Single-point linkage analysis did not show any evidence of linkage both for the three markers in strong LD alone and for the whole marker set (Fig. 1). However, using the three markers that are in strong LD and affected sib-pair only data, multipoint linkage analysis showed false-positive evidence of linkage for both model-free and parametric linkage analyses that incorporated a heterogeneity parameter (Fig. 1). This confirmed the observation by Huang et al. [6]. Including parents in the multipoint analysis eliminated the false-positive evidence (data not shown). The false-positive evidence induced by LD can be gradually reduced by adding markers that are not in LD to either or both sides of the three core markers that are in strong LD, and it seemed a better "rescue" effect can be achieved by adding markers to both sides than to a single side (Table 2). With all 20 markers, there is no evidence of linkage (maximal LOD score at the peak position: 0.34 ± 0.2).
