One of the early difficulties in replicating genetic associations observed in candidate gene studies was the fact that different groups would study different markers in the same region. Because the LD among these markers was poorly understood, results from multiple studies could increase rather than decrease confusion. The initial study may have seen an association with SNP A, but the second study did not genotype that SNP, and instead saw an association with SNP B, which was not genotyped in the original study. As the number of SNPs typed per region increased, “moving the goalposts” in this fashion contributed to the problem of persistent false positives in the candidate gene literature; by chance some SNP in the region (not necessarily the SNP that was statistically significant in other studies) would have p<0.05, and this would be (incorrectly) proclaimed replication.[34] In response to this problem, guidelines for replication in genetic association studies now call for exact replication. The same marker—or, if technical difficulties preclude this, a perfect or near-perfect proxy for the original marker—should be genotyped across all studies and analyzed
