Abstract

Spirometric measures of pulmonary function have been shown to be highly heritable and evidence for major genes influencing forced expiratory volume in 1 s (FEV1) and forced vital capacity (FVC) have been reported. A genome scan of pulmonary traits in the Framingham Heart Study identified a region on chromosome 6qter with evidence for linkage to FEV1 and the FEV1/FVC ratio. For this study, additional markers were genotyped in the region to refine the location of linkage and test for association. Variance component linkage analysis was performed using GENEHUNTER, and family-based association tests were performed using FBAT. The chromosome 6 telomeric region provided significant evidence of linkage with the additional markers, resulting in a maximum multipoint LOD score of 5.0 for FEV1 at 184.5 cM. LOD scores for FVC and the FEV1/FVC ratio were also above 1.0 in this region. Evidence for association with FEV1 and FVC was observed with D6S281 at 190 cM. The strongest effect was seen with the 224 allele, which was associated with higher levels of FEV1 and FVC in allele carriers compared with those carrying other alleles. This study supports the presence of a gene influencing pulmonary function on the q-terminus of chromosome 6 in the region of 184 cM (D6S503) to 190 cM (D6S281).

INTRODUCTION

The influence of genetic factors on pulmonary function has been explored in several independent studies by evaluating quantitative measures of lung size and rate of airflow. The most commonly used measures for quantifying pulmonary function are forced expiratory volume in 1 s (FEV1) and forced vital capacity (FVC) obtained during spirometry. FEV1 is influenced by both lung size and airflow obstruction, while FVC, a measure of lung size, is less affected by airway obstruction. The ratio of FEV1 to FVC is a measure of airflow obstruction that is independent of lung size, and it is the measure used most often to identify obstructive pulmonary disease. Heritability estimates in excess of 40% have been reported for both FEV1 and FVC in several studies (15). Evidence for a major gene influencing FEV1 was reported in a segregation analysis of the Family Heart Study cohort (5), and evidence for a major gene influencing FVC was reported in a segregation analysis of the Humboldt Family Study (6). In the Framingham Heart Study, a segregation analysis of FEV1 did not show evidence for a major gene, but did indicate that multiple genes with varying effect sizes may influence FEV1 (7).

A genome scan was performed on the pulmonary traits FEV1, FVC and the ratio of FEV1/FVC in a cohort of 330 families in the Framingham Heart Study (8). The heritability estimates generated from the families in the genome scan were 35, 49 and 26% for FEV1, FVC and ratio of FEV1/FVC, respectively. Variance component linkage analysis for the genome scan was performed using the Sequential Oligogenic Linkage Analysis Routines (SOLAR) software package (9) and identified a region on the q-terminus of chromosome 6 with suggestive linkage to FEV1 (multipoint LOD score=2.4) and some evi-dence for linkage to the ratio of FEV1/FVC (multipoint LOD=1.4) (8). In the present study, additional microsatellite markers were genotyped on chromosome 6q in a subset of the families contributing to the genome scan. A total of 182 families from the genome scan were selected for follow-up based on the availability of DNA for additional genotyping in at least three members of the family (see Materials and Methods). The goals of the present study were to further refine the region of linkage by evaluating new markers in the region and to evaluate whether lung function was associated with specific marker alleles.

RESULTS

Table 1 reports the characteristics of the sample contributing to these analyses, comprising all 1231 participants with spirometry data. The characteristics of this follow-up sample are similar to the original sample included in the genome scan (8), although the mean pulmonary function in the cohort participants is slightly reduced in the current subset. The correlations among the traits after adjustment for age, age2, body mass index (BMI), height, smoking status and pack-years are shown in Table 2. All of the correlations were significant at P<0.005.

Linkage analysis

To compare results in the fine-mapping sample with those observed in the genome scan, multipoint variance components analysis in SOLAR was performed using only the four markers from the genome scan in the 182 families selected for the present analysis. Using the same family structure from the genome scan, analysis of the 182 families included 2810 individuals in the family file. A multipoint LOD score of 2.87 was observed for linkage to FEV1 at 184 cM, which was increased over the original result in the 330 families with an LOD score of 2.4. Subsequent linkage analysis was based on a family structure that was modified to eliminate individuals who were not informative in the linkage analysis and split nine families that were too large for the computation of identity-by-descent (IBD) using an exact multipoint algorithm (see Materials and Methods). The resulting sample consisted of 191 families and 1606 individuals. Of those removed, 98.3% did not have any genotyping data and 80.5% did not have phenotypic data. Analyzing the same four markers in SOLAR with the modified family data produced a LOD score of 3.0 for FEV1 at 186 cM.

