Introduction

Asthma has been described as a 'noninfectious epidemic' affecting about 155 million individuals worldwide and rising. Quantitative variation in the levels of total immunoglobulin E (IgE) has been shown to be an important mediating factor for this as well as other complex atopic disorders such as eczema and hay fever.^1,2

One major problem encountered when studying nonMendelian, complex phenotypes such as asthma is that of definition.³ Studying intermediate quantitative phenotypes^4,5 such as the level of circulating specific or nonspecific IgE as a proxy for asthma has many potential clinical and analytical advantages. The most important is the correlation between the more genetically and environmentally homogenous phenotype and the underlying DNA sequences. Although there are cases where this is not particularly true,⁶ generally, this could result in significant gains in power, as fewer genes of major effect will be sought, less genetic and environmental heterogeneity will be observed, and less error will be generated in definition and measurement of the phenotype.⁷ Studying twin pairs offer additional, potential advantages.^8,9 These include: (i) removing any confounding effects of age; (ii) controlling for differences in pre- and postnatal circumstances of gestation and rearing;¹⁰ and (iii) lower risk of nonpaternity compared to that of siblings.¹¹ Moreover, since the increase in the prevalence of asthma and other atopic disorders have been linked to clean Westernised environments, diet and childhood infection, the twin design will allow us to test whether such environmental factors that should be common to twin pairs contribute significantly to variation in total IgE.^12,13

Numerous studies have implicated the 5q21–33 region in the inheritance of asthma, IgE levels (specific and nonspecific), eosinophil levels, bronchial hyper-reactivity, and other immune-mediated disorders. Several immune-related genes reside in this region including the cytokine gene cluster (IL-13, IL-4, IL-5, IL-3), IL-9, CD14, the ₂-adregnergic receptor (ADRB2), IL-12B, and hepatitis A virus receptor (HAVCR-1) (see Figure 3). The presence of these genes makes this region a major candidate region for the study of the immunogenetics of asthma and its quantitative proxy, total IgE. However, despite numerous previous reports of linkage and association of asthma, atopy and serum IgE levels to genes within the 5q21–33 region, definitive results are still not available. This was highlighted recently in two separate meta-analysis conducted by the International Consortium on Asthma Genetics (COAG).^14,15 Both studies explicitly looked at total serum IgE levels and reported negative linkages to polymorphisms within this region.

Figure 3.

Genetic (cM) and physical (Mb) map showing the genomic and marker organisation within the 5q21–33 region. Several immune-related genes reside in this region including the cytokine gene cluster (IL- 13, IL- 4, IL- 5, IL- 3, interferon regulatory factor- 1 (IRF- 1), cell division cycle 25 (CDC25C), granulocyte- macrophage colony-stimulating factor (GMCSF)), T-cell-specific transcription factor- 7 (TCF7), IL- 9, early growth factor response-1 (EGR1), CD14, fibroblast growth factor-1 acidic (FGF1), lymphocyte-specific glucocorticoid receptor (GRL), the ₂-adregnergic receptor (ADRB2), colony-stimulating factor-1 receptor (CSF1R), platelet-derived growth factor receptor (PDGFR), IL-12B, and hepatitis A virus receptor (HAVCR-1). Marker order, genetic and physical distances were obtained using software provided by the university of California, Santa Cruz, http://www.genome.ucsc.edu/.

Full figure and legend (55K)

To further address these conflicting results, we measured total IgE in a total of 568 monozygotic (MZ) and dizygotic (DZ) healthy female twin pairs and genotyped 11 microsatellite markers across a 26 cM region spanning the 5q21–33 region. Total IgE was shown to be highly heritable with a narrow sense heritability estimate of 65% (confidence intervals (CI): 58–71% (previous reports have ranged from 46 to 80%)). Although evidence for linkage was not observed to any of the 11 microsatellites, the results from the omnibus test of association showed positive association for D5S393, and D5S673, which were shown to account for 5 and 11% of the total variance in IgE levels, respectively.

Results

The age range of the twins under study was 18–80 years. The mean age of the MZ twins was 47.715.6 years and for the DZ twins it was 50.013.0 years.

The mean values and standard deviations for the MZ (83.2227.9 kU/l) and DZ (91.91202.2 kU/l) twin populations were comparable, although means were slightly higher in the DZ group. On testing for the differences in the means between the MZ and DZ groups, we showed that although the differences were significant this did not affect any of the results (data not shown).

The intraclass (twin) correlations for IgE suggest a strong additive genetic effect, as the ICC for the DZ twins (0.28) is about half the MZ twins (0.63) intraclass correlations. The results of the model fitting analysis are shown in Table 1. The AE model (ascribing variance because of additive genetic and unique environment) represented the best-fitting and most parsimonious model for IgE (C and D could be dropped without any significant change in ²). The heritability of IgE was estimated at 65% (CI=58–71%). The remainder of the variance was attributable to unique environmental factors (35%, CI=29–42%). Age showed a small and insignificant contribution to the variance of total IgE (0.3%, P>0.05).

Table 1 - Path analysis and genetic model fitting results for variation in total IgE levels.

Full table

Two-point and multipoint linkage analysis was carried out using the computer package Genehunter 2.0. No evidence for linkage was found using traditional two-point or multipoint Haseman and Elston, maximum likelihood nonparametric or variance components procedures to any of the 11 markers spanning the 5q21–33 region (results not shown).

