New population-based exome data question the pathogenicity of some genetic variants previously associated with Marfan syndrome

Background: Marfan syndrome (MFS) is a rare autosomal dominantly inherited connective tissue disorder with an estimated prevalence of 1:5,000. More than 1000 variants have been previously reported to be associated with MFS. However, the disease-causing effect of these variants may be questionable as many of the original studies used low number of controls. To study whether there are possible false-positive variants associated with MFS, four in silico prediction tools (SIFT, Polyphen-2, Grantham score, and conservation across species) were used to predict the pathogenicity of these variant. Results: Twenty-three out of 891 previously MFS-associated variants were identified in the ESP. These variants were distributed on 100 heterozygote carriers in 6494 screened individuals. This corresponds to a genotype prevalence of 1:65 for MFS. Using a more conservative approach (cutoff value of >2 carriers in the EPS), 10 variants affected a total of 82 individuals. This gives a genotype prevalence of 1: 79 (82: 6494) in the ESP. A significantly higher frequency of MFS-associated variants not present in the ESP were predicted to be pathogenic with the agreement of >= 3 prediction tools, compared to the variants present in the ESP (p = 3.5 x 10(-15)). Conclusions: This study showed a higher genotype prevalence of MFS than expected from the phenotype prevalence in the general population. The high genotype prevalence suggests that these variants are not the monogenic cause of MFS. Therefore, caution should be taken with regard to disease stratification based on these previously reported MFS-associated variants.