Patterns of RFLP association were studied, to identify gene regions influencing quantitative variation in lipid and lipoprotein traits (coronary artery disease [CAD] risk factors or metabolically related traits). Subjects (118 female and 229 male; age 20-59 years) were selected for health. Multiple RFLPs were used to sample variability in regions around genes for apolipoprotein (apo) B (restriction enzymes HincII, PvuII, EcoRI, and XbaI), apo AI-CIII-AIV (BamHI, XmnI, TaqI, PstI, SstI, and PvuII) and cholesterol ester transfer protein (TaqI). Separate analyses were done by gender. The sample was truncated at mean +/- 4 SD, to remove extreme outliers. There was no significant gender difference in RFLP genotype frequency distribution. After trait-level adjustment to maximize removal of concomitant variability, analysis of variance was used to estimate the percentage trait phenotypic variance explained by measured variability in the gene regions studied. Fewer gene regions were involved in m en, with less influence on quantitative trait variation than in women, in whom hormone use affected association patterns. Gender differences imply that pooling genders or adjusting data for gender effects removes genetic information and should be avoided. The association patterns show that variability around the candidate genes modulates trait levels: the genes are contributors to the genetics of CAD risk variables in a healthy sample.
