Human Molecular Genetics, 2002, Vol. 11, No. 12 1477-1485
© 2002 Oxford University Press
Common haplotypes in five genes influence genetic variance of LDL and HDL cholesterol in the general population
1Franz Volhard Clinic, HELIOS Kliniken, Berlin, Germany, 2Max Delbrück Center for Molecular Medicine, Medical Faculty of the Charité, Humboldt University of Berlin, Germany and 3ValiGen SA, Tour Neptune 92086 Paris-La-Defense, France
Received March 1, 2002; Accepted March 27, 2002
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
We studied the association between common haplotypes in six relevant lipid metabolism genes with plasma lipid levels. We selected single-nucleotide polymorphisms (SNPs) in the cholesterol ester transfer protein (CETP), lipoprotein lipase (LPL), hepatic triglyceride lipase (HL), low-density lipoprotein cholesterol receptor (LDLR), apolipoprotein E (ApoE) and lecithin-cholesterol acyltransferase (LCAT) genes, and studied 732 individuals from 184 German families. Total cholesterol (TC), low-density lipoprotein cholesterol (LDL) and high-density lipoprotein cholesterol (HDL) were similar to those reported in other European and American populations. Haplotypes derived from SNP combinations resulted in more significance and of a higher degree than did single SNPs in the genotype–phenotype association analysis. Reduction of the polygenic variance attributable to haplotypes was estimated using variance components analysis. Under the biometrical genetic model, allelic association of haplotypes was highly significant for HDL, LDL and the LDL/HDL ratio. The residual kinship correlation was reduced accordingly. The ApoE gene had a strong effect on trait variation; however, the other genes also contributed substantially. An epistatic interaction could not be demonstrated in this sample. The data are consistent with the notion that common genetic variants influence common traits.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Cardiovascular disease is the leading cause of death worldwide, and serum lipid concentrations are a major risk factor (1). Major mutations have been described in genes coding for the low-density lipoprotein cholesterol receptor (LDLR), apolipoprotein B and, more recently, LDLR adaptor protein (2,3). These conditions are highly informative, but rare. Important but uncommon variants in the genes for the cholesterol ester transfer protein (CETP), lipoprotein lipase (LPL), hepatic triglyceride lipase (HL), lecithin-cholesterol acyltransferase (LCAT) and apolipoprotein E (ApoE) have also been described (4–8). Lipid concentrations in the general population, even in the absence of major mutations, are also highly heritable (9). Nevertheless, the identification of new sequence variations contributing to dyslipidemia has proved difficult (10). We modeled the interactions of six of the above lipid metabolism genes by means of differential equations (11). We tested the model by simulating the effects of known disease mutations as well as the effects of a high-fat diet, and observed that the predictions corresponded very well to published measurements. The aim of the present study was to estimate the genetic variance in lipid concentrations contributed by common SNPs and haplotypes in six of the above established gene loci in normal persons. We employed a genetic–biometric model of variance decomposition (12–16). We studied German families phenotyped in terms of total cholesterol (TC), low-density lipoprotein cholesterol (LDL), high-density lipoprotein cholesterol (HDL), the LDL/HDL ratio and triglyceride (TGL) concentrations.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
The demographic data and the raw lipid data are shown in Table 1. As expected, cholesterol and body mass index (BMI) were greater in the parental generation. Men had higher cholesterol values than women – with the exception of HDL, where matters were reversed. These effects were significant in regression analyses (data not shown) for age (LDL, HDL and LDL/HDL ratio), age2 (TC, LDL and LDL/HDL ratio) and gender (TGL, LDL, HDL and LDL/HDL ratio). We next examined the within and between-family variation with an analysis of variance, as shown in Table 2. The variation between families was greater than within families – as we should expect with strong genetic and/or shared environmental influence on the trait. The TGL relationship was somewhat weaker, since this variable is more strongly influenced by the last meal and the time of day at which the blood is obtained.
