high LD. This indicated that including markers that are not in strong LD ensures that the haplotype frequencies are closer to those expected under the linkage equilibrium assumption and thus may help to reduce false-positive linkage findings. We also found that the LD effect is severe only when the majority of the markers being jointly examined are in strong LD. Single-point linkage analysis is not affected by LD. Therefore, given the relatively accurate allele frequencies that can readily be obtained for single marker, single-point linkage analysis can be used as a check for any suspicious false positives by comparing results to multipoint analysis. However, when a very large number of SNPs are studied, a possibility remains that allele frequency estimates for individual SNPs might be biased perhaps either by unrecognized strong stratification in the sample or by nonrandom errors introduced during processing. A potential further check is the confirmation of linkage at multiple SNPs in a region, as well as absence of linkage signal for most of the remainder of the genome. With current advances in high-throughput genotyping technology, high density marker data are easily generated. Caution must be taken when applying traditional linkage analysis to dense markers where strong