Although there was no evidence of linkage of markers to IgE, results from the multiallelic omnibus association test showed positive association for D5S393 (P=0.038) and D5S673 (P=0.000002). When corrected for multiple testing, only D5S673 was statistically significant at the nominal threshold of P0.0045. Allelewise association was observed for markers D5S393, D5S643, D5S207, D5S210, D5S410, and D5S673. However, after correction for multiple testing none of the allele showed significant association at the nominal level of P0.0009. Although Bonfferoni corrections are overly conservative for correlated variables the testing of many markers and many alleles means that the probability of type I error is inflated, thus making adjustments to nominal significance levels necessary. No evidence for confounding because of population stratification (at P0.05) was found for any tests of association (not presented). The results of tests of associations are summarised in Table 2.

Table 2 - Association analysis for IgE to 11 microsatellite markers within the 5q21–33 region.

Full table

Power calculations were performed using both our own data combined with best-case scenarios for the unobserved quantitative trait locus (QTL). The results showed that we had 7% and 15% power (=0.05) to detect linkage between a marker lying adjacent to a quantitative trait locus accounting for 5 and 10% of the phenotypic variation, respectively. At the level of 0.0001, we have virtually no power to detect linkage between a marker and QTL. The power in our sample to detect association between a marker and putative gene was shown to be highly dependent on the unobserved QTL parameters. As a best-case scenario, we set the degree of disequilibrium between marker and QTL to one (D'=1) and equated the marker allele to the QTL allele frequency (M=Q=0.5). The power to detect association or linkage disequilibrium between a marker and QTL is over 80% for a sample size consisting of 200 sibling pairs (=0.0000001, =0.8). However, slight deviations in the level of disequilibrium (D'= 0.8; =0.37) or the ratio of marker to QTL allele frequency (M=0.4, Q=0.7: =0.2) resulted in substantial loss in power. The results of the power calculations are summarised in Table 3.

Table 3 - Power to detect linkage and association of a marker to a QTL accounting for 10% of the total variance in a phenotype with narrow sense heritability of 65% for a range of recombination fractions (), linkage disequilibria (D'), and QTL (Q) and marker (M) allele frequency ratios.

Full table

To assess the extent of linkage disequilibrium within the 5q21–33 region, we calculated D' for markers D5S471 to D5S410, using either a family-based technique or a method that utilises unrelated individuals to infer haplotypes and estimate their frequency. Pairwise D' estimates were highly correlated using either family data or unrelated individuals (Spearman rank-order correlation [r] = 0.72). The results are presented graphically in Figure 2. Although numerous marker pairs were associated (P<0.05), no marker pair showed a D' value greater than 0.35. We then tested the relation between (i) the genetic and physical distances, (ii) the physical distance and the extent of LD, and (iii) genetic distance and LD, between the 11 microsatelites (results not shown). While physical and genetic distances were highly correlated (r=0.94, P<0.00001), their relation with the amount of LD were highly unpredictable (r=-0.0138, (P=0.92) and 0.1276 (P=0.36) for physical and genetic distances respectively). For example, the second highest D' was measured between markers D5S393 and D5S673 (D'=0.24, P=0.03), which are separated by 21.5 Mb of DNA (16 cM), and whereas, for markers D5S436 and D5S402, which are separated by 1.3 Mb (0 cM), D' was calculated at only 0.06 (P=0.07). Moreover, significant LD, judged on the basis of the P-value from Fisher's exact test, did not necessarily mean high D', as the two were shown to be in moderate negative correlation (r=-0.49, P<0.0001).

Figure 2.

Visual representation of the pairwise linkage disequilibrium (D') across the 5q21–33 region. The data were plotted using the graphical overview of linkage disequilibrium (GOLD) program. Plots are symmetric across X- and Y-axis intersect and the axis represent (grid lines) markers 5' (D5S471) to 3' (D5S410); see Figure 3 for ordering of the markers. The colour keys represent regions of minimal LD (dark), to regions of maximal LD (light). D' was based on inferred haplotype frequencies using either unrelated individuals (expectation-maximisation algorithm) or the full sib-pair genotype data (simulated annealing algorithm). Alleles with frequency less than 5% were pooled for all markers. The Spearman rank-order correlation between D' values estimated by using either family data or unrelated individuals was r_s=0.72.

Full figure and legend (120K)

Discussion

The inheritance pattern of asthma and IgE has been under vigorous investigation for many years. Numerous candidate gene studies and genomewide scans have been performed on both outbred and isolated populations.^{16,17,18,19,20,21,22,23,24,25,26,27,28} Moreover, the long arm of chromosome 5q stretching from q21 to q33 represents one of the best-characterised regions for both linkage and association to asthma and atopy, as well as many other Mendelian and nonMendelian immune-related diseases.^{29,30,31,32,33,34} In this study, we used the classical twin design to (i) estimate the relative contributions of genes and environment to variation in total IgE levels, (ii) assess genetic linkage, and (iii) examine allelic association of 11 microsatellite markers spanning the 5q21–33 region to total IgE.