|
|
The SNPs and their frequency distribution in family founders are shown in Table 3. The observed allele frequency of SNPs reported as ‘common’ ranged from rare to frequent (0–49%). Hardy–Weinberg equilibrium was the case for all the SNPs, supporting the existence of random mating in this population. Figure 1 is the genomic map of the SNPs obtained from public databases (17) in four of the genes. Only one out of three SNPs in LCAT was polymorphic at a low frequency. LCAT is therefore not shown. The major ApoE haplotypes have been described earlier. Figure 2A shows the linkage disequilibrium of 55 SNP pairs expressed as Pearson correlation coefficient
. The average linkage disequilibrium decreased to 50% after about 25 kb, with HL contributing more ‘rapid’ and the other loci less ‘rapid’ decay. Figure 2B gives the haplotype counts in unrelated founders.
|
|
|
Table 4 shows an analysis of the five gene loci in terms of their individual influence on the phenotypes. A P-entry indicates which of the loci is by itself able to account for a given portion of the polygenic variance. The haplotypes gave persistently more significant results than single SNPs. Furthermore, some haplotypes showed significances that were not detected by the SNPs. CETP influenced HDL and LDL/HDL ratio. LPL influenced HDL. LDLR influenced LDL and LDL/HDL ratio. ApoE exerted influence on all the phenotypes. HL influenced LDL and LDL/HDL ratio.
|
Table 5 displays the coefficients of additive genetic variance of haplotypes at those loci that showed a significant effect in Table 4. Some of these coefficients differ significantly from zero. Others are either close to zero or are based on an insufficient number to obtain statistical significance. Only a few instances of significant stratification were observed for alleles that showed no total association (with one exception). In most cases, such as the effect of LDL receptor haplotypes on LDL/HDL ratio for example, the most frequent haplotypes formed a mean effect ‘center’ with value close to zero (+0.03 units). The less common LDL receptor haplotypes grouped around the center with a raising or lowering effect on the phenotype (between +0.34 and -0.40 units). However, in the case of CETP haplotypes on HDL, the most frequent haplotype was at the ‘periphery’ with lowered HDL values (-0.26 units), whereas the ensemble of less frequent haplotype variants moved HDL in the opposite direction.
|
Table 6 shows the combined association effect of all SNPs or all haplotypes on the various phenotypes. Again, the SNP haplotypes were more informative than single SNPs, even with the different degrees of freedom. The likelihood ratio tests were highly significant for HDL, LDL, and LDL/HDL ratio, but not for TC and TGL. Comparison of columns 3 (all loci) and 4 (without ApoE) demonstrates that the P-level drops, but is generally still significant, when association of alleles other than ApoE is tested. This finding indicates that ApoE had a strong influence on those traits. However, the other genes also influenced the lipid phenotypes considerably. This result is in keeping with the likelihood ratio tests shown in Table 4.
|
We next compared the maximum-likelihood estimates of environmental (Ve) and polygenic (Vg) variance provided by the null model, which includes no genotypes, and the full model, which incorporates the SNPs and the haplotypes (Vreg). The comparison permits an assessment of the contributions of genetic and non-genetic factors to the phenotype variation. The ‘environmental’ component Ve was consistently around 60% of the total variance. The inclusion of allelic association effects on HDL and LDL phenotypes in the full model reduced the residual kinship resemblance (Vg residual) substantially. A striking feature was that the combination of LDL and HDL as a ratio reduced the estimate of the residual resemblance Vg to 5%. The quantity Vreg may be inflated by the presence of polymorphisms without association effect. Therefore we next applied a correction for overfitting (Vreg adjusted). Comparison of this adjusted Vreg with the residual Vg (columns 8 and 10 of the top part of Table 6) shows that the contribution of associated haplotype alleles to the genetic variance is substantial for HDL and LDL and even dominant for the LDL/HDL ratio. This finding suggests that the LDL/HDL ratio describes a large amount of the genetic variation of the normal lipoprotein variability in this sample.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Our results show the influence of genetic diversity on the complex network of lipoprotein metabolism in the confines of common SNPs and haplotypes. SNP haplotypes were more informative than SNPs considered separately. In our analysis, the ‘environmental’ effect of the cholesterol phenotypes ranged between 54% and 71% and was greater than that observed in previous studies (9,18). This finding may be related to sample heterogeneity across all Germany and variation in laboratory techniques. Haplotype variation in five lipid genes attracted a substantial part of the genetic variance for HDL, LDL and LDL/HDL ratio. This finding was highly significant.
