The use of inconclusive or incomplete haplotypic information has long been considered a major source of errors in imputed frequencies, especially in chromosomal regions or populations with limited haplotypic knowledge. To test this hypothesis, we used r2 SNP-by-SNP disequilibrium measure available in HapMap release (HapMap Public Release #22, 2007). Initially, we determined markers of our filtered dataset that were also evaluated in the HapMap database. This group is composed of 317,255 polymorphisms and it will be used in our further analysis. For each marker, a haplotypic block was determined using the complete set of markers showing a linkage disequilibrium measure to the specific marker (See methods for further details). Each haplotypic block is composed by a limited set of marker-to-marker r2 statistics. In each block, four different descriptive statistics were evaluated: mean, median, maximum value and variance of the values of the r2 statistic. We plotted separately the four statistics against the observed bias between association statistics of the complete set of markers (Figure 4). The analysis of the plotting organized in Figure 4 outlined that the use of the
