ABSTRACT
Genome-wide linkage studies have been extensively used to identify chromosomal regions which may harbor susceptibility genes for complex diseases. Early enthusiasm for such studies has been replaced by the realization that most complex disease genes have only a minor effect on risk, and consequently many linkage studies have low power to detect such genes (Risch and Merikangas, 1996). This phenomenon was well illustrated by a compilation of 101 genome-wide linkage studies in 31 diseases, which found that few studies achieved significant evidence for linkage, and there was little replication within each disease (Altmu¨ller et al., 2001). Replication of linkage is an important concept in genome-wide linkage studies: that two studies produce high (if not significant) lod scores in the same approximate region lends further weight to the results. This ad hoc method of comparing results across studies is formalized in meta-analysis, which provides statistical evidence for the co-localization of linkage evidence across studies. Meta-analysis can also provide a solution to the lack of power in individual studies: combining weak evidence of linkage from several studies may show an overall significant effect.
