Abstract Background: PPARγ is a transcription factor important for adipogenesis and adipocyte differentiation. Data from animal studies suggest that PPARγ may be involved in breast tumorigenesis, but results from epidemiologic studies on the association between PPARγ variation and breast cancer risk have been mixed. Recent data suggest that soy isoflavones can activate PPARγ. We investigated the interrelations of soy, PPARγ, and mammographic density, a biomarker of breast cancer risk in a cross-sectional study of 2,038 women who were members of the population-based Singapore Chinese Health Study Cohort. Methods: We assessed mammographic density using a computer-assisted method. We used linear regression to examine the association between 26 tagging single-nucleotide polymorphisms (SNP) of PPARγ and their interaction with soy intake and mammographic density. To correct for multiple testing, we calculated P values adjusted for multiple correlated tests (PACT). Results: Out of the 26 tested SNPs in the PPARγ, seven SNPs were individually shown to be statistically significantly associated with mammographic density (PACT = 0.008–0.049). A stepwise regression procedure identified that only rs880663 was independently associated with mammographic density which decreased by 1.89% per-minor allele (PACT = 0.008). This association was significantly stronger in high-soy consumers as mammographic density decreased by 3.97% per-minor allele of rs880663 in high-soy consumers (PACT = 0.006; P for interaction with lower soy intake = 0.017). Conclusions: Our data support that PPARγ genetic variation may be important in determining mammographic density, particularly in high-soy consumers. Impact: Our findings may help to identify molecular targets and lifestyle intervention for future prevention research. Cancer Epidemiol Biomarkers Prev; 21(4); 635–44. ©2012 AACR.