Clinical evidence indicates that polycystic ovary syndrome (PCOS) has a heritable basis (1, 2), at least in part, which could result from a genetic etiology, epigenetic changes, or an admixture of the two causes. The apparent heritable nature of the syndrome has led investigators to search for the gene (or genes) that contribute to development of the PCOS phenotype. One impetus for genetic studies of PCOS was to unravel the etiology of the syndrome and provide greater precision to the diagnosis. Unfortunately, the phenotypic variability of the syndrome represents a major hindrance to that goal because the PCOS phenotype is based on criteria marked by a constellation of features, currently defined by the Rotterdam criteria (3). Thus, the PCOS phenotype is not a clearly delineated disease entity, and there is imprecision of diagnosis. Patients diagnosed with PCOS may satisfy the diagnostic criteria by having different features of the disease. Stated differently, PCOS can be conceptualized as a pathophysiological state of acyclic ovarian function rather than a single distinct disorder. This imprecision in diagnosis raises the strong possibility that PCOS with different root causes will be lumped into the same “diseased category” for comparison with normal controls. Genetic studies of PCOS are also confounded by variability between different investigators, especially if the study design or diagnostic criteria used across studies are not the same. The group of Ewens et al. (4) has used standard diagnostic criteria throughout their studies (oligomenorrhea and increased testosterone levels) to distinguish cases from controls. A potential problem is that any variability in diagnostic criteria between investigators may impact the ability of putative candidate genes identified in one population to be replicated in other populations if the initial population studied was phenotypically dissimilar, as a result of different diagnostic criteria. Investigators conducting genetic studies of PCOS have used several different approaches in different populations. Some have used a candidate gene approach. Other investigators have used linkage studies, either of case-control or family-based sib pairs. Association studies have also been conducted. Another approach is the transmission disequilibrium test—briefly, a family-based test of association designed to minimize error due to population stratification (5). Despite a decade of collaborative efforts, including analysis of 37 candidate genes (6), efforts to identify loci associated with PCOS have been plagued by a lack of replication between association and linkage studies (reviewed in Ref. 7). It seems clear that neither follistatin (8) nor lamin a/c (9) is a critical PCOS gene. So what genes are associated with PCOS? Surprisingly, in 2010 the best evidence of an association of a single gene (locus) with PCOS is the dinucleotide repeat microsatellite marker D19S884 (4, 6, 10–13). The D19S884 marker lies within intron 55 of the fibrillin-3 gene (FBN3, NM_032447). Allele 8 of the D19S884 marker contains 17 “GT” repeats and has been associated with PCOS. In contrast, allele 9 containing 18 GT repeats and allele 13 with 22 GT repeats show no association with PCOS. Fibrillin is a large extracellular matrix protein that forms microfibrils and contains epidermal growth factorlike modules, some of which have a calcium binding consensus sequence (calcium binding-epidermal growth factor-like modules). Related fibrillin family members regulate TGF action, but the reason as to why fibrillin-3 hasbeenassociatedwithPCOSremainsunclear, especially