Background: Elevated plasma lipoprotein level plays a crucial role in the development of coronary artery disease. There is growing evidence that plasma triglycerides (TG) level is related with athrogenic lipoprotein and TG level may be influenced by the interaction of lipid metabolism genes. The goal of this study was to determine whether there is evidence that the plasma TG is influenced by interactive effects of these genes and to identify partitions of multilocus genotypes that predict interindividual variations in plasma TG levels. Methods and Results: We genotyped 12 SNPs from six relevant lipid metabolism genes (ApoE, ABCA1, CETP, LIPC, ApoA5, and MTP) in a sample of 350 unrelated males (mean age 50.6 y). In a single locus analysis, only the SNPs in ApoA5 were related with TG levels (R² value 0.07, p<0.05). However, using restricted partition methods, we identified the evidence of interaction among multilocus genotypes and loci contributing epistatically to plasma TG levels. The partitioning of the multilocus genotypes into the two best sets is shown in figure and the genotype partitions, CETP Taq1B*ApoA5 G-3A and LIPC C-514T*ApoE112 C/A, explained the TG heterogeneity by 8.1% and 1.9%, respectively. Conclusion: The present study shows the gene-gene interaction effect (epistasis) on plasma TG level supporting the idea that the genetic architecture of quantitative trait is complex and a variety of genetic effects should be considered.