We observed a robust haplotype effect on the genetic variance of the LDL/HDL ratio. This observation is of particular interest, because the LDL/HDL ratio is particularly clinically relevant. We hasten to indicate that the result is dependent on the genetic model employed and subject to sampling uncertainty. The well-known influence of ApoE was confirmed. However, other loci also contributed significantly to the allelic association with traits, independent of ApoE. We could not find epistatic effects, because nonlinear terms were not significant in our models (not shown).
We used the family structure for haplotype construction (19). The ApoE gene had only three haplotypes, since we studied only the two widely known SNPs in that gene for the purpose of providing a positive control. Within the LCAT gene only one out of three SNPs showed slight variability. Thus, this gene was of no utility for our haplotype analysis. In CETP, LDLR, HL and LPL, the number of observed haplotypes in our population was far less than predicted by the linkage equilibrium assumption. This finding is in keeping with other recently published data (20–24). Furthermore, we extend to lipid-relevant genes the recent findings on other loci that the genetic diversity is characterized by an erratic pattern of SNP-pair linkage disequilibrium. This linkage disequilibrium decays within a distance of about 25 kb. Full linkage disequilibrium appeared only when the distance was 1 kb or less in our sample of ‘founder’ individuals. However, only a few of the theoretically possible haplotypes were actually observed and accounted for about 90% of the sample.
Many previous epidemiological studies (9) – including our own on monozygotic/dizygotic twins (18) – have shown that the basic indicators of human lipid physiology are subject to a strong genetic causal component. Whether this common physiological diversity is caused by common genetic variants as suggested by Collins et al. (25), or whether the diversity is due to numerous rare variants of greater influence, is unknown. If the former hypothesis is true, then the few haplotypes found in a population-wide family-based sample are obvious candidates for functional alleles or are markers of functional alleles. The former is probably the case with the haplotypes of ApoE.
In our family analysis, we were able to reject with great confidence the null hypothesis that the genetic variants are not correlated with lipid phenotypes. We found by variance decomposition that there were family clusters of lipid traits, because the between-family variation was almost twice as strong as the within-family variance. These data support the degree of familial aggregation due to genetic and non-shared environmental influence estimated by Higgins (9). We took care to eliminate families with monogenic conditions from our study. The selected families were recruited from an index person who had atherosclerosis determined by cardiac catheterization. That index person was not included in this analysis. This creates a sampling bias against families without such a relative. The subjects from the selected nuclear families in our analysis did not undergo catheterization to exclude atherosclerosis. However, none had symptoms to warrant that diagnostic procedure, which is an ascertainment bias in the opposite direction, in favor of families with smaller cardiovascular risk. We think that these effects balance each other and the net bias is small, for four reasons: (i) in the majority of cases, the nuclear family was sampled from the unrelated spouses of the index person; (ii) our sample lipid values are very similar to those of representative studies such as Framingham in the USA and the PROCAM study in Germany (26,27); (iii) the index person was never a member of a selected nuclear family – she/he rather was, in a minority of cases, a second- or third-degree relative; (iv) since cardiovascular disease is the commonest cause of death worldwide (1), being related to such an individual is not unusual in the general population. We are therefore confident that the results are relevant for lipid variability in the general population, albeit not exactly representative for a population-based sample.
The family sample was of sufficient size to show (at least for the most frequent haplotypes) to what extent and in which direction the lipid values were different from the average. This observation supports the common-variation hypothesis for the cause of the physiologically and clinically most important plasma cholesterol fractions: HDL and LDL. It seems advisable to confirm our findings on a different sample with additional loci and more SNPs.
