Choice of enzymes for mapping based on CpG islands in the human genome

Abstract

The frequencies of sites for rare-cutting restriction enzymes in 2.9 million bp of human genomic DNA sequence in the EMBL database have been determined and compared with the expected frequencies. Rare cutters can be divided into four groups based on certain features of their recognition sites. Mlu, I, Nru I, Spl I, and Pvu I are predicted to cleave genomic DNA most infrequently, which is borne out by the fragment lengths observed for Mlu I and Nru I. Thus, these four enzymes are ideal for making long-range maps based on pulsed-field electrophoresis. Other enzymes like Not I are useful for making more detailed maps. Finer maps for identification of CpG islands and associated genes should involve several rare cutters including Eag I, Sac II and Bss HII. A cluster of sites for at least two such enzymes is a good indicator of a CpG island, and 78% of the island-associated genes can be located in this way.