Multipoint linkage analysis was then performed using the restructured pedigrees and a map containing both the genome scan markers and three new markers. Using SOLAR, the maximum multipoint LOD score for FEV1 was 3.64 at 182 cM. Thus, the addition of three new markers increased the multipoint LOD score by 0.64 LOD units. Using GENEHUNTER, the maximum multipoint LOD score for FEV1 was 5.0 at 184.5 cM. Table 3 reports the two-point LOD scores for FEV1 from SOLAR and GENEHUNTER and indicates the location where each marker was genotyped. The two-point LOD scores from the programs were very similar. The multipoint peak (shown in Figure 1) is positioned closest to marker D6S503, but the highest two-point LOD scores occurred at the markers flanking it. The D6S503 marker was one that was only genotyped in the first set of 53 families sent to MGS, and thus is missing in the majority of families. We explored whether the difference in the magnitude of the multipoint LOD scores using the two software packages was related to the missing data at D6S503 by reanalyzing multipoint linkage after excluding that marker entirely. Excluding D6S503, the multipoint LOD scores based on six markers were 3.46 for SOLAR and 4.35 for GENEHUNTER. The difference in the LOD scores was reduced by almost half of a LOD unit by removing D6S503, suggesting that the algorithms for computing multipoint IBD in the presence of missing marker data accounts for some of the difference in these LOD scores.

All of the pulmonary function measures were analyzed for linkage in this region using GENEHUNTER. Table 3 reports the two-point LOD scores and Figure 1 portrays the multipoint LOD scores generated using GENEHUNTER for FEV1, FVC and the FEV1/FVC ratio phenotype, both untransformed and normalized by ranking. The multipoint LOD scores for FVC and the ratio phenotypes are above 1.0 for the region that defines the 1-LOD score support interval (10) for FEV1 between 180 and 190 cM. The maximum multipoint LOD for FEV1 is located at 184.5 cM, but the peak multipoint LOD for the other traits within the 1-LOD support interval is at the position of D6S1027 at 187 cM. At this position, the multipoint LOD scores are 1.76 for FVC, 2.39 for the untransformed ratio and 1.38 for the ranked ratio. However, the ratio phenotypes both produced their highest multipoint LOD score at the position of D6S1273 at 173 cM, where the LOD scores were 2.88 and 1.41 for the ratio and ranked ratio, respectively.

Family-based association tests

The P-values from FBAT for multi-allelic tests are reported in Table 4, where the left half of the table reports results testing the null hypothesis of no linkage and no association (FBAT) and the right half of the table reports results testing the null hypothesis of no association in the presence of linkage (EV-FBAT). The reported P-values were not adjusted for multiple testing because of the non-independence of tests across correlated traits and microsatellite markers with varying degrees of linkage disequilibrium. The marker on chromosome 6 at 190 cM (D6S281) provided evidence of linkage and association with both FEV1 and FVC (P=0.014 and P=0.016, respectively) in multi-allelic tests, but not to the FEV1/FVC ratio (P=0.345). Two-point LOD scores for this marker were 0.95 for FEV1 and 0.22 for FVC. Evaluating this marker using the empirical variance option (EV-FBAT) to test the null hypothesis of no association in the presence of linkage resulted in P-values for FEV1 and FVC of 0.043 and 0.037, respectively, and the FEV1/FVC ratio had a borderline significant result with a P-value of 0.054.

Di-allelic analysis of D6S281 at 190 cM identified three alleles to be contributing to the association with pulmonary function. Table 5 reports the di-allelic FBAT and EV-FBAT results for the three alleles with P-values less than 0.05 for any of the three traits. The 222 allele and the 224 allele were significantly associated with FEV1 using both FBAT and EV-FBAT. The 222 and 224 alleles were informative in 12 and 29 families, respectively. The 224 allele was also significantly associated with FVC using both FBAT and EV-FBAT, and the association between the 222 allele and FVC was nearly significant. Additionally, allele 216 was significantly associated with the FEV1/FVC ratio using EV-FBAT. FEV1 and FVC were not significantly associated with allele 216.

The alleles implicated by FBAT analysis were further evaluated using regression in SOLAR, which accounts for the family relationships (see Materials and Methods). The frequency of the 222 allele was 1.2% and no participants were homozygous for the 222 allele; thus analysis was performed comparing presence or absence of the 222 allele. Participants carrying a copy of the 222 allele had a higher mean FEV1 by 116 ml and FVC by 133 ml compared with participants who did not carry a 222 allele, but these differences were not statistically significant. The frequency of the 224 allele was 3.3% with only one 224 homozygote, so analysis was performed comparing the presence or absence of the 224 allele. Participants carrying a copy of the 224 allele had significantly higher FEV1 by 130 ml (P=0.047) and FVC by 194 ml (P=0.007), compared with participants who did not carry any 224 alleles. The frequency of allele 216 was 2.8% and allele carriers had 91 ml higher mean FEV1 (P=0.17) and 121 ml higher mean FVC (P=0.10) compared to those without this allele.