In general, the direction of effects is consistent with the known physiological effects of these genes. CETP serves the purpose of cholesterol-ester transfer from HDL into the ApoB-100-containing lipoproteins (VLDL/LDL). Thus, showing that this gene exercises an influence on HDL and the LDL/HDL ratio was not unexpected. A deficiency in CETP is associated with high HDL levels (4). We found that the LPL gene had a significant influence on HDL. This result is consistent with the report by Reymer et al. (28) indicating that a mutation in the gene is associated with reduced HDL values and premature atherosclerosis. We were interested to observe that haplotypes in the LDLR had a strong influence on LDL concentrations and the LDL/HDL ratio. Haplotype analyses in normolipidemic individuals from RFLP analyses of the LDLR gene have been performed previously (29,30). These studies also suggested that the gene causing familial hypercholesterolemia has an effect on LDL in the normal population. Our results extend these earlier data. Had we merely focused on single SNPs, this result would not have been identified. Earlier studies reporting negative results focused on individual variants of the gene. The strong effect of ApoE was expected. This gene and its variants serve as a positive control in our study. HL has multiple important effects on lipid metabolism. A variant in the gene has been associated with increased HDL (6). We found a significant effect of certain HL haplotypes on LDL and the LDL/HDL ratio.
To the best of our knowledge, our family study is the first multiple SNP haplotype analysis in multiple genes to examine their utility in explaining genetic variance on LDL and HDL in the general population. We studied quantitative traits. For qualitative traits an approach has been published by Ott and Hoh (31). SNPs occur approximately every 1000 bp and putatively explain differences between unrelated persons and related individuals who are not identical twins. The fact that we can address a substantial fraction of the genetic variance of HDL, LDL and LDL/HDL ratio in terms of frequent haplotypes is reassuring. It is generally expected that single common variants can be individually responsible for only minor effects. Major effects have clinical consequences that result in their detection. These variants are known, and were not the purpose of our study. Since cardiovascular death is the most common cause of death worldwide (1), the vast majority of affected individuals have lipid phenotypes that are ‘normal’, but small deviations may have a long-term effect on the risk profile of an individual.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Study population
We selected 184 extended families from our genetic fieldworking program according to methods outlined elsewhere (32). All subjects completed a medical questionnaire and were examined by their family physician or by us. Our university's committee on human subjects approved the protocol, and written informed consent was obtained from all participants. Families with familial hypercholesterolemia, familial-combined hyperlipidemia or familial hypertriglyceridemia were excluded, as were families with patients who had known secondary causes of hyperlipidemia. We focused instead on families without known primary or secondary lipid disturbances. The persons in this analysis were either second- or higher-degree-related to someone with heart disease or were recruited via the family of the spouses of a patient with heart disease. However, none had symptoms of angina pectoris or had known heart disease. The extended families were split into 222 nuclear families that contained at least two adult children. A total of 732 subjects were included in the study.
Genotyping and measurement of lipid traits
TC, HDL and TGL were measured in serum by automated methods and LDL was calculated by the Friedewald formula. Our DNA extraction procedures are outlined elsewhere (18). SNPs were selected from the literature and from public databases (17). SNP analysis was performed by DNA sequencing. Cycle sequencing with an ABI Prism 3700 Analyzer (Applied Biosystems) using the BigDye Terminator cycle sequencing kit (Applied Biosystems) has been outlined elsewhere (33,34).
TC, LDL, HDL and TGL were corrected for age and gender. The correction diminished the total variance by 18%, 12%, 12%, 16% and 14% for TC, TGL, HDL, LDL and the LDL/HDL ratio, respectively. The data were subjected to variance component analysis (12–16). This analysis requires a Gaussian normal distribution, and therefore a log-transformation was applied to the data. To make sure that this transformation did not mask association effects, we also applied the non-parametric Kruskal–Wallis test to the untransformed original data.
We first inspected the demographic data and assured ourselves that these results were similar to known population values. We next tested for familiar phenotype clustering in these data (Table 2). The 732 individuals from 222 nuclear families were subjected to variance analysis of their log-transformed phenotype values. For convenience, the latter were scaled to a mean of zero and unit variance.
