High density lipoprotein cholesterol (HDL-C) isinversely associated with coronary heart disease and has a genetic component; however, linkage to HDL-C is not conclusive. Subfractions of HDL, such as HDL 3 -C, may be better phenotypes for linkage studies. Using HDL 3 -C levels measured on 907 Framingham Heart Study subjects from 330 families around 1987, we conducted a genome-wide variance components linkage analysis with 401 microsatellite markers spaced ف 10 centimorgan (cM) apart. Nine candidate genes were identified and annotated using a bioinformatics approach in the region of the highest linkage peak. Twentyeight single nucleotide polymorphisms (SNPs) were selected from these candidate genes, and linkage and family-based association fine mapping were conducted using these SNPs. The highest multipoint log-of-the-odds (LOD) score from the initial linkage analysis was 3.7 at 133 cM on chromosome 6. Linkage analyses with additional SNPs yielded the highest LOD score of 4.0 at 129 cM on chromosome 6. Family-based association analysis revealed that SNP rs2257104 in PLAGL1 at ف 143 cM was associated with multivariable adjusted HDL 3 ( P ؍ 0.03). Further study of the linkage region and exploration of other variants in PLAGL1 are warranted to define the potential functional variants of HDL-C metabolism. The inverse association of high density lipoprotein cholesterol (HDL-C) concentrations and risk of coronary heart disease was initially observed in the 1960s and 1970s and is now well established (1). Several studies have provided significant evidence supporting strong familial aggregation of HDL-C levels, but inconsistent results have been reported on whether this is attributable to a gene of major effect (2-6).HDL is a heterogeneous mixture of particles of different sizes, densities, and compositions (7,8). Differential chemical precipitation separates HDL particles into two major subfractions: HDL 2 and HDL 3 . As thoroughly reviewed by Wilson (9), a number of studies have found that lower concentrations of HDL 2 -C and HDL 3 -C are significantly associated with increased coronary heart disease risk, suggesting an important role of these two subfractions in coronary heart disease. However, there have been inconsistent findings across populations. To obtain a better understanding of the genetic influences on HDL levels, some studies have examined the more metabolically homogenous subfractions. A segregation study using families of patients who underwent coronary arteriography found that a major gene explains 34% of the total variation in HDL 3 -C and 9% in HDL-C, but the variation in HDL 2 -C is likely mostly the result of environmental factors (10). Using gradient gel electrophoresis to further differentiate three subclasses (HDL 3a , HDL 3b , and HDL 3c ) within HDL 3 and two subclasses (HDL 2a and HDL 2b ) within HDL 2 , a genome-wide linkage study (11) was performed on the five subclasses, and significant linkage [log-of-the-odds (LOD) Ͼ 3] to HDL 2a and suggestive linkage (LOD Ͼ 2) to HDL 3a and...