Properties of linkage disequilibrium (LD) maps - PubMed (original) (raw)

Properties of linkage disequilibrium (LD) maps

Weilhua Zhang et al. Proc Natl Acad Sci U S A. 2002.

Abstract

A linkage disequilibrium map is expressed in linkage disequilibrium (LD) units (LDU) discriminating blocks of conserved LD that have additive distances and locations monotonic with physical (kb) and genetic (cM) maps. There is remarkable agreement between LDU steps and sites of meiotic recombination in the one body of data informative for crossing over, and good agreement with another method that defines blocks without assigning an LD location to each marker. The map may be constructed from haplotypes or diplotypes, and efficiency estimated from the empirical variance of LD is substantially greater for the rho metric based on evolutionary theory than for the absolute correlation r, and for the LD map compared with its physical counterpart. The empirical variance is nearly three times as great for the worst alternative (r and kb map) as for the most efficient approach (rho and LD map). According to the empirical variances, blocks are best defined by zero distance between included markers. Because block size is algorithm-dependent and highly variable, the number of markers required for positional cloning is minimized by uniform spacing on the LD map, which is estimated to have approximately equal 1 LDU per locus, but with much variation among regions. No alternative representation of linkage disequilibrium (some of which are loosely called maps) has these properties, suggesting that LD maps are optimal for positional cloning of genes determining disease susceptibility.

PubMed Disclaimer

Figures

Fig 1.

Fig 1.

Graph of the LD map of 6p21.3 (A) and meiotic recombination (B) reported by Jeffreys et al. (8), oriented from pter to qter. The dotted line is a rough estimate of recombination within the major blocks defined by recombination hot spots as centimorgan (cM)/Mb = 0.04. Such low levels make definition of the small steps arbitrary and, therefore, of doubtful utility for positional cloning.

Fig 2.

Fig 2.

Graph of the LD map of 5q31 (A) and comparison with the 11 blocks (B) inferred from latent variables by Daly et al. (9). This is a more typical region, not selected by recombination hot spots. It illustrates the high frequency of small steps (e.g., between blocks 2 and 3, 5 and 6, and 8 and 9) and, therefore, the subjectivity of block definition. It remains to be established whether splitting or lumping is more favorable to positional cloning, or irrelevant.

Similar articles

Cited by

References

    1. Haldane J. B. S. & Smith, C. A. B. (1947) Ann. Eugen. 14, 10-31. - PubMed
    1. Mohr J., (1954) A Study of Linkage in Man (Munksgaard, Copenhagen).
    1. Morton N. E. (1956) Am. J. Hum. Genet. 8, 80-96. - PMC - PubMed
    1. Ohta T. & Kimura, M. (1969) Genet. Res. 13, 47-55.
    1. Cepellini R., Curtoni, E. S., Mattiuz, P. L., Miggiano, V., Scudeller, G. & Serra, A., (1967) Histocompatibility Testing (Munksgaard, Copenhagen).

Publication types

MeSH terms

LinkOut - more resources