The result of phenotypic analysis of IgE show that majority of the variation in total IgE is because of genetic polymorphisms working in an additive manner. There is no indication at all, at least in female Caucasoids, that common environmental factors play a role in governing IgE variation. This result agrees with similar previous findings for both asthma and total IgE variation.³⁵

The association of D5S393 with total IgE replicates numerous previous studies and highlights the importance of this region in mediating IgE levels.³⁶ The closest genes to this marker are IL-9 and CD14. IL-9 is produced mainly by T cells and plays a pivotal role in mast cell development and enhancement of IgE production by B cells. CD14 is found mainly on monocytes and granulocytes and serves as a high-affinity receptor for bacterial lipopolysaccharide (LPS). Monocytes and granulocytes are both implicated in the pathogenesis of allergy and atopy (see Figure 1). One of the earliest studies on the inheritance pattern of IgE, carried out on 92 families ascertained from northern Holland, showed strong linkage between marker D5S436 and IgE using both traditional Lod score and affected sib-pair analysis (Lod=3.61). CSFR1 (Lod=1.13) and microsatellite marker D5S393 (Lod=1.93) also showed suggestive linkage signals.³⁷ Following the recruitment of another 108 families, the same group replicated the previous results on chromosome 5q using more powerful variance components methods (Lod=2.46, P=0.0004). They further showed that this region, which was flanked by markers D5S666, and D5S402, explained 37% of the variance in IgE.²⁵ In another study there was evidence of linkage for markers IL4R1, IRF1, IL-9, D5S393 and D5S399 to total IgE but not to specific IgE suggesting that polymorphisms within the 5q21–33 region could be contributing significantly to the variation of the non-MHC restricted pathway of IgE production.³⁸ The importance of CD14 polymorphisms and several other markers as modifiers of IgE and atopy has also been highlighted in several populations of British ancestry.^39,40

Figure 1.

Immunoglobulin-E (IgE) production can occur through two distinct pathways. The first is primarily antigen-specific (cognate) and requires major histocompatibility (MHC) class II-restricted antigen presentation to T helper 2 (Th2) cells with the concomitant release of interleukin-4 (IL-4) and/or IL-13, a process facilitated by IL-5. This causes immunoglobulin (Ig) heavy-chain class switching which in turn results in the specific IgE antibody (Ab) responses. Nonspecific IgE is produced as a result of a non-MHC-restricted (noncognate) IL-4 response involving mainly FcRI⁺ cells such as mast cells, basophils, and eosinophils. Production of IFN by Th1 cells can antagonise Th2-mediated responses. APC=antigen-presenting cell.

Full figure and legend (85K)

The strong association observed between IgE and the marker D5S673 in our study is interesting in light of a recent study that exploited the high degree of sequence homology between the mouse and human genomes, to scan the human 5q23–35 syntenic region in the mouse, for asthma susceptibility genes. The results highlighted the T-cell and airway phenotype regulator (TAPR) locus, which controls the development of airway hyperactivity and T-cell production of IL-4 and IL-13 in the mouse. TAPR cosegregates with sequence variants that encode the T-cell membrane proteins (TIMs) (41). The human orthologue of TIM is the receptor for hepatitis A virus (HAVCR-1), which maps to the 5q33.3 region adjacent to the D5S673 microsatellite marker. These results agree with the hypothesis that rising trends in asthma and atopy, observed in developing countries over the past 100 years, is because of reduced levels of childhood infection. Another strong candidate gene in the vicinity of D5S673 is IL-12B. IL-12 is a proinflamatory cytokine produced by macrophages, which acts on B cells, natural killer (NK) cells, and monocytes to induce proliferation and cytokine synthesis. Interestingly, the main cytokine produced as a result of IL-12 action is gamma-interferon (IFN) which downregulates Th2-mediated, antibody responses while upregulating Th1 cytotoxic responses (see Figure 1).

In light of such strong evidence supporting a role for the 5q21–33 region in controlling variation in IgE and susceptibility to asthma or atopy, there have also been numerous studies that have failed to highlight 5q21–33 as a major mediator of atopy per se.⁴² This was highlighted recently in a meta-analysis carried out on a combined data set from 11 different studies worldwide (2400–2600 full sibling pairs) where tests of linkage of total IgE to markers within this region were negative.¹⁵ This agreed with a previous meta-analysis which found evidence of linkage to asthma within the cytokine gene cluster around microsatellite marker D5S1505 (LOD=2.61, P=0.0025), but failed to do so for atopy variables including total IgE levels.¹⁴ Furthermore, over 50 independent analyses conducted on these data sets as part of the Genetic Analysis Workshop 12 (published as a special supplement of Genetic Epidemiolology 2001) also failed to find consistent evidence for linkage across these studies. The differing analytical tools used, reports of empirical rather than asymptotic P-values, use of different genetic maps, alternative definitions of disease, potential genotyping errors in earlier studies, and the presence of significant genetic heterogeneity within different populations are just some of the reasons as to why such discrepancies exist in the literature for asthma as well as many other common complex disease.

Potential caveats in the interpretation of our experimental data are (i) the degree of haematological chimerism which might confound the twin model, (ii) generisability of the results to males and/or nonCaucasian populations, (iii) the cross-sectional nature of the cohort that precludes real insight into disease a etiology at different age groups, and (iv) the potential utility of microsatellite markers in tests of association. Significant haematological chimerism in DZ twins is relatively rare. However, examinations of monochorionic twin placentas reveal the presence of vascular anastomoses between the two foetal circulations in around 10% of cases; such anastomoses rarely occur in dichorionic placentas.

