammals, and the occurrence of significant variation in Msp I and Taq I sites supports the view that methylated cytosine residues are hotspots for mutation in mammalian DNA.