Statistical genetic analysis
We further constructed haplotypes in all nuclear families using an E-M algorithm that employs the nuclear family information (19). To take advantage of the family structures, we performed an association analysis of quantitative traits by use of variance component models for family data. The analysis uses the QTDT program: http://www.well.ox.ac.uk/asthma/QTDT (12–15). We conducted a regression analysis with single SNPs or SNP haplotypes as independent variables and lipid phenotypes as dependent variables. Genotypes were scored as 0, 1 or 2 according to the number of alleles present. The candidate alleles as covariates under test get a coefficient of additive effect and modify the mean of the phenotype. The random effects are modelled into a variance–covariance structure with a diagonal variance term, Ve, which collects environmental variance and part of the polygenic background, and an off-diagonal covariance term, Vg, which collects polygenic and shared environmental correlation according to the kinship structure (13). Confounding family stratification was tested by splitting the individual genotype into a between-family and a within-family component. Significance was tested as a 
2—test derived from the likelihood ratio (LR) (12–15) with due regard to the degrees of freedom (df) compared with a model restricted by a null hypothesis.
The total variance (=1) was partitioned into Ve, Vg and, in the full model, Vreg (variance attributed to SNPs or haplotypes due to regression). As was predicted from theoretical analysis (13), the power to detect a weak additive genetic influence of single loci (typically around 10% of the total variance) as linkage effect was not sufficient in our data set.
Test hierarchy
Our six genes code for proteins whose essential role in the lipid pathways is well established. Thus, the hypothesis that we tested was whether or not genetic variation in these genes bears on the phenotypic variation. We applied a test hierarchy that first tested the null hypothesis that the phenotype was not affected by the genotype configuration at a given locus (Table 4). Only after the rejection of this global null hypothesis did we test single regression coefficients in order to decide which of the numerous individual SNPs or SNP haplotypes show significant association on its own. This strategy reduced the problem of multiple testing to five global tests and subsequent selection of contributing individual factors. Nominal P-values less than 0.05 denote association of phenotype with certain alleles at a given locus. When individually significant, the regression coefficients obtained under the full biometric model stand for an additive genetic effect of the tested allele. Their values represent the resulting deviation (either positive or negative) from the total population mean, expressed per allele in units of standard deviation of the log-transformed phenotype values.
Finally, several tests were constructed in which the full model included all single SNPs or all SNP haplotypes as independent factors. We did not eliminate genes with insignificant effects (Table 4) to maintain similar variance estimates between the traits. We tested the null model both with and without ApoE to determine whether or not significant associations in addition to that provided by ApoE were present. These tests yielded maximum-likelihood estimates of the variance components Ve (predominantly ‘environmental’), Vg (predominantly ‘polygenic’ background) and Vreg (mainly due to association). The overfitting due to redundant parameters was estimated by applying an adjustment recommended elsewhere (35).
|
ACKNOWLEDGEMENTS |
|---|
This project was supported by grants-in-aid from the Deutsche Forschungsgemeinschaft (H.K. and J.R.) and the German Human Genome Project (F.C.L. and J.R.).
|
FOOTNOTES |
|---|
* To whom correspondence should be addressed at: Franz Volhard Clinic, Wiltberg Strasse 50, 13125 Berlin, Germany. Tel: +49 30 9417 2202; Fax: +49 30 9417 2206; Email: luft{at}fvk-berlin.de
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
1 Murray, C.J.L. and Lopez, A.D. (1997) Mortality by cause for eight regions of the world: global burden of disease study. Lancet, 349, 1269–1276.[Web of Science][Medline]
2 Umans-Eckenhausen, M.A., Defesche, J.C., Sijbrands, E.J., Scheerder, R.L. and Kastelein, J.J. (2001) Review of first 5 years of screening for familial hypercholesterolaemia in the Netherlands. Lancet, 357,165–168.[Web of Science][Medline]
3
Garcia, C.K., Wilund, K., Arca, M., Zuliani, G., Fellin, R., Maioli, M., Calandra, S., Bertolini, S., Cossu, F., Grishin, N. et al. (2001) Autosomal recessive hypercholesterolemia caused by mutations in a putative LDL receptor adaptor protein. Science, 292, 1394–1398.