Monochorionic twin placentation occurs in around two-thirds of MZ twins. Thus, we would expect significant haematological chimerism to occur in about 7% of the MZ group.⁴³ There is no way of knowing how relevant this is to the peripheral populations in adults but we expect any effect to be reflected in increased phenotypic similarity within an MZ pair. This would have the effect of slightly raising the heritability estimate. There is evidence that the expression of asthma and associated intermediate phenotypes such as total serum IgE levels are gender-dependent^44,45,46 in most Caucasoid populations and there is also evidence pointing towards differing pathophysiological mechanisms leading to atopy operating in different age groups.³ However, until suitable cohorts are collected and well-designed longitudinal studies are carried out, definitive answers remain at large. Finally, the utility of microsatellite markers in tests of association remains debatable. The high mutation rate of microsatellites could mean that a single microsatellite allele may represent several distinct haplotypes, confounding the detection of association between a microsatellite allele and a QTL. Unfortunately, practically and theoretically this remains a virtually untouched territory and needs more attention if the vast amount of data collected so far by the genetic community is to be taken full advantage of.

Sample size is a thorny issue affecting the sensitivity of any analysis. Our power calculations show that even assuming a recombination fraction of zero between the marker and QTL (=0), approximately 5000 sib-pairs are required to detect linkage to a gene accounting for 10% of phenotypic variation and an overall phenotypic heritability of 65%. For a locus accounting for 5% of the variance, the number of sibling pairs roughly doubles to 10 000. Our sample size has very low power to detect linkage to loci of small effect (5–10%) within the 5q21–33 regions. However, previous positive linkage reports utilising the same markers or markers in linkage to the markers in our study have either had similar or even smaller sample sizes than ours and have often utilised less powerful analytical techniques. These discrepancies could be because of the presence of genetic heterogeneity in asthma and/or IgE in different populations. However, the limitations of linkage analysis can be overcome by the association approach so long as appropriate genetic markers are examined. Power for linkage disequilibrium analysis is governed by not only the sample size as the extent of allelic association and allele frequency of marker and trait loci (which are determined by population-specific forces such as demography and selection) but can also have substantial effects on power.^47,48,49 Our power calculations and association results highlight and agree with previous claims that association studies are more powerful than linkage-only studies in detecting loci of small effect, provided that marker allele frequencies are matched to the QTL allele frequencies and that they are in strong allelic association.⁵⁰

Finally, the erratic distribution of LD throughout the 5q21–33 region and its weak association with both physical as well as genetic distances, follow from numerous previous studies that have shown that there is limited ability to predict the distance between two polymorphisms based solely on the amount of LD that exits between them.^51,52 This perhaps is not surprising when we take all the demographic (migration, selective sweeps, population bottlenecks) and genetic factors (recombination rates, selection) that can affect local patterns of LD into consideration.^53,54,55 However, as more recent data has shown, fine mapping of QTLs should be facilitated by LD mapping once local hot and cold regions of recombination and gene conversion are characterised and accounted for when choosing the density of markers required to carry out association mapping of complex traits.⁵⁶

The 5q21–33 region is rich with genes with pivotal roles in the functioning of the human immune system. We have shown, contrary to recent findings, that this region remains a major candidate for IgE level variation in Caucasoids and remains a target for further work in fine mapping and positional cloning of the causative polymorphisms affecting asthma and atopy. Our results show that at least two loci within this region are responsible for variation in IgE and possible susceptibility to atopic disorders. Although difficulties exist in the interpretation of mouse experiments, because of biological redundancy and in the unnatural behaviour of systems in controlled environments, our results also conform to recent findings in the murine model of human asthma.

Materials and methods

A total of 266 MZ (or identical) and 302 DZ (or fraternal) female Caucasoid twin pairs aged 18–80 years from the St Thomas' UK adult twin registry participated in the study. The twin pairs were ascertained from the general population through a national media campaign in the United Kingdom.⁵⁷ Participating twins were unaware of the specific diseases of interest or hypothesis being tested and informed consent was obtained from all subjects. Zygosity was determined by standardised questionnaire⁵⁸ and confirmed by a combination of DNA analyses, which included full genome scans, the ABI FESTR kit (Perkin Elmer-Applied Biosystems, UK) and the Forensic Science Services QUAD system (Birmingham, UK).

IgE measurements

Levels of total IgE were measured in duplicates by radioallergosorbent tests (RAST) using the Pharmacia CAP system. Further tests were performed to resolve any discrepancies between replicate assays. Log transformation of total IgE level was carried out prior to analysis to obtain an approximately normal distribution.