DISCUSSION

The results of the fine-mapping and association analyses of FEV1, FVC and the ratio of FEV1/FVC in the Framingham Heart Study support evidence for a gene on the q-terminus of chromosome 6 that influences pulmonary function. Multipoint linkage analysis to FEV1 that provided suggestive evidence of linkage in the original genome scan was enhanced with the addition of three new markers, and reanalysis of the data with the revised pedigree structure has provided evidence of linkage that is statistically significant at the genome-wide significance level (11). The resulting LOD score of 5.0 at 184.5 cM for FEV1, colocalizing with LOD scores above 1 for FVC and the FEV1/FVC ratio, strongly suggests the presence of a gene that may influence airflow and lung size. The results of family-based association tests reveal evidence for association with both FEV1 and FVC at D6S281 (190 cM). The 224 allele of D6S281 was significantly associated with higher mean levels of FEV1 and FVC, and allele 222 also demonstrated an association with higher FEV1 and FVC.

The multipoint LOD score results reported are largely those generated using the software package GENEHUNTER. In the presence of a high proportion of missing data at a single marker (D6S503), the multipoint algorithm in GENEHUNTER was utilizing the adjacent marker information to predict missing genotypes on the basis of the observed haplotypes in the family, and that information was used for the estimate of multipoint IBD. In contrast, the SOLAR algorithm did not use haplotype information for the estimation of multipoint IBD. Recent reports comparing the multipoint IBD algorithms in the two packages suggest that the hidden Markov model approach implemented in GENEHUNTER increases power to detect linkage over SOLAR's multipoint IBD estimates based on applying regression formula to results from a single-point algorithm (12,13). For these reasons, we have chosen to primarily report the multipoint results from GENEHUNTER.

FBAT was used to test for association with pulmonary traits in a region implicated by linkage analysis. On chromosome 6, significant evidence for linkage to FEV1 was observed, but the evidence of linkage to FVC in this region does not meet criteria for suggestive linkage (11). The appropriate null hypothesis for FBAT differs based on the presence of linkage. For FEV1, the appropriate test is for association in the presence of linkage, but in the absence of suggestive linkage to FVC, it may be more appropriate to test the null hypothesis of no linkage and no association. Doing so, we found that FEV1 produces a significant result indicating association with D6S281, and results for FVC indicate statistically significant linkage and association with this marker. Assuming linkage and testing association for FVC, we found that the evidence for association persists at D6S281. However, the reported P-values were not adjusted for multiple comparisons because of the non-independence of tests, and thus these findings may include false-positive results.

These findings of linkage to FEV1 at 184.5 cM (D6S503) and association at 190 cM (D6S281) on chromosome 6 suggest that a putative gene influencing pulmonary function may be located in the region flanked by these two markers. The two-point linkage to FEV1 is strongest at 187.23 cM (D6S1027), but that marker is not associated with any of the traits in FBAT analyses. One possible explanation for this result is that the most common allele of D6S1027 is in linkage disequilibrium with a gene influencing the pulmonary measures. The marker at 190 cM (D6S281) shows evidence for linkage and association using FBAT, where the power to detect linkage is increased if there is an association (14), and alleles with low frequency were identified to be associated with FEV1 and FVC. The two-point LOD score of 0.95 for FEV1 at D6S281 may be reduced due to a lower marker heterozygosity of 0.67, compared with 0.77 seen for D6S1027, as this affects the informativeness of the marker and thus the ability to detect linkage (15). This reasoning might suggest that a gene is closer to D6S281 than the linkage peak. However, allelic association at D6S281 was observed with rare alleles that were among the largest number of repeats for the microsatellite. If the microsatellite expansion was a more recent evolutionary event, it is possible that these alleles exhibit linkage disequilibrium across a larger region than would be expected of older mutations, and the underlying functional polymorphism may be some distance from the observed association. Additional mapping with single nucleotide polymorphisms (SNPs) will be necessary to evaluate the extent of linkage disequilibrium in the region.

The evidence for a gene on the q-terminus of chromosome 6 influencing pulmonary function is supported by the robustness of the finding to all three measures, although the linkage and association results support somewhat different interpretations for the effect of a putative gene. FEV1 is influenced by overall lung volume, but is also sensitive to the presence of airflow obstruction independent of changes in lung volume, due either to dysynaptic lung growth or to accelerated loss of lung function after lung growth is complete. In contrast, FVC is a measure of lung size that is little influenced by airflow obstruction unless that obstruction is severe. The ratio of the two traits is an index of airflow obstruction that is independent of lung size. The linkage results suggest that a gene in the region primarily influences FEV1 and may influence airflow obstruction, while association results suggest a gene that may influence both airway caliber and lung size. Prior studies in the Framingham Heart Study population have shown a substantially higher heritability for cross-sectional pulmonary function measures than for longitudinal decline in pulmonary function (16), suggesting that the heritable component of cross-sectional pulmonary function, as analyzed in this study, may reflect developmental factors rather than acquired functional decline. However, as the subjects of this study are primarily middle-aged and older adults, genetic factors influencing the rate of decline in lung function during adulthood remains a possible explanation of the observed findings.