4 Carmena-Ramon, R., Ascaso, J.F., Real, J.T., Najera, G., Ordovas, J.M. and Carmena, R. (2001) Association between the TaqIB polymorphism in the cholesteryl ester transfer protein gene locus and plasma lipoprotein levels in familial hypercholesterolemia. Metabolism, 50, 651–656.[Medline]
5
Ikeda, Y., Takagi, A., Nakata, Y., Sera, Y., Hyoudou, S., Hamamoto, K., Nishi, Y. and Yamamoto, A. (2001) Novel compound heterozygous mutations for lipoprotein lipase deficiency. A G-to-T transversion at the first position of exon 5 causing G154V missense mutation and a 5' splice site mutation of intron 8. J. Lipid Res., 42, 1072–1081.
6
Zambon, A., Deeb, S.S., Brown, B.G., Hokanson, J.E. and Brunzell, J.D. (2001) Common hepatic lipase gene promoter variant determines clinical response to intensive lipid-lowering treatment. Circulation, 103, 792–798.
7
Miettinen, H.E., Gylling, H., Tenhunen, J., Virtamo, J., Jauhiainen, M., Huttunen, J.K., Kantola, I., Miettinen, T.A. and Kontula, K. (1998) Molecular genetic study of Finns with hypoalphalipoproteinemia and hyperalphalipoproteinemia: a novel Gly230 Arg mutation (LCAT[Fin]) of lecithin:cholesterol acyltransferase (LCAT) accounts for 5% of cases with very low serum HDL cholesterol levels. Arterioscler. Thromb. Vasc. Biol., 18, 591–598.
8
Hagberg, J.M., Wilund, K.R. and Ferrell, R.E. (2000) APO E gene and gene–environment effects on plasma lipoprotein-lipid levels. Physiol. Genom., 4, 101–108.
9 Higgins, M. (2000). Epidemiology and prevention of coronary heart disease in families. Am. J. Med., 108, 387–395.[Web of Science][Medline]
10
Garcia, C.K., Mues, G., Liao, Y., Hyatt, T., Patil, N., Cohen, J.C., Hobbs, H.H. (2001) Sequence diversity in genes of lipid metabolism. Genome Res., 11, 1043–1052.
11 Knoblauch, H., Schuster, H., Luft, F.C. and Reich J. (2000) A pathway model of lipid metabolism to predict the effect of genetic variability on lipid levels. J. Mol. Med., 78, 507–515.[Web of Science][Medline]
12 Fulker, D.W., Cherny, S.S., Sham, P.C. and Hewitt, J.K. (1999) Combined linkage and association analysis for quantitative traits. Am. J. Hum. Genet., 64, 259–267.[Web of Science][Medline]
13 Sham, P.C., Cherny, S.S., Purcell, S. and Hewitt, J.K. (2000) Power of linkage versus association analysis of quantitative traits, by use of variance components models, for sibship data. Am. J. Hum. Genet., 66, 1616–1630.[Web of Science][Medline]
14 Abecasis, G.R., Cookson, W.O.C. and Cardon, L.R. 2000. Pedigree tests of transmission equilibrium. Eur. J. Hum. Genet., 8, 545–551.
15 Abecasis, G.R., Cardon, L.R. and Cookson, W.O. (2000) A general test of association for quantitative traits in nuclear families. Am. J. Hum. Genet., 66, 279–292.[Web of Science][Medline]
16 Almasy, L. , Williams, J.T., Dyer, T.D. and Blangero, J. (1999) Quantitative trait locus detection using combined linkage/disequilibrium analysis. Genet. Epidemiol.,17, 31–36.
17 The International SNP Map Working Group. (2001) A map of human genome sequence variation containing 1.42 million single nucleotide polymorphisms. Nature, 409, 928–933.[Medline]
18
Knoblauch, H., Busjahn, A., Munter, S., Nagy, Z., Faulhaber, H.D., Schuster, H. and Luft, F.C. (1997) Heritability analysis of lipids and three gene loci in twins link the macrophage scavenger receptor to HDL cholesterol concentrations. Arterioscler. Thromb. Vasc. Biol., 17, 2054–2060.
19 Rohde, K. and Fuerst, R. (2001) Haplotyping and estimation of haplotype frequencies for closely linked biallelic multilocus genetic phenotypes including nuclear family information. Hum. Mutat., 17, 289–295.[Web of Science][Medline]
20 Reich, D.E., Cargill, M., Bolk, S., Ireland, J., Sabeti, P.C., Richter, D.J., Lavery T., Kouyoumjian, R., Farhadian, S.F., Ward, R. and Lander, E.S. (2001) Linkage disequilibrium in the human genome. Nature, 411, 199–204.[Medline]
21 Pritchard, J.K. and Przeworski, M. (2001) Linkage disequilibrium in humans: models and data. Am. J. Hum. Genet., 69, 1–14[Web of Science][Medline]
22 Abecasis, G.R. (2001) Extent and distribution of linkage disequilibrium in three genomic regions. Am. J. Hum. Genet., 68, 191–197[Web of Science][Medline]
23 Dunning, A.M., Durocher, F., Healey, C.S., Teare, M.D., McBride, S.E., Carlomagno, F., Xu, C.F., Dawson, E., Rhodes, S., Ueda, S. et al. (2000) The extent of linkage disequilibrium in four populations with distinct demographic histories. Am. J. Hum. Genet., 67, 1544–1554.[Web of Science][Medline]
24 Taillon-Miller, P., Bauer-Sardina, I., Saccone, N.L., Putzel, J., Laitinen, T., Cao, A., Kere, J., Pilia, G., Rice, J.P. and Kwok, P.Y. (2000) Juxtaposed regions of extensive and minimal linkage disequilibrium in human Xq25 and Xq28. Nat. Genet., 25, 324–328.[Web of Science][Medline]
25
Collins, F.S., Guyer, M.S. and Charkravarti, A. (1997) Variations on a theme: cataloging human DNA sequence variation. Science, 278, 1580–1581.
26 Jaross, W., Assmann, G., Bergmann, S. and Schulte H. (1994) Comparison of risk factors for coronary heart disease in Dresden and Munster. Results of the DRECAN (Dresden Cardiovascular Risk and Nutrition) study and the PROCAM (Prospective Cardiovascular Munster) study. Eur. J. Epidemiol., 10, 307–315.[Web of Science][Medline]
27 Kannel, W.B., Castelli, W.P., Gordon, T. and McNamara, P.M. (1971) Serum cholesterol, lipoproteins and the risk of coronary heart disease. The Framingham study. Ann. Intern. Med., 74, 1–12.
28 Reymer, P.W., Gagne, E., Groenemeyer, B.E., Zhang, H., Forsyth, I., Jansen, H., Seidell, J.C., Kromhout, D., Lie, K.E., Kastelein, J. et al. (1995) A lipoprotein lipase mutation (Asn291Ser) is associated with reduced HDL cholesterol levels in premature atherosclerosis. Nat. Genet., 10, 28–34.[Web of Science][Medline]
29 Schuster, H., Humphries, S., Rauh, G., Held, C., Keller, C., Wolfram, G. and Zöllner, N. (1990) Association of DNA-haplotypes in the human LDL-receptor gene with normal serum cholesterol levels. Clin. Genet., 38, 401–409.[Medline]
30 Haviland, M.B., Ferrel, R.E. and Sing, C.F. (1997) Association between common alleles of the low-density lipoprotein receptor gene region and inter-individual variation in plasma lipid and apolipoprotein levels in a population-based sample from Rochester, Minnesota. Hum. Genet., 99, 108–114.[Medline]
31 Ott, J. and Hoh, J. (2001) Statistical multilocus methods for disequilibriuim analysis in complex traits. Hum. Mutat., 17, 285–288.[Medline]
32 Schuster, H., Lamprecht, A., Junghans, C., Dietz, B., Baron, H., Nothnagel, M., Muller-Myhsok, B. and Luft, F.C. (1998) Approaches to the genetics of cardiovascular disease through genetic field work. Kidney Int., 53, 1449–1454.[Web of Science][Medline]
33 McCombie, W.R., Heiner, C., Kelley, J.M., Fitzgerald, M.G. and Gocayne, J.D. (1992) Rapid and reliable fluorescent cycle sequencing of double-stranded templates. DNA Seq., 2, 289–296.[Medline]
34 Rosenthal, A. and Charnock-Jones, D.S. (1992) New protocols for DNA sequencing with dye terminators. DNA Seq., 3, 61–64.[Medline]
35 Rawlings, J.O., Pantula, S.G. and Dickey, D.V. (1998) Applied Regression Analysis, 2nd edn. Springer-Verlag, New York.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  K. J. Gaulton, C. J. Willer, Y. Li, L. J. Scott, K. N. Conneely, A. U. Jackson, W. L. Duren, P. S. Chines, N. Narisu, L. L. Bonnycastle, et al. Comprehensive Association Study of Type 2 Diabetes and Related Quantitative Traits With 222 Candidate Genes Diabetes, November 1, 2008; 57(11): 3136 - 3144. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Zhu, H. M. Tucker, K. E. Grear, J. F. Simpson, A. K. Manning, L. A. Cupples, and S. Estus A common polymorphism decreases low-density lipoprotein receptor exon 12 splicing efficiency and associates with increased cholesterol Hum. Mol. Genet., July 15, 2007; 16(14): 1765 - 1772. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Meletiadis, S. Chanock, and T. J. Walsh Human Pharmacogenomic Variations and Their Implications for Antifungal Efficacy Clin. Microbiol. Rev., October 1, 2006; 19(4): 763 - 787. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. M. Mutch, W. Wahli, and G. Williamson Nutrigenomics and nutrigenetics: the emerging faces of nutrition FASEB J, October 1, 2005; 19(12): 1602 - 1616. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Posada, T. J. Maxwell, and A. R. Templeton TreeScan: a bioinformatic application to search for genotype/phenotype associations using haplotype trees Bioinformatics, May 1, 2005; 21(9): 2130 - 2132. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. R. Templeton, T. Maxwell, D. Posada, J. H. Stengard, E. Boerwinkle, and C. F. Sing Tree Scanning: A Method for Using Haplotype Trees in Phenotype/Genotype Association Studies Genetics, January 1, 2005; 169(1): 441 - 453. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Isaacs, F. A. Sayed-Tabatabaei, O. T. Njajou, J. C. M. Witteman, and C. M. van Duijn The -514 C->T Hepatic Lipase Promoter Region Polymorphism and Plasma Lipids: A Meta-Analysis J. Clin. Endocrinol. Metab., August 1, 2004; 89(8): 3858 - 3863. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. C. Luft Geneticism of Essential Hypertension Hypertension, June 1, 2004; 43(6): 1155 - 1159. [Full Text] [PDF]
|
|
|
|
|
|
H. Knoblauch, A. Bauerfeind, M. R. Toliat, C. Becker, T. Luganskaja, U. P. Gunther, K. Rohde, H. Schuster, C. Junghans, F. C. Luft, et al. Haplotypes and SNPs in 13 lipid-relevant genes explain most of the genetic variance in high-density lipoprotein and low-density lipoprotein cholesterol Hum. Mol. Genet., May 15, 2004; 13(10): 993 - 1004. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Morabia, E. Cayanis, M. C. Costanza, B. M. Ross, M. S. Flaherty, G. B. Alvin, K. Das, and T. C. Gilliam Association of extreme blood lipid profile phenotypic variation with 11 reverse cholesterol transport genes and 10 non-genetic cardiovascular disease risk factors Hum. Mol. Genet., November 1, 2003; 12(21): 2733 - 2743. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. M. Boekholdt and J. F. Thompson Natural genetic variation as a tool in understanding the role of CETP in lipid levels and disease J. Lipid Res., July 1, 2003; 44(6): 1080 - 1093. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||