Molecular methods

DNA was extracted from whole blood by a phenol/chloroform method. Microsatellite-marker-based genotyping was undertaken using standard ABI Prism™ (PE Biosystems) fluorescence-based genotyping methodologies.^59,60 Specifically for this study, marker loci were amplified in 5 l single-plex PCRs in 384-well microtitre plates. Amplification products were pooled, precipitated, combined with loading buffer, formamide and an internal size standard (GS400HD, PE Biosystems), and then size-separated and detected using ABI PrismTM 377 automated sequencers (PE Biosystems). Locus-specific alleles were sized, identified, and coded as previously described,⁶¹ then re-coded to correlate exactly with the alleles published in GDB (http://www.gdb.org/). Alleles not found in GDB were given new codes, using the 'Allele Gold-Standardisation' function of Phenobase™, a proprietary database of clinical and genetic information (Gemini Genomics, UK). Allele sizes, initial codes, and final designations were processed and maintained within Phenobase™.

Statistical methods

Twin methodology makes use of the fact that MZ twins share identical genotypes, whereas DZ twins share on average only 50% of their segregating genes. It is assumed that both types of twins on average share their common family environment to the same extent so any greater similarity between MZ compared with DZ twins reflects genetic influences. A higher MZ than DZ intraclass correlation (r) provides a first impression of the magnitude of genetic influence, which is based on the classic Falconer formula to estimate broad-sense heritability: h²=2(r_MZ–r_DZ). Under an additive genetic model, r_DZ should be half of r_MZ. Nonadditive genetic components and common environmental factors would increase and decrease this ratio, respectively.⁶²

Path analysis and genetic model fitting of IgE

Genetic model fitting (path analysis) can be used for the decomposition of the observed phenotypic variance into its genetic and environmental components. Briefly, the total genetic effect on a phenotype can be decomposed into that because of the additive effects of alleles at multiple loci (A), that because of the dominance effects (D) and that because of epistatic interaction (I) between loci. Similarly, the environmental effects can be decomposed into that because of common environmental influences (C) shared by twins and that because of environmental influences unique (E) to each individual in a twin pair. The latter also contains measurement error. Dividing each of these components by the total variance yields the different standardised components of variance; for example, the heritability (h²) that can be defined as the ratio of additive genetic variance to total phenotypic variance.⁶³ Age was incorporated into the model as a covariate and its influence quantified on the phenotype. See the following references and references therein for detailed description of path analysis specific to the analysis of twin data.^63,64

Variance components genetic model

We used the program QTDT⁶⁵ to calculate maximum likelihood estimates of the means _b and _w, variance components _a², _pg², _e² and test models of linkage (_a²=0), association (_a=0), population stratification (_b-_w=0) and linkage in the presence of association (_a²=0|_a). See the following references and references therein for detailed description of the methodology involved.^66,67 As MZ twin pairs are genetically identical, they provide biased information when included in any test of linkage. However, MZ twin pairs can be included both in variance component tests of linkage to provide an accurate estimate of the residual polygenic component, _pg², and tests of association provided only one person from each twin pair is used. Single-point and multipoint identical by descent (IBD) probabilities were calculated using the program Genehunter 2.0.⁶⁸ Genehunter 2.0 was also used to carry out two-point and multipoint linkage analysis by standard Haseman & Elston regression method, the maximum likelihood estimation maximisation (EM) method and the nonparametric method. Permutations, implemented within QTDT, were carried out to test the sensitivity of tests to phenotypic distributions as not all individuals with IgE measurement had complete genotypic data. A total of 10 000 permutations were performed for each test.

Background linkage disequilibrium within the 5q21–33 region

Several measures of LD exist.⁶⁹ We calculated D'⁷⁰ between markers D5S471 and D5S410, a region spanning 26 cM corresponding to 36 Mb of DNA (Figure 2, see Figure 3 for marker ordering). Haplotype frequencies were inferred using genotypic information either on one person from each twin pair via the expectation-maximisation algorithm as implemented in utility programs 'Faster Permutation & Model-free analysis' (PFEH+), 'Faster Estimating Haplotypes' (FEH+), and '2-locus Linkage Disequilibrium' calculator (2LD)^71,72 or on related sibling pairs via the simulated-annealing algorithm as implemented in Simwalk 2.8.⁷³ Pairwise disequilibrium values were plotted using the Graphical Overview of Linkage Disequilibrium (GOLD) program.⁷⁴

Simulations

To estimate the power of our sample size to detect linkage and/or association, we ran a series of power calculations using a combination of both our phenotypic and genotypic information and a best-case QTL scenario. The following information was used: 250 fully informative sib-pairs, sib–sib correlation of 0.33, QTL additive variance of 0.1, residual shared variance of 0.55 (h²_phenotypic–h_QTL²), residual nonshared variance (1–r_mz=e²) of 0.35, recombination fraction () between QTL and marker of 0, equal QTL and marker allele frequency (0.5), perfect disequilibrium between marker and QTL (D'=1), for 80% power () and type I error rates () at 0.0001 and 0.0000001 to detect linkage and association, respectively. For the software used to carry out the simulations see (http://statgen.iop.kcl.ac.uk/gpc/).