These findings of linkage and association arising from the evaluation of different pulmonary measures support a hypothesis that a putative gene on chromosome 6q27 may affect growth and development of the lungs. It remains unclear whether this gene had differential effects on the development of the lung parenchyma and the airways. Linkage analysis with FEV1 resulted in a LOD score of 5.0, strong evidence for a gene influencing airway size, but association was stronger with overall lung size, where a 194 mL higher FVC was associated with the presence of allele 224 of D6S281 (P=0.007). One candidate gene in the region of the linkage peak is SMOC2: secreted modular calcium-binding protein 2, which is of interest because it contains a Kazal-type serine protease inhibitor domain, and protease inhibitors have been previously linked to pulmonary disease. Polymorphisms of the alpha-1-antitrypsin protease inhibitor represent the only genetic defect known to cause obstructive pulmonary disease (17). Additional studies are planned that will explore association to these pulmonary measures with SNPs in this candidate gene and throughout the region.

MATERIALS AND METHODS

Sample

The Framingham Heart Study was established in 1948 as a prospective follow-up study of a sample of adult residents of Framingham, MA, USA. Approximately half of the town's residents in the age range of 28–62 years old were enrolled in the study for a sample of 5209 persons. The participants have undergone clinical examinations on a biennial basis. Children of the original Framingham Heart Study participants and their spouses were recruited into the Framingham Offspring Cohort in 1971. Recruitment yielded a sample of 5124 offspring and spouses. The offspring and spouses have undergone clinical examinations on a quadrennial basis since their second examination (1980–1984). The offspring cohort includes 3548 children of original cohort participants and 1576 spouses of the offspring. In 1987, collection of DNA samples for genetic studies was initiated in both the original and offspring cohorts. All study participants were examined under the screening and examination protocol approved by the Boston University Institutional Review Board and have signed approved consent forms. The Mammalian Genotyping Service (MGS) in Marshfield, Wisconsin typed a total of 401 markers spaced at approximately a 10 cM interval across the genome in 1702 individuals from the 330 largest extended Framingham Heart Study families.

Based on the results of a genome scan of the pulmonary traits (8), the linkage to FEV1 and the FEV1/FVC ratio found on the q-terminus of chromosome 6 was selected for follow-up. Not all of the 330 families that were included in the genome scan had adequate amounts of DNA remaining to accommodate additional genotyping. Families from the original 330 were selected for fine-mapping based on the availability of DNA for genotyping additional markers in (1) sibships with two or more siblings and at least one parent, or (2) sibships of three or more. A total of 182 families met these criteria, which included 1115 individuals with DNA whose reported family relationships had already been verified in the genome scan data using the sib_kin program of the ASPEX package. This sample include 66% of the individuals and 55% of the families genotyped for the genome-wide scan.

Three additional markers were selected for genotyping based on their positions relative to the four markers genotyped in the genome scan between 166 and 187 cM on chromosome 6. The microsatellite markers selected were chosen because they were well characterized on the Marshfield and Genethon genetic maps, had the highest heterozygosity at a given position, and performed well in trial PCR reactions. Several additional markers amplified poorly, and thus were not genotyped in the families. Genotyping was performed using an ABI Prism 3700 and the accompanying Genotyper software (Applied Biosystems, Foster City, CA, USA) in the Framingham Heart Study Genetics Laboratory at Boston University School of Medicine. The markers typed were located between 173 and 190 cM on chromosome 6. These markers and their heterozygosities are: D6S1273, 0.68; D6S264, 0.71; and D6S281, 0.67. The order and cM position of markers was specified according to MGS genetic maps (http://research.marshfieldclinic.org/genetics/). These markers were evaluated for Mendelian inconsistencies using the program GENTEST available as a precursor to INFER in the PEDSYS package (Southwest Foundation for Biomedical Research, San Antonio, TX, USA). When Mendelian inconsistencies were identified for a marker in a family, the genotyping information for that marker was removed for the entire nuclear family.

Phenotype

The pulmonary phenotypes used for this analysis are the same as those utilized in the genome scan (8). Spirometry data for the original cohort were collected between 1958 and 1961 using a 13.5 l Collins water-filled survey spirometer and multiplied by a standard correction factor of 1.1 to correct for body temperature and pressure, saturated (BTPS). In later years, spirometry in the cohort and offspring study was performed using a 6 l Collins water-filled survey spirometer connected to an Eagle II microprocessor that provided an automatic correction for BTPS and provided quality assurance information. The pulmonary phenotype was defined as the mean value of the observed trait at two clinical exams, adjusted for covariates. If an individual had only a single observation at the two clinical exams, then that single observation was used. For cohort participants, spirometry data from exam cycles 5 or 6 (conducted 1958–1961) and cycle 13 (conducted 1974–1975) were used to generate the mean value. In offspring participants, spirometry data from cycle 3 (conducted 1984–1987) and cycle 5 (conducted 1992–1995) were used to generate the mean value. These exam cycles were selected because they were thought to have the highest quality spirometry data. The mean value of the traits produced higher sibling correlations than did the measures at a single time point.

