As expected, the analysis based on the SNP subset showed decreased IC and reduced linkage signals compared with the SNP full set, which suggested that the difference in IC might be one key factor that contributed to the observed difference in the two scans. This was consistent with the conclusion from John et al. [2], who examined possible reasons for the observed difference between the scans using the SNPs and the microsatellites comprehensively, including the genotyping errors of the SNPs and the microsatellites, the possible errors in the two maps used, the presence of linkage disequilibrium (LD), and the differences in IC. We have also investigated the possible effect of genotyping errors on the linkage results. Out results suggested that the impact of genotyping errors was quite small for the COGA dataset, which may be due to the small genotyping error rate (0.37% for the microsatellites and 0.20% for the SNPs) and may not be generalized to other data sets. It is worth noting that for the full SNP set with an average spacing of 0.35 cM, it is highly