References

1. Holgate ST, Corne J, Jardieu P, Fick RB, Heusser CH. 30 years with IgE. Munksgaard: Copenhagen, 1998.
2. Kon OM, Sihra BS, Compton CH, Leonard TB, Kay AB, Barnes NC. Randomised, dose-ranging, placebo-controlled study of chimeric antibody to CD4 (keliximab) in chronic severe asthma. Lancet 1998; 352: 1109–1113. | Article | PubMed | ISI | ChemPort |
3. Martinez FD. Complexities of the genetics of asthma. Am J Respir Crit Care Med 1997; 156 (4) (Part 2): S117–S122.
4. Nadeau JH, Frankel WN. The roads from phenotypic variation to gene discovery: mutagenesis versus QTLs. Nat Genet 2000; 25: 381–384. | Article | PubMed | ISI | ChemPort |
5. Stoll M, Cowley Jr AW, Tonellato PJ et al. A genomic-systems biology map for cardiovascular function. Science 2001; 294: 1723–1726. | Article | PubMed | ISI | ChemPort |
6. Silverman EK, Mosley JD, Rao DC et al. Linkage analysis of alpha 1-antitrypsin deficiency: lessons for complex diseases. Hum Heredity 2001; 52: 223–232. | Article |
7. Schork NJ. Genetics of complex disease: approaches, problems, and solutions [Review]. Am J Respir Crit Care Med 1997; 156(4 Part 2): S103–S109. | PubMed | ISI | ChemPort |
8. Martin N, Boomsma D, Machin G. A twin-pronged attack on complex traits [Review]. Nat Gene 1997; 17: 387–392.
9. Ahmadi KR, Hall MA, Norman PJ et al. Genetic determinism in the relationship between human CD4+ and CD8+ T lymphocyte populations? Genes Immun 2001; 2: 381–387. | Article |
10. Jones CA, Kilburn SA, Warner JA, Warner JO. Intrauterine environment and fetal allergic sensitization. Clin Exp Allergy 1998; 28: 655–659. | Article | PubMed | ISI | ChemPort |
11. Abecasis GR, Cherny SS, Cardon LR. The impact of genotyping error on family-based analysis of quantitative traits. Eur J Hum Gene 2001; 9: 130–134.
12. Matricardi PM. Infections preventing atopy: facts and new questions. Allergy 1997; 52: 879–882.
13. Strachan DP. Lifestyle and atopy. Lancet 1999; 353: 1457–1458. | Article |
14. Lonjou C, Barnes KC, Chen H et al. A first trial of retrospective collaboration for positional cloning in complex inheritance: assay of the cytokine region in chromosome 5 by the consortium on asthma genetics (COAG). Proc Natl Acad Sci USA 2000; 97: 10942–10947. | Article | PubMed | ChemPort |
15. Palmer LJ, Barnes KC, Burton PR et al. Meta-analysis for linkage to asthma and atopy in the chromosome 5q31–33 candidate region. Hum Mol Genet 2001; 10: 891–899. | Article |
16. Moffatt MF, Hill MR, Schou C et al. Genetic linkage of T-cell receptor / complex to specific IgE responses. Lancet 1994; 343: 1597–1600. | Article | PubMed | ISI | ChemPort |
17. Daniels SE, Bhattacharrya S, James A et al. A genome-wide search for quantitative trait loci underlying asthma. Nature 1996; 383: 247–250. | Article | PubMed | ISI | ChemPort |
18. Barnes KC, Neely JD, Duffy DL et al. Linkage of asthma and total serum IgE concentration to markers on chromosome 12q: evidence from Afro-Caribbean and Caucasian populations. Genomics 1996; 37: 41–50. | Article | PubMed | ISI | ChemPort |
19. Borish L, Aarons A, Rumbyrt J, Cvietusa P, Negri J, Wenzel S. Interleukin-10 regulation in normal subjects and patients with asthma. J Allergy Clin Immunol 1996; 97: 1288–1296. | Article | PubMed | ISI | ChemPort |
20. Sandford A, Weir T, Pare P. The genetics of asthma. Am J Respir Crit Care Med 1996; 153: 1749–1765. | PubMed | ISI | ChemPort |
21. CSGA. A genome-wide search for asthma susceptibility loci in ethnically diverse populations. The Collaborative Study on the Genetics of Asthma (CSGA). Nat Genet 1997; 15: 389–392. | Article | PubMed | ISI |
22. Palmer LJ, Pare PD, Faux JA et al. Fc epsilon R1-beta polymorphism and total serum IgE levels in endemically parasitized Australian aborigines. Am J Hum Genet 1997; 61: 182–188. | PubMed | ISI | ChemPort |
23. Thomas NS, Wilkinson J, Holgate ST. The candidate region approach to the genetics of asthma and allergy. Am J Respir Crit Care Med 1997; 156 (4 Part 2): S144–S151. | PubMed |
24. Wjst M, Fischer G, Immervoll T et al. A genome-wide search for linkage to asthma. German Asthma Genetics Group Genomics 1999; 58: 1–8. | Article | PubMed | ISI | ChemPort |
25. Xu J, Postma DS, Howard TD et al. Major genes regulating total serum immunoglobulin E levels in families with asthma. Am J Hum Genet 2000; 67: 1163–1173. | Article | PubMed | ISI | ChemPort |
26. Ober C, Leavitt SA, Tsalenko A et al. Variation in the interleukin 4-receptor alpha gene confers susceptibility to asthma and atopy in ethnically diverse populations. Am J Hum Genet 2000; 66: 517–526. | Article | PubMed | ISI | ChemPort |
27. Laitinen T, Daly MJ, Rioux JD et al. A susceptibility locus for asthma-related traits on chromosome 7 revealed by genome-wide scans in a founder population. Nat Genet 2001; 28: 87–91. | Article | PubMed | ISI | ChemPort |
28. Cookson WOC, Ubhi B, Lawrence R et al. Genetic linkage of childhood atopic dermatitis to psoriasis susceptibility loci. Nat Genet 2001; 27: 372–373. | Article | PubMed | ISI | ChemPort |
29. Hastbacka J, de la Chapelle A, Kaitila I, Sistonen P, Weaver A, Lander E. Linkage disequilibrium mapping in isolated founder populations: diastrophic dysplasia in Finland (published erratum appears in Nat Genet 1992; 2: 343). Nat Genet 1992; 2: 204–211. | Article | PubMed | ISI | ChemPort |
30. Jabs EW, Li X, Lovett M et al. Genetic and physical mapping of the Treacher Collins Syndrome locus with respect to loci in the chromosome 5q31 region. Genomics 1993; 18: 7–13. | Article | PubMed | ISI | ChemPort |
31. Rihet P, Traore Y, Abel L, Aucan C, Traore-Leroux T, Fumoux F. Malaria in humans: Plasmodium falciparum blood infection levels are linked to chromosome 5q31–q33. Am J Hum Genet 1998; 63: 498–505. | Article | PubMed | ISI | ChemPort |
32. Rioux JD, Stone VA, Daly MJ et al Familial eosinophelia maps to the cytokine gene cluster on human chromosomal region 5q31–33. Am J Hum Genet 1998; 63: 1086–1094. | Article | PubMed | ISI | ChemPort |
33. Chavanas S, Garner C, Bodemer C et al. Localization of the Netherton syndrome gene to chromosome 5q32, by linkage analysis and homozygocity mapping. Am J Hum Genet 2000; 66: 914–921. | Article | PubMed | ISI | ChemPort |
34. Rioux JD, Daly MJ, Silverberg MS et al. Genetic variation in the 5q31 cytokine gene cluster confers susceptibility to Crohn disease. Nat Genet 2001; 29: 223–228. | Article | PubMed | ISI | ChemPort |
35. Skadhauge LR, Christensen K, Kyvik KO, Sigsgaard T. Genetic and environmental influence on asthma: a population-based study of 11,688 Danish twin pairs. Eur Respir J 1999; 13: 8–14. | Article | PubMed | ISI | ChemPort |
36. Bosken CH, Ko YC, Shete S et al. Adaptations of linkage and association methods for the study of asthma, a complex trait. Genet Epidemiol 2001; 21 (Suppl 1): S89–S96.
37. Meyers DA, Postma DS, Panhuysen CI et al. Evidence for a locus regulating total serum IgE levels mapping to chromosome 5. Genomics 1994; 23: 464–470. | Article | PubMed | ISI | ChemPort |
38. Marsh DG, Neely JD, Breazeale DR et al. Linkage analysis of IL4 and other chromosome 5q31.1 markers and total serum immunoglobulin E concentrations. Science 1994; 264: 1152–1156. | Article | PubMed | ISI | ChemPort |
39. Gao PS, Mao XQ, Baldini M et al. Serum total IgE levels and CD14 on chromosome 5q31 [letter]. Clin Genet 1999; 56: 164–165. | Article | PubMed | ISI | ChemPort |
40. Walley AJ, Wiltshire S, Ellis CM, Cookson WO. Linkage and allelic association of chromosome 5 cytokine cluster genetic markers with atopy and asthma associated traits. Genomics 2001; 72:15–20. | Article | PubMed | ISI | ChemPort |
41. McIntire JJ, Umetsu SE, Akbari O et al. Identification of Tapr (an airway hyperreactivity regulatory locus) and the linked Tim gene family. Nat Immunol 2001; 2: 1109–1116. | Article | PubMed | ISI | ChemPort |
42. Mathias RA, Freidhoff LR, Blumenthal M et al. Genomewide linkage analyses of total serum IgE using variance components analysis in asthmatic families. Genet Epidemiol 2001; 20: 340–355. | Article | PubMed | ISI | ChemPort |
43. Philips DIW. Fetal programming or genes? In: Spector TD, Snieder H, MacGregor AJ (eds). Advances in Twin and Sib-pair Analysis. Oxford University Press: Oxford, 2000: pp. 79–89.
44. Shah A. Asthma and sex. Indian J Chest Dis Allied Sci 2001; 43: 135–137.
45. Fagan JK, Scheff PA, Hryhorczuk D, Ramakrishnan V, Ross M, Perskey V. Prevalence of asthma and other allergic diseases in an adolescent population:association with gender and race. Ann Allergy Asthma Immunol 2001; 86: 177–184.
46. Bjornson CL, Mitchell I. Gender differences in asthma in childhood and adolescence. J Gender Specific Med 2000; 3: 57–61.
47. Muller-Myhsok B, Abel L. Genetic analysis of complex diseases. Science 1998; 275: 1328–1329.
48. Kruglyak L. Prospects for whole-genome linkage disequilibrium mapping of common disease genes. Nat Genet 1999; 22: 139–144. | Article | PubMed | ISI | ChemPort |
49. Cardon LR, Abecasis GR. Some properties of a variance components model for fine-mapping quantitative trait loci. Behav Genet 2000; 30: 235–243. | Article | PubMed |
50. Ahmadi KR, Hall MA, Norman PJ, Vaughan RW, Spector TD, Lanchbury JS. A simple genotype-quantitative phenotype relationship in humans: a case study of T cell receptor V6.7. Am J Hum Genet 69 (4), 1312, 2001.
51. Collins A, Lonjou C, Morton NE. Genetic epidemiology of single-nucleotide polymorphisms. Proc Nat Acad Sci USA 1999; 96: 15 173–15 177.
52. Reich DE, Cargill M, Bolk S et al. Linkage disequilibrium in the human genome. Nature 2001; 411: 199–204. | Article | PubMed | ISI | ChemPort |
53. Laan M, Pääbo S. Demographic history and linkage disequilibrium in human populations. Nat Genet 1997; 17:435–438. | Article | PubMed | ISI | ChemPort |
54. Thompson EA, Neel JV. Allelic disequilibrium and allele frequency distribution as a function of social and demographic history. Am J Hum Genet 1997; 60: 197–204. | PubMed | ISI | ChemPort |
55. Harpending H, Rogers A. Genetic perspectives on human origins and differentiation. Annu Rev Genom Hum Genet 2000; 1: 361–385. | Article | ISI | ChemPort |
56. Daly MJ, Rioux JD, Schaffner SF, Hudson TJ, Lander ES. High resolution haplotype structure in the human genome. Nat Genet 2001; 29: 229–232. | Article | PubMed | ISI | ChemPort |
57. Andrew T, Hart DJ, Snieder H, de Lange M, Spector TD, MacGregor AJ. Are twins and singletons comparable? A study of disease-related and life-style characteristics in adult woman. Twin Res 2001; 4: 464–477. | Article | PubMed | ChemPort |
58. Goldsmith HH. A zygosity questionnaire for young twins: a research note. Behav Genet 1991; 21: 257–269. | PubMed | ISI | ChemPort |
59. Reed PW, Davis JL, Copeman JB et al. Chromosome-specific microsatellite sets for flourescence-based, semi-automated genome mapping. Nat Genet 1994; 7: 343–344. | Article |
60. Pritchard LE, Kawaguchi Y, Reed PW et al. Analysis of the CD3 gene region and type I diabetes: application of flourescence-based technology to linkage disequilibrium mapping. Hum Mol Genet 1995; 4: 197–202. | Article | PubMed | ISI | ChemPort |
61. Reed PW, Cucca F, Jenkins SC et al. Evidence for a type I diabetes susceptibility locus (IDDM10) on human chromosome 10p11–q11. Hum Mol Genet 1997; 6: 1011–1016. | Article | PubMed | ISI | ChemPort |
62. Falconer M, Introduction to Quantitative Genetics, 4th edn. Longman: New York, 1996.
63. Neale MC, Cardon L. Methodologies for Genetic Studies of Twins and Families. Kluwer Academic Publishers: Dordrecht; 1992.
64. Hall MA, Ahmadi KR, Norman PJ et al. Genetic influence on peripheral blood T lymphocyte levels. Genes Immun 2000; 1: 423–427. | Article | PubMed |
65. Abecasis GR, Cardon LR, Cookson WO. A general test of association for quantitative traits in nuclear families. Am J Hum Genet 2000; 66: 279–292. | Article | PubMed | ISI | ChemPort |
66. Amos CI. Robust variance-components approach for assesing genetic linkage in pedigrees. Am J Hum Genet 1994; 54: 535–543. | PubMed | ISI | ChemPort |
67. Fulker DW, Cherny SS, Sham PC, Hewitt JK. Combined linkage and association sib-pair analysis for quantitative traits. Am J Hum Genet 1999; 64: 259–267. | Article | PubMed | ISI | ChemPort |
68. Pratt SC, Daly MJ, Kruglyak L. Exact multipoint quantitative-trait linkage analysis in pedigrees by variance components. Am J Hum Genet 2000; 66: 1153–1157. | Article | PubMed | ISI | ChemPort |
69. Hedrick PW. Gametic disequilibrium measures: proceed with caution. Genetics 1987; 117: 331–341. | PubMed | ISI | ChemPort |
70. Lewontin RC. The interaction of selection and linkage. I. General considerations; heterotic models.Genetics 1964; 49: 49–67. | PubMed | ISI | ChemPort |
71. Xie X, Ott J. Testing linkage disequilibrium between a disease gene and a marker loci. Am J Hum Genet 1993; 53: 1107. | PubMed |
72. Zhao JH, Curtis D, Sham PC. Model-free analysis and permutation tests for allelic associations. Hum Heredity 2000; 50: 133–139. | Article |
73. Sobel E, Lange K. Descent graphs in pedigree analysis: application to haplotyping, location scores, and marker sharing statistics. Am J Hum Genet 1996; 58: 1323–1337. | PubMed | ISI | ChemPort |
74. Abecasis GR, Cookson WO. GOLD—graphical overview of linkage disequilibrium. Bioinformatics 2000; 16: 182–183. | Article | PubMed | ISI | ChemPort |

Acknowledgements

We thank all the twins for participating and the clinical nurses and administration team for their work. KRA is supported by a Chronic Diseases Research Foundation Fellowship.