The traits were adjusted by gender within cohort or offspring samples, for the effects of age, age2, BMI (kg/m2), height, dummy variables indicating never, former or current smoking status, and pack-years for all former and current smokers. The relationship of these variables to FEV1 in this sample has been described previously (7). The covariates age, age2, height, weight (for the calculation of BMI), and pack-years of smoking were included in the same manner as the trait as a mean value or single observation. The categorical smoking status was determined from the earlier exam when data from two exams were available. The residuals from the regression models were standardized to a mean 0 and standard deviation of 1, and the distribution of each of the traits was evaluated. The distribution of the ratio of FEV1/FVC was somewhat kurtotic (kurtosis=2.37; SAS version 8), whereas the distributions for the FEV1 and FVC traits did not deviate from normality to such a degree that a transformation was needed. Since a variance components analysis can be sensitive to non-normality in the data (18), the distribution of the ratio trait was normalized using the ranking procedure in SAS to create normalized deviates. Both the residual of the FEV1/FVC ratio and the normalized deviate of the residual were evaluated in the linkage analysis. The correlations among the four adjusted phenotypes were calculated.

Linkage analysis

Variance components linkage analysis was initially performed in SOLAR for comparison with results from the genomewide scan. Since the genome scan, Framingham investigators (LDA, NLH-C) revised the pedigrees using the program PEDTRIM available as part of the PEDSYS software package. The restructuring of families was designed to eliminate family members who were not contributing information to the linkage analysis, so that smaller families could be analyzed using the exact multipoint algorithm available in the GENEHUNTER software package (1921). The variance component methods in both SOLAR and GENEHUNTER partition the total phenotypic variance into components due to genetic and environmental effects, but the algorithms for computing multipoint identity by descent estimates differ. Whereas GENEHUNTER is a multipoint algorithm, SOLAR is a single-point algorithm that uses a linear approximation to estimate multipoint identity by descent (13). The restructured pedigrees were analyzed for linkage in both software packages. The null hypothesis being tested is that the additive genetic variance due to a putative quantitative trait locus (QTL) equals zero. This null model does not include a QTL effect, but models the polygenic effects on the phenotype. The null hypothesis of no linkage is tested by comparing the likelihood of the null (polygenic) model with that of a model that estimates the additive genetic variance at the putative QTL.

Family-based association tests

All of the markers were evaluated using Family Based Association Tests (FBAT) (22,23). Family-based association tests the null hypothesis of no linkage and no association, or may test the null hypothesis of no association in the presence of linkage using an empirical variance estimator that adjusts for the correlation among sibling marker genotypes (24). One advantage of testing these composite null hypotheses in a family-based test is that they protect against type I error arising from population stratification. Rejecting the composite null hypothesis of no linkage and no association requires that both conditions be met, as the test has no power to detect the alternative hypothesis if either condition is not met (22). FBAT analyses extend the methodology of the transmission disequilibrium test (25) to evaluate nuclear families including both affected and unaffected offspring, and it can be applied to quantitative trait data as well as data on dichotomous outcomes. FBAT conditions on the observed traits and parental genotypes, and where parental data are missing, conditions on the offspring genotype configuration specify the distribution of a score statistic. The conditional distribution is used to calculate the mean and variance of each family's contribution to a general score statistic (23,26).

FBAT was used to evaluate all of the markers on chromosome 6 for each of the three phenotypes. Multi-allelic tests were performed using an additive genetic model to identify markers with evidence for both linkage and association with any of the phenotypes or for association in the presence of linkage. Markers significant at P<0.05 in a multi-allelic test were then evaluated with a di-allelic test under an additive genetic model. The di-allelic test compares each allele individually against all others collapsed into a single category to determine which specific allele(s) at the marker show association. The null hypothesis of no association in the presence of linkage was tested using the empirical variance option in FBAT. The empirical variance option provides an unbiased test for association in the presence of linkage by accounting for the correlation among sibling marker genotypes (24). This option is particularly relevant for evaluating association with FEV1 at the markers on chromosome 6qter.

The alleles significantly associated with the trait in FBAT were then further evaluated to determine the magnitude and direction of association with FEV1 and FVC. Since the alleles identified were relatively rare, groups defined by the presence or absence of the allele were compared. The allelic indicator variable was incorporated as a covariate into a SOLAR polygenic regression model that used the unadjusted spirometric measures as the trait and included the same covariates as were previously described. The polygenic model incorporates the family structure and covariates to estimate the heritability of the trait and the proportion of trait variance explained by the covariates. This analysis provided an estimate of the allele's effect on FEV1 and FVC measured in liters and tested the statistical significance of the allele's effect in a model that accounts for the non-independence of family members.

ACKNOWLEDGEMENTS

Supported by the Framingham Heart Study of the National Heart, Lung and Blood Institute (supported by NIH/NHLBI Contract N01-HC-25195) and the Research Enhancement Award Program from the Boston Veteran's Administration Medical Research Service.

*

To whom correspondence should be addressed at: Boston University School of Medicine, 715 Albany Street, B-601, Boston, MA 02118, USA. Tel: +1 6176385105; Fax: +1 6176388076; Email: jwilk@bu.edu

Figure 1. Multipoint LOD score results obtained from GENEHUNTER for pulmonary function measures on chromosome 6qter. Marker locations are indicated by the arrows and correspond to inflection points on the LOD score plots. The asterisks indicate markers genotyped in this follow-up study.

Figure 1. Multipoint LOD score results obtained from GENEHUNTER for pulmonary function measures on chromosome 6qter. Marker locations are indicated by the arrows and correspond to inflection points on the LOD score plots. The asterisks indicate markers genotyped in this follow-up study.

Table 1.

Descriptive characteristics of the pulmonary measures and covariates (mean±standard deviation; percentage where indicated)

 Cohort Offspring 
 Males(n=190) Females(n=238) Males(n=400) Females(n=403) 
Age 56.1±7.2 57.8±7.5 48.1±9.8 49.2±9.9 
Height (m) 1.71±0.06 1.57±0.06 1.76±0.07 1.62±0.06 
BMI (kg/m226.8±3.1 27.0±4.7 27.8±4.0 26.2±5.7 
FEV1 (l) 3.03±0.68 2.26±0.50 3.65±0.74 2.61±0.58 
FVC (l) 3.90±0.71 2.73±0.54 4.81±0.87 3.38±0.64 
Ratio FEV1/FVC 0.75±0.10 0.79±0.08 0.76±0.07 0.77±0.08 
Percentage current smokers 59.5 35.3 26.5 28.5 
Percentage former smokers 15.3 7.1 38.3 34.7 
Pack-years 36.5±20.6 17.4±15.0 26.0±24.5 17.7±17.2 
 Cohort Offspring 
 Males(n=190) Females(n=238) Males(n=400) Females(n=403) 
Age 56.1±7.2 57.8±7.5 48.1±9.8 49.2±9.9 
Height (m) 1.71±0.06 1.57±0.06 1.76±0.07 1.62±0.06 
BMI (kg/m226.8±3.1 27.0±4.7 27.8±4.0 26.2±5.7 
FEV1 (l) 3.03±0.68 2.26±0.50 3.65±0.74 2.61±0.58 
FVC (l) 3.90±0.71 2.73±0.54 4.81±0.87 3.38±0.64 
Ratio FEV1/FVC 0.75±0.10 0.79±0.08 0.76±0.07 0.77±0.08 
Percentage current smokers 59.5 35.3 26.5 28.5 
Percentage former smokers 15.3 7.1 38.3 34.7 
Pack-years 36.5±20.6 17.4±15.0 26.0±24.5 17.7±17.2 
Table 2.

Correlation coefficients between phenotypes after adjustment for age, age2, BMI, height, smoking status, and pack-years

 FEV1 FVC Ratio 
FVC 0.793   
Ratio 0.487 −0.067  
Normalized ratio 0.455 −0.100 0.977 
 FEV1 FVC Ratio 
FVC 0.793   
Ratio 0.487 −0.067  
Normalized ratio 0.455 −0.100 0.977 
Table 3.

Two-point LOD scores from SOLAR and GENEHUNTER for FEV1 and GENEHUNTER two-point LOD scores for other pulmonary measures

Marker Position (cM) Typeda SOLAR GENEHUNTER 
   FEV1 FEV1 FVC Ratio Ratiob 
D6S305 166.39 0.22 0.23 0.16 0.16 0.0 
D6S1277 173.31 0.74 0.74 0.73 0.74 0.43 
D6S1273 173.31 0.81 0.88 0.14 1.87 1.29 
D6S264 179.07 2.31 2.30 0.22 1.07 0.70 
D6S503 184.51 1.07 1.09 0.12 0.85 0.63 
D6S1027 187.23 2.67 2.64 0.92 1.83 1.16 
D6S281 190.14 1.01 0.95 0.22 0.29 0.24 
Marker Position (cM) Typeda SOLAR GENEHUNTER 
   FEV1 FEV1 FVC Ratio Ratiob 
D6S305 166.39 0.22 0.23 0.16 0.16 0.0 
D6S1277 173.31 0.74 0.74 0.73 0.74 0.43 
D6S1273 173.31 0.81 0.88 0.14 1.87 1.29 
D6S264 179.07 2.31 2.30 0.22 1.07 0.70 
D6S503 184.51 1.07 1.09 0.12 0.85 0.63 
D6S1027 187.23 2.67 2.64 0.92 1.83 1.16 
D6S281 190.14 1.01 0.95 0.22 0.29 0.24 

aM: marker genotyped at Mammalian Genotyping Services; F: marker genotyped at Framingham Heart Study Genetics Laboratory.

bNormalized residual of the ratio of FEV1/FVC.

Table 4.

Multi-allelic results for all markers testing null hypotheses of no linkage and no association (FBAT) and no association in the presence of linkage (empirical variance EV-FBAT)

Marker Distance (cM) FBAT P-values EV-FBAT P-values 
  FEV1 FVC Ratio FEV1 FVC Ratio 
D6S305 166.39 0.115 0.084 0.683 0.252 0.117 0.590 
D6S1277 173.31 0.485 0.895 0.251 0.653 0.898 0.297 
D6S1273 173.31 0.069 0.106 0.340 0.091 0.115 0.413 
D6S264 179.07 0.942 0.863 0.544 0.958 0.868 0.526 
D6S503 184.51 0.424 0.097 0.505 0.690 0.155 0.619 
D6S1027 187.23 0.710 0.986 0.602 0.683 0.976 0.663 
D6S281 190.14 0.014 0.016 0.345 0.043 0.037 0.054 
Marker Distance (cM) FBAT P-values EV-FBAT P-values 
  FEV1 FVC Ratio FEV1 FVC Ratio 
D6S305 166.39 0.115 0.084 0.683 0.252 0.117 0.590 
D6S1277 173.31 0.485 0.895 0.251 0.653 0.898 0.297 
D6S1273 173.31 0.069 0.106 0.340 0.091 0.115 0.413 
D6S264 179.07 0.942 0.863 0.544 0.958 0.868 0.526 
D6S503 184.51 0.424 0.097 0.505 0.690 0.155 0.619 
D6S1027 187.23 0.710 0.986 0.602 0.683 0.976 0.663 
D6S281 190.14 0.014 0.016 0.345 0.043 0.037 0.054 
Table 5.

Di-allelic results for marker D6S281 alleles with a P-value≤0.05 for at least one pulmonary measure

Allele FBAT P-values EV-FBAT P-values 
 FEV1 FVC Ratio FEV1 FVC Ratio 
216 0.962 0.306 0.059 0.957 0.212 0.009 
222 0.007 0.063 0.130 0.035 0.052 0.084 
224 0.018 0.017 0.487 0.025 0.021 0.456 
Allele FBAT P-values EV-FBAT P-values 
 FEV1 FVC Ratio FEV1 FVC Ratio 
216 0.962 0.306 0.059 0.957 0.212 0.009 
222 0.007 0.063 0.130 0.035 0.052 0.084 
224 0.018 0.017 0.487 0.025 0.021 0.456 

References

1
Lebowitz, M.D., Knudson, R.J. and Burrows, B. (
1984
) Familial aggregation of pulmonary function measurements.
Am. Rev. Respir. Dis.
 ,
129
,
8
–11.
2
Astemborski, J., Beaty, T. and Cohen, B. (
1990
) Variance components analysis of forced expiration in families.
Am. J. Med. Genet.
 ,
21
,
741
–753.
3
Coultas, D.B., Hanis, C.L., Howard, C.A., Skipper, B.J. and Samet, J.M. (
1991
) Heritability of ventilatory function in smoking and nonsmoking New Mexico Hispanics.
Am. Rev. Respir. Dis.
 ,
144
,
770
–775.
4
Schilling, R.S., Letai, A., Hui, S., Beck, G., Schoenberg, J. and Bouhuys, A. (
1997
) Lung function, respiratory disease, and smoking in families.
Am. J. Epidemiol.
 ,
106
,
274
–283.
5
Wilk, J.B., Djousse, L., Arnett, D.K., Rich, S.S., Province, M.A., Hunt, S.C., Crapo, R.O., Higgins, M. and Myers, R.H. (
2000
) Evidence for major genes influencing pulmonary function in the NHLBI Family Heart Study.
Genet. Epidemiol.
 ,
19
,
81
–94.
6
Chen, Y., Rennie, D.C., Lockinger, L.A. and Dosman, J.A. (
1997
) Major genetic effect on forced vital capacity: The Humboldt Family Study.
Genet. Epidemiol.
 ,
14
,
63
–76.
7
Givelber, R.J., Couropmitree, N.N., Gottlieb, D.J., Evans, J.C., Levy, D., Myers, R.H. and O'Conner, G.T. (
1998
) Segregation analysis of pulmonary function among families in the Framingham Study.
Am. J. Respir. Crit. Care Med.
 ,
157
,
1445
–1451.
8
Joost, O., Wilk, J.B., Cupples, L.A., Harmon, M., Shearman, A.M., Baldwin, C.T., O'Connor, G.T., Myers, R.H. and Gottlieb, D.J. (
2002
) Genetic loci influencing lung function: a genomewide scan in the Framingham Study.
Am. J. Respir. Crit. Care Med.
 ,
165
,
795
–799.
9
Almasy, L. and Blangero, J. (
1998
) Multipoint quantitative trait linkage analysis in general pedigrees.
Am. J. Hum. Genet.
 ,
62
,
1198
–1211.
10
Conneally, P.M., Edwards, J.H., Kidd, K.K., Lalouel, J.M., Morton, N.E., Ott, J. and White, R. (
1985
) Report of the committee on methods of linkage analysis and reporting.
Cytogenet. Cell Genet.
 ,
40
,
356
–359.
11
Lander, E. and Kruglyak, L. (
1995
) Genetic dissection of complex traits: guidelines for interpreting and reporting linkage results.
Nature Genet.
 ,
11
,
241
–247.
12
Ekstrom, C.T. (
2001
) Power of multipoint identity-by-descent methods to detect linkage using variance component models.
Genet. Epidemiol.
 ,
21
,
285
–298.
13
Sobel, E., Sengul, H. and Weeks, D.E. (
2001
) Multipoint estimation of identity-by-descent probabilities at arbitrary positions among marker loci in general pedigrees.
Hum. Hered.
 ,
52
,
121
–131.
14
Xu, J., Zheng, S.L., Hawkins, G.A., Faith, D.A., Kelly, B., Isaacs, S.D., Wiley, K.E., Chang, B., Ewing, C.M., Bujnovszky, P. et al. (
2001
) Linkage and association studies of prostate cancer susceptibility: Evidence for linkage at 8p22–23.
Am. J. Hum. Genet.
 ,
69
,
341
–343.
15
Heard-Costa, N.L., Demissie, S., DeStefano, A.L., Knowlton, B.A., Maher, N.E., Myers, R.H., Volcjak, J.S., Wilk, J.B. and Cupples, L.A. (
2001
) Influence of marker heterozygosity and genetic heterogeneity on fine mapping.
Genet. Epidemiol.
 ,
21
(Suppl. 1),
S467
–S472.
16
Gottlieb, D.J., Wilk, J.B., Harmon, M., Evans, J.C., Joost, O., Levy, D., O'Connor, G.T. and Myers, R.H. (
2001
) Heritability of longitudinal change in lung function: the Framingham Study.
Am. J. Respir. Crit. Care Med.
 ,
164
,
1655
–1659.
17
Tobin, M.J., Cook, P.J.L. and Hutchinson, D.C.S. (
1983
) Alpha-1-antitrypsin deficiency: the clinical and physiological features of pulmonary emphysema in subjects homozygous for Pi type Z: a survey by the British Thoracic Association.
Br. J. Dis. Chest
 ,
77
,
14
–27.
18
Allison, D.B., Neale, M.C., Zannolli, R., Schork, N.J., Amos, C.I. and Blangero, J. (
1999
) Testing the robustness of the likelihood-ratio test in a variance-component quantitative-trait loci-mapping procedure.
Am. J. Hum. Genet.
 ,
65
,
531
–544.
19
Kruglyak, L., Daly, M.J., Reeve-Daly, M.P. and Lander, E.S. (
1996
) Parametric and nonparametric linkage analysis: a unified multipoint approach.
Am. J. Hum. Genet.
 ,
58
,
1347
–1363.
20
Pratt, S.C., Daly, M.J. and Kruglyak, L. (
2000
) Exact multipoint quantitative-trait linkage analysis in pedigrees by variance components.
Am. J. Hum. Genet.
 ,
66
,
1153
–1157.
21
Markianos, K., Daly, M.L. and Kruglyak, L. (
2001
) Efficient multipoint linkage analysis through reduction of inheritance space.
Am. J. Hum. Genet.
 ,
68
,
963
–977.
22
Laird, N.M., Horvath, S. and Xu, X. (
2000
) Implementing a unified approach to family-based tests of association.
Genet. Epidemiol.
 ,
19
(Suppl. 1),
S36
–S42.
23
Horvath, S., Xu, X. and Laird, N.M. (
2001
) The family based association test method: strategies for studying general genotype-phenotype associations.
Eur. J. Hum. Genet.
 ,
9
,
301
–306.
24
Lake, S.L., Blacker, D. and Laird, N.M. (
2000
) Family-based tests of association in the presence of linkage.
Am. J. Hum. Genet.
 ,
67
,
1515
–1525.
25
Spielman, R.S., McGinnis, R.E. and Ewens, W.J. (
1993
) Transmission test for linkage disequilibrium: The insulin gene region and insulin-dependent diabetes mellitus (IDDM).
Am. J. Hum. Genet.
 ,
52
,
506
–516.
26
Rabinowitz, D. and Laird, N.M. (
2000
) A unified approach to adjusting association tests for population admixture with arbitrary pedigree structure and arbitrary missing marker information.
Hum. Hered.
 ,
50
,
211
–223.