Although we had a moderately large sample available for analysis, our study is limited by the lack of a separate replication sample. Our sample was, as noted, ascertained on the basis of other SD traits or lack thereof, whereas many other studies have selected for ND subjects. It is not clear to what extent this is a limitation for identifying risk alleles, but it is a difference with respect to many other (but far from all) published ND studies. Also, we note that the SAGE sample (but not the Yale-Penn sample) has been studied previously in GWAS with respect to several ND-related traits.
