© 2005 American Thoracic Society doi: 10.1164/rccm.2502006
Haplotypes and Asthma
a Wellcome Trust Centre for Human Genetics Oxford, United Kingdom The asthma susceptibility gene GPR154 (G-protein–coupled receptor 154) was recently positionally cloned in families from founder populations in Finland and Canada (1). Novel genes such as GPR154 are of interest because of what they tell us about basic mechanisms of disease, and the risk for disease that they confer in the general population. GPR154 has similarities to oxytocin and vasopressin receptors and produces calcium channel activity when bound by its peptide ligand (2, 3). GPR154 is abundant in the retina and hypothalamus, but is also expressed in airway epithelium and upregulated by inflammation (1). This type of expression shares similarities to other genes which influence epithelial diseases such as asthma, psoriasis, and inflammatory bowel disease (4). The influence of polymorphisms in GPR154 on susceptibility to childhood asthma is reported in this issue of the AJRCCM (pp. 1089–1095) in an article by Melén and colleagues (5). The authors examined seven single nucleotide polymorphisms (SNPs) from a region of 70 kilobases containing GRP154. The seven SNPs were chosen because previous data suggested that they summarized the genetic information from the region (1). They were genotyped in children aged 4 from the BAMSE birth cohort of urban Swedish schoolchildren, as well as the multicenter PARSIFAL study of children from rural areas in Sweden, Switzerland, Holland, Germany, and Austria. Both the examination of the SNPs on their own and their combination into haplotypes demonstrated modest association with susceptibility to childhood asthma. The SNPs and haplotypes showing association differed from those identified in the original study (1), and the results also differed between the BAMSE and PARSIFAL subjects. This type of weakly positive result is not unexpected, and indeed has been the experience with many genes positionally cloned for complex diseases. BAMSE and PARSIFAL are population samples, and therefore asthma in their subjects would be mild, and possibly different from the more severe asthma recruited into genetic linkage and positional cloning studies. Europe is not homogeneous genetically, and there is a recognized gradient of polymorphisms across the continent in many genes. This difference is observable in the haplotype frequencies reported for the different components of the PARSIFAL study (Table E4 in the online supplement to Reference 5). Alpine farms and Stockholm city apartments differ markedly in their microbial and allergen contents, and strong environmental heterogeneity is present amongst the different populations in the study. While genetic admixture can give spurious false positive results, the likely outcome of environmental admixