Replication of Linkage of Familial Combined Hyperlipidemia to Chromosome 1q With Additional Heterogeneous Effect of Apolipoprotein A-I/C-III/A-IV Locus
The NHLBI Family Heart Study
Abstract—Familial combined hyperlipidemia (FCHL), the most common familial dyslipidemia, is implicated in up to 20% of cases of premature coronary heart disease. Although underlying mutations for FCHL have yet to be identified, several candidate genes/regions have been identified. A positive linkage to chromosome 1q markers has been reported, with the highest lod score of 5.93 occurring at a location between D1S104 and D1S1677. Using the same diagnostic criteria, the Family Heart Study (FHS) has defined 71 FCHL families, comprising 170 cases, for a total of 137 possible affected sibling pairs. The FCHL criteria require elevation in serum low density lipoprotein cholesterol and triglyceride levels within the family, with at least 2 affected first-degree relatives. Markers D1S104 and D1S1677 were typed, and significant allele sharing was found in FCHL sibships (multipoint lod score with use of the model from the Finnish study was 2.52, and multipoint nonparametric score was 2.48; P=0.007), replicating linkage in this chromosome 1 region. In addition, previously reported linkage of FCHL to apolipoprotein A-I/C-III/A-IV has been investigated in FHS families. FHS results revealed positive but nonsignificant allele sharing among FCHL sibships with apolipoprotein A-I/C-III/A-IV by use of marker D11S4127 (nonparametric linkage score 1.11, P=0.13). Two-locus analyses of D1S104 and D11S4127 suggested possible heterogeneity rather than epistasis, with a maximum 2-locus lod score of 3.05. A nonparametric 2-locus analysis revealed significant improvement in the 2-locus versus single-locus scores. Finally, no linkage was found with markers near the lipoprotein lipase gene region.
- Received March 29, 2000.
- Accepted June 26, 2000.
Familial combined hyperlipidemia (FCHL) is the most common familial dyslipidemia, with a prevalence of 1% to 2%,1 and is implicated in up to 20% of cases of premature coronary heart disease (CHD).2 FCHL criteria require elevation of serum total cholesterol and triglyceride levels within the family, with at least 2 affected first-degree relatives.3 Studies of the genetics of FCHL have been complicated by uncertain phenotype definition, likely genetic heterogeneity, and unknown mode of inheritance. These difficulties have thus far prevented the identification of underlying mutations for FCHL. Several loci have produced positive results, although often with mixed findings in replication studies.
Perhaps the most promising candidate region is on chromosome 1q21-23. Pajukanta et al4 recently published positive linkage to chromosome 1q markers in 250 individuals (115 affected) from 31 Finnish FCHL families, with the highest lod score of 5.93 occurring between D1S104 and D1S1677. Of additional interest, Castellani et al5 have since identified a region on mouse chromosome 3, syntenic to human 1q21-23, causing traits that mimic FCHL in the HcB-19 mouse strain.
Linkage in FCHL families has also been reported to apolipoprotein A-I/C-III/A-IV on chromosome 11.6 That study ascertained 7 FCHL families through a proband carrying an allele of the XmnI restriction fragment length polymorphism, a marker in linkage disequilibrium with A-I/C-III/A-IV. Support for this finding has been reported in 18 families with premature coronary artery disease and FCHL7 ; however, the linkage to A-I/C-III/A-IV in that study was to the presence of small dense LDL particles, a trait significantly associated with FCHL. In addition to these linkage studies, positive linkage disequilibrium has also been reported between A-I/C-III/A-IV and hypertriglyceridemia.8 Negative linkage reports of A-I/C-III/A-IV in FCHL families also appear in the literature, suggesting possible genetic heterogeneity or false-positive results in the positive reports.9 10 11 One study has proposed more complex models to explain the contribution of A-I/C-III/A-IV to FCHL; these models involve epistatic interactions of specific haplotypes at the locus.12
The lipoprotein lipase (LPL) gene is also an attractive candidate for FCHL. Decreased activity of LPL in subjects with FCHL has been shown,13 and positive associations have been reported between FCHL and genetic variants in the LPL promoter and in exon 6.14 15 16 However, other studies have reported negative findings for FCHL with LPL.17 18
In the National Heart, Lung, and Blood Institute (NHLBI) Family Heart Study (FHS), there are 71 FCHL families, comprising 170 cases, for a total of 137 affected sibling pairs. Markers in all 3 of these candidate regions have been genotyped on these sibships and on a larger set of FHS families for whom lipid traits can be studied. Linkage to FCHL and interactions of the regions were tested in the NHLBI FHS sample. Our replication study used the exact genetic model of the Finnish study, which was derived from the population frequency of FCHL. In addition, we tested each region by using nonparametric methods.
Subjects were drawn from the NHLBI FHS,19 a multicenter population-based study investigating genetic and nongenetic causes of CHD, atherosclerosis, and cardiovascular risk factors. Subjects were drawn from 4 centers, including Framingham, Mass; Salt Lake City, Utah; Minneapolis, Minn; and Forsyth County, NC. After written informed consent was obtained, a total of 5381 subjects from 1245 families (657 high CHD risk and 588 random) completed an extensive clinical examination. Clinical data obtained included the following: anthropometry, blood pressure, blood samples (for laboratory tests and DNA), ECG, pulmonary function, ultrasound cardiovascular measures, and questionnaires supplying information about socioeconomic status, lifestyle, medical history, medication use, food intake, physical exercise, family structure, and psychosocial characteristics. For details, see the NHLBI FHS Manuals of Procedure.20
The primary focus of the present study was on the FCHL sibships within the FHS. To define FCHL, probands were classified into the following categories: (1) type 2a (>90th percentile in LDL cholesterol [LDL-C] and <90th percentile in triglycerides), (2) type 2b (>90th percentile in LDL-C and >90th percentile in triglycerides, and (3) type 4 (<90th percentile in LDL-C and >90 percentile in triglycerides). A sibship was classified as having FHCL if at least 2 first-degree relatives in the family were type 2b or if 2 first-degree relatives had at least 2 different FCHL phenotypes. No adjustments were made for lipid-lowering medications to avoid uncertainty in the magnitude of the medication effect. With this conservative strategy, any individuals in the FHS sample who would have been classified as affected and were not taking medications were not included in the analysis. Of the subjects with FCHL, 2 reported that they were currently on lipid-lowering medications, but they still had high enough values to be classified as affected. Ten of the FCHL subjects reported having had diabetes.
This diagnosis resulted in 71 FHS families with at least 1 sibling pair affected with FCHL (170 total cases, 137 affected sibling pairs). The distribution of affected siblings in these families was as follows: 46 sibships with 1 affected pair, 13 sibships with 3 affected cases (3 pairs), 2 sibships with 4 affected cases (6 pairs), 1 sibship with 6 affected cases (15 pairs), 2 sibships with 1 affected pair plus another half sib, 1 sibship with 1 affected pair plus 2 other half sibs, and 5 sibships with 1 affected half sib pair.
In addition, quantitative lipid traits (triglycerides, LDL-C, HDL cholesterol [HDL-C], and total cholesterol) were studied in the larger set of families in the FHS that had been genotyped for candidate gene studies. This genotyped sample included sibships drawn from FHS families having the following: (1) at least 1 CHD sib pair, (2) at least 1 pair of sibs both in the upper 80th percentile of the intima medial carotid wall thickness distribution, (3) FCHL sib pairs, (4) at least 1 hypertensive sib pair, (5) all of the African-American sample, and (6) 454 unrelated random subjects for association studies and allele frequency estimation.
For measurement of lipids, subjects were asked to fast for 12 hours before their clinic visit. Evacuated tubes with no additives were used to collect samples for lipid study. Blood samples were spun at 3000g for 10 minutes at 4°C and then stored at −70°C until sufficient numbers of samples accumulated for shipment to the Central Biochemistry Laboratory at the University of Minnesota for processing. For most subjects, LDL was estimated by use of standard methods.21 For subjects with triglyceride levels >400 mg/dL, LDL was measured by ultracentrifugation. HDL, total cholesterol, and triglyceride concentrations were measured by a Roche COBAS FARA high-speed centrifugal analyzer (Roche Diagnostic Systems). HDL-C was measured after precipitation of other lipoprotein fractions by dextran sulfate.22 Because of skewness, triglyceride concentration was logarithmically transformed before further analysis.
Other measures were used as covariates in the present study, including age, sex, body mass index, smoking, alcohol, fat intake, fruit/vegetable intake, hematocrit, estrogen use, and research center. Anthropometric measurements were collected with subjects wearing scrub suits. Weight was measured by use of a balance scale, and height was measured by use of a vertical ruler mounted to a wall. Subjects were asked to bring medications to their clinic visit, and medication use was assessed through a review of these medications and through an interview. All other variables were collected through interviews performed by trained interviewers. Individuals on cholesterol-lowering medications were omitted from the quantitative analysis. Table 1⇓ presents descriptive statistics for the subset of subjects with FCHL and all genotyped FHS subjects. The table reports lipid levels after adjustment for covariates; this adjustment is described below.
Genotyping for all markers was performed by using polymerase chain reaction amplification of genomic DNA. Primer sequences are found in the Genethon database (http://www.genethon.fr) for D1S104, D11S4127, and D8S282 and in the CHLC database (http://www.chlc.org) for D1S1677. D1S1677 is mapped 0.4 cM from D1S104; both markers were used in the Pajukanta et al4 study of FCHL. D11S4127 is mapped 0.0 cM from the A-I/C-III/A-IV gene region (lod score 13.0) according to the CEPH database (http://www.cephb.fr.ogi-bin). Marker LPL1GTR2, located in the 5′ region of the LPL gene, is described in detail in Odelberg and White.23 D8S282 is 3 cM from the LPL gene and is described in Wu et al.24 Polymerase chain reactions for the markers were carried out by use of standard techniques and under standard conditions optimized for each marker. Exact concentrations and conditions are available on request from the authors. Characteristics showing the informativeness of the markers estimated from the FHS sample are shown in Table 2⇓.
FCHL sibships were analyzed by using parametric and nonparametric methods. The parametric model was identical to that used in the Pajukanta4 study, in an effort to replicate the significant linkage of that study to the chromosome 1 region. This model was derived from the general population frequency of FCHL. The gene frequency for the disease allele was fixed at 0.006 for the dominant model and 0.1095 for the recessive model. Subjects were coded as either affected or unknown to avoid the specification of reduced penetrance. The pedigrees were analyzed by use of the FASTLINK modification25 of the LINKAGE program.26 Simulation analyses of representative markers with 75% heterozygosity were performed to determine the probability that observed results were false positives. The simulations were performed on 1000 replicates of the FCHL families by assuming the dominant linkage model with the SLINK program.27 28 Nonparametric analyses on the FCHL sample were performed with the use of GENEHUNTER.29
Two-locus analysis was performed by use of D1S104 and D11S4127. The correlation of the nonparametric linkage GENEHUNTER scores for each family at each locus was computed as described in Cox et al.30 We used the correlation to dictate the most likely model to test by use of the weighting scheme described by Cox et al. The resulting negative correlation indicated that although the data were consistent with a heterogeneity model, epistasis was unlikely. Investigation of scores revealed a pattern in which high scores for D1S104 were coupled with lower or negative scores for D11S4127, and vice versa, although negative-negative and positive-positive sibships were observed. Approximately one quarter (18 of 70) of the sibships showed positive scores for both markers, suggesting a model in which either locus might give rise to the phenotype but without an epistatic interaction. We tested this model by weighting all families with a positive score for either locus with a 1 and by weighting families with negative or zero scores at both loci with a 0. We then performed a new GENEHUNTER analysis. Significance of the increase in the score was assessed through taking the difference between the original total score and the weighted total score multiplied by 2 log(10). This statistic is distributed as a χ2 with 1 df, and probability values from this test will be conservative.30
Parametric 2-locus lod scores were also computed by use of the Pedigree Analysis Package (PAP).31 The 2 loci were assumed to be unlinked, and dominant inheritance was assumed at both loci, with affection resulting from a mutation at either locus (heterogeneity model). Again, unaffected cases were coded as unknown.
Quantitative lipid traits were analyzed on the full FHS genotyping sample by use of MapMaker Sibs.32 We report the nonparametric quantitative trait locus score, which requires no assumptions about the distribution of the phenotypic differences. The scores approximate a standard normal distribution. We also report the traditional lod score from the Haseman-Elston regression method.33A
Analyses of FCHL Sibships
Two-point lod scores are shown in Table 3⇓. The observed maximum score of 2.06 for D1S104 was tested by simulation analyses. In a simulation assuming no linkage, the average maximum score was 0.03, and only 3 scores of 1000 were greater than the observed score of 2.06. The additional information gained from the adjacent marker D1S1677 increased the score in a multipoint analysis to 2.52, but the location of the maximum was not between the 2 markers as it was in the study of Pajukanta et al.4 A multipoint simulation using 1000 replicates and assuming no linkage generated no lod score >1.0 across the region, indicating that the observed score of 2.52 is unlikely to be a false positive. Similar findings are present for the nonparametric analyses (see Table 4⇓).
Nonparametric scores for D1S104 and D11S4127 from each family were generated by use of the GENEHUNTER program. These scores were significantly negatively correlated (r=−0.30, P=0.01). To test a heterogeneity model in which a mutation at either locus could produce the disease, we weighted the families 0 if scores for both loci were negative and 1 if the score for either locus was positive. A new GENEHUNTER run with this weighting generated the scores shown in Table 4⇑. The significance of the 2-locus model was computed by use of a conservative χ2 test and provides significance for the multipoint chromosome 1 score and for D11S4127. In addition to this nonparametric analysis, the 2-locus parametric analysis using PAP produced a maximum lod score of 3.05 with disease locus 1 fixed at D1S104 (θ=0) and disease locus 2 fixed at D11S4127 (θ=0).
Analysis of Quantitative Lipid Traits in FHS Sibships
Before genetic analysis, adjustment for each quantitative lipid trait was performed by use of the general linear model procedure in SAS.33B Adjustment was performed separately by sex. For men, primary effects for LDL-C, total cholesterol, and triglycerides were age, age squared, body mass index, hematocrit, and research center. Additional effects for men were fruit/vegetable consumption (LDL-C and total cholesterol), drinking status (total cholesterol), fat consumption (triglycerides), and exercise per day (triglycerides). HDL-C in men showed associations with body mass index, hematocrit, smoking status, and drinking status. For women, primary effects for LDL-C and total cholesterol were age, age squared, body mass index, hematocrit, menopause status, and research center. For HDL-C and triglycerides, primary effects were body mass index, hematocrit, exercise per day, smoking status, drinking status, estrogen replacement, and research center. Additional effects in women were estrogen replacement (LDL-C), fruit/vegetable consumption (total cholesterol), fat intake (HDL-C), and menopause status (triglycerides). Residual scores from these regressions were used in the subsequent genetic analysis. Table 5⇓ shows the MapMaker Sibs results for the adjusted quantitative traits. No significant results were found.
The present study confirms evidence for a predisposing locus for FCHL on chromosome 1q21-23 in the NHLBI FHS. The maximum multipoint score reached significance for a replication (maximum lod score 2.52, nonparametric multipoint score 2.48). As suggested in Lander and Kruglyak,34 the significance threshold of a linkage originally found in a region up to 20 cM should be P=0.01. Because we additionally tested 2 models (dominant and recessive), the lod score threshold for significance should be increased by log10k, where k is the number of models tested (in this case, from 2.0 to ≈2.3). The smaller family size and the sib-pair approach likely contribute to the lower lod scores found in the present study versus the original Finnish study. In the present study, maxima occur either at D1S104 or are centromeric, rather than occurring between D1S104 and D1S1677, as reported by Pajukanta et al.4 There are a number of interesting candidate genes near this linkage region identified in the Pajukanta study, including apolipoprotein A-II (5 cM from D1S104). A gene encoding a secretory protein of adipose tissue (adipocyte-specific apM-1) has been subsequently mapped to 1q21.3-1q23.35
Weak positive scores were also obtained for the A-I/C-III/A-IV region by use of marker D11S4127 (nonparametric score 1.11, P=0.13). Parametric 2-locus linkage analysis using PAP gave a maximum lod score of 3.05 for D1S104 and D11S4127 by assuming a heterogeneity model in which a dominant mutation at either locus could produce FCHL. A nonparametric 2-locus analysis was performed by weighting families in GENEHUNTER as described by Cox et al.30 The weighted analyses gave significant improvements in scores over the single-locus analysis with use of a conservative χ2 test. Both methods suggest heterogeneous effects of the chromosome 1 region and the A-I/C-III/A-IV region in the FHS sample. A replication study in a family sample with larger extended pedigrees would likely clarify this finding.
Quantitative lipid traits (LDL-C, HDL-C, total cholesterol, and triglycerides) defined in the larger FHS sample produced consistently negative results with the candidate markers. Therefore, we conclude that these candidate loci are unlikely to account for major gene effects that have been suggested in segregation analyses of lipid traits. A possibility remains that the distribution of quantitative lipid traits might contain a more homogeneous subgroup at upper extremes for LDL-C, total cholesterol, and triglycerides and at the lower extreme for HDL-C. The 25th percentiles, as defined by the entire FHS sample, were chosen as cutoffs that would produce sufficient numbers of sibling pairs for analysis. Parametric and nonparametric analyses of these 25th percentiles produced positive results for triglyceride concentration in the chromosome 1 candidate region, although results were lower than those for the FCHL phenotype (data not shown). However, an investigation of the sib pairs analyzed revealed that there was substantial overlap between the 88 sib pairs defined by the upper 25th percentile triglyceride concentration and the FCHL sibships; 50 of 88 (57%) of the pairs were also FCHL pairs. Our result with high triglycerides in FHS is therefore likely to be a reflection of our FCHL results with this region.
The replication in the FHS sample of the chromosome 1q finding is important for this genetically heterogeneous disorder. An effect of this magnitude in our diverse multicenter study, consistent also with results in the Finnish population, suggests a common predisposing allele for FCHL on chromosome 1q. The heterogeneous A-I/C-III/A-IV effect was suggested in previous discrepant linkage studies reviewed in the introductory section, and this region remains controversial. The A-I/C-III/A-IV effect in our data was more pronounced when analyzed together with D1S104, suggesting the utility of 2-locus strategies in the detection of loci for this heterogenous trait.
Participating institutions and principal staff of the study are as follows: Forsyth County/University of North Carolina/Wake Forest University: Gerardo Heiss, Stephen Rich, Greg Evans, James Pankow, H.A. Tyroler, Jeannette T. Bensen, Catherine Paton, Delilah Posey, and Amy Haire; University of Minnesota Field Center: Donna K. Arnett, Aaron R. Folsom, Larry Atwood, James Peacock, and Greg Feitl; Boston University/Framingham Field Center: R. Curtis Ellison, Richard H. Myers, Yuqing Zhang, Andrew G. Bostom, Luc Djoussé, Jemma B. Wilk, and Greta Lee Splansky; University of Utah Field Center: Steven C. Hunt, Roger R. Williams (deceased), Paul N. Hopkins, Hilary Coon, and Jan Skuppin; Coordinating Center, Washington University, St. Louis: Michael A. Province, D.C. Rao, Ingrid B. Borecki, Yuling Hong, Mary Feitosa, Jeanne Cashman, and Avril Adelman; Central Biochemistry Laboratory, University of Minnesota: John H. Eckfeldt, Catherine Leiendecker-Foster, Michael Y. Tsai, and Greg Rynders; Central Molecular Laboratory, University of Utah: Mark F. Leppert, Jean-Marc Lalouel, Tena Varvil, Lisa Baird, Tami Elsner, Jim Yehle, Kristen Gruenthal, Chris Pappas, and Craig Barnitz; and NHLBI Project Office: Phyliss Sholinsky, Millicent Higgins (retired), Jacob Keller (retired), Sarah Knox, and Lorraine Silsbee.
This research was supported by National Institute of Mental Health grant MH-52055 and by NHLBI cooperative agreement grants U01 HL-56563, U01 HL-56564, U01 HL-56565, U01 HL-56566, U01 HL-56567, U01 HL-56568, and U01 HL-56569. This article is presented on behalf of the investigators of the NHLBI Family Heart Study. We thank the subjects participating in this study.
Grundy SM, Chait A, Brunzell JD. Familial combined hyperlipidemia workshop. Arteriosclerosis. 1987;7:203–207.
Genest JJ Jr, Martin-Munley SS, McNamara JR, Ordovas JM, Jenner J, Myers RH, Silberman SR, Wilson PW, Salem DN, Schaefer EJ. Familial lipoprotein disorders in patients with premature coronary artery disease. Circulation. 1992;85:2025–2033.
Goldstein JL, Schrott HG, Hazzard WR, Bierman EL, Motulsky AG. Hyperlipidemia in coronary heart disease, II: genetic analysis of lipid levels in 176 families and delineation of a new inherited disorder, combined hyperlipidemia. J Clin Invest. 1973;52:1544–1568.
Alayee H, Aouizerat BE, Cantor RM, Dallinga-Thie GM, Krauss RM, Lanning CD, Rotter JI, Lusis AJ, de Bruin TW. Families with familial combined hyperlipidemia and families enriched for coronary artery disease share genetic determinants for the atherogenic lipoprotein phenotype. Am J Hum Genet. 1998;63:577–585.
Surguchov AP, Page GP, Smith L, Patsch W, Boerwinkle E. Polymorphic markers in apolipoprotein C-III gene flanking regions and hypertriglyceridemia. Arterioscler Thromb Vasc Biol. 1996;16:941–947.
Wijsman EM, Brunzell JD, Jarvik GP, Austin MA, Motulsky AG, Deeb SS. Evidence against linkage of familial combined hyperlipidemia to the apolipoprotein A-I/C-III/A-IV gene complex. Arterioscler Thromb Vasc Biol. 1998;18:215–226.
Tahvanainen E, Pajukanta P, Porkka K, Nieminen S, Ikavalko L, Nuotio I, Taskinen MR, Peltonen L, Ehnholm C. Haplotypes of the apoA-I/C-III/A-IV gene cluster and familial combined hyperlipidemia. Arterioscler Thromb Vasc Biol. 1998;18:1810–1817.
Dallinga-Thie GM, van Linde-Sibenius Trip M, Rotter JI, Cantor RM, Bu X, Lusis AJ, de Bruin TW. Complex genetic contribution of the apo AI-CIII-AIV gene cluster to familial combined hyperlipidemia: identification of different susceptibility haplotypes. J Clin Invest. 1997;99:953–961.
Babirak SP, Brown BG, Brunzell JD. Familial combined hyperlipidemia and abnormal lipoprotein lipase. Arterioscler Thromb. 1992;12:1176–1183.
Hoffer MJ, Bredie SJ, Snieder H, Reymer PW, Demacker PN, Havekes LM, Boomsma DI, Stalenhoef AF, Frants RR, Kastelein JJ. Gender-related association between the −93T→G/D9N haplotype of the lipoprotein lipase gene and elevated lipid levels in familial combined hyperlipidemia. Atherosclerosis. 1998;138:91–99.
Yang WS, Nevin DN, Iwasaki L, Peng R, Brown BG, Brunzell JD, Deeb SS. Regulatory mutations in the human lipoprotein lipase gene in patients with familial combined hyperlipidemia and coronary artery disease. J Lipid Res. 1996;37:2627–2637.
Reymer PW, Groenemeyer BE, Gagne E, Miao L, Appelman EE, Seidel JC, Kromhout D, Bijvoet SM, van de Oever K, Bruin T, et al. A frequently occurring mutation in the lipoprotein lipase gene (Asn291Ser) contributes to the expression of familial combined hyperlipidemia. Hum Mol Genet.. 1995;4:1543–1549.
Gagne E, Genest J Jr, Zhang H, Clarke LA, Hayden MR. Analysis of DNA changes in the LPL gene in patients with familial combined hyperlipidemia. Arterioscler Thromb. 1994;14:1250–1257.
Pajukanta P, Porkka KV, Antikainen M, Taskinen MR, Perola M, Murtomaki-Repo S, Ehnholm S, Nuotio I, Suurinkeroinen L, Lahdenkari AT, et al. No evidence of linkage between familial combined hyperlipidemia and genes encoding lipolytic enzymes in Finnish families. Arterioscler Thromb Vasc Biol. 1997;17:841–850.
Higgins M, Province M, Heiss G, Eckfeldt J, Ellison RC, Folsom AR, Rao DC, Sprafka M, Williams R. NHLBI family heart study: objectives and design. Am J Epidemiol. 1996;143:1219–1228.
NHLBI Family Heart Study: Manuals of Procedure. St. Louis, Mo: Washington University; 1993.
Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem. 1972;18:499–502.
Warnick GR, Benderson J, Albers JJ. Dextran sulfate-Mg2+ precipitation procedure for quantitation of high-density-lipoprotein cholesterol. Clin Chem. 1982;28:1379–1388.
Lathrop GM, Lalouel J-M, Julier C, Ott J. Strategies for multilocus linkage analysis in humans. Proc Natl Acad Sci U S A. 1984;81:3443–3446.
Ott J. Computer-simulation methods in human linkage analysis. Proc Natl Acad Sci U S A. 1989;86:4175–4178.
Weeks DE, Ott J, Lathrop GM. SLINK: a general simulation program for linkage analysis. Am J Hum Genet. 1990;47(suppl):A204. Abstract.
Hasstedt SJ. Pedigree Analysis Package, Revision 4.0. Salt Lake City, Utah: Department of Human Genetics, University of Utah; 1994.
SAS User’s Guide, Version 6.03. Cary, NC: Statistical Analysis Institute; 1988.
Schaffler A, Orso E, Palitzsch KD, Buchler C, Drobnik W, Furst A, Scholmerich J, Schmitz G. The human apM-1, an adipocyte-specific gene linked to the family of TNF’s and to genes expressed in activated T cells, is mapped to chromosome 1q21.3-q23, a susceptibility locus identified for familial combined hyperlipidemia (FCH). Biochem Biophys Res Commun. 1999;260:416–425.