The first large genome-wide association (GWA) study of systemic sclerosis (SSc) recently included 2296 SSc patients and 5014 healthy controls from four case–control series of Caucasian individuals and involved the analysis of about 280 000 single nucleotide polymorphisms (SNPs).1 Outside the HLA region, SNPs mapping to the known TNPO3 - IRF5 and STAT4 regions showed the strongest association but three other loci reached genome-wide significance (namely CD247 locus in 1q22–23, CDH7 in 18q22 and EXOC2 - IRF4 near 6p25) although only one ( CD247 ) could be replicated in the second set of samples. Given the major challenge of separating the many false-positive associations from the few true-positive associations with disease in GWA studies, resulting in a stratification bias, a critical strategy is replication of results in independent samples.2 Therefore, we herein provide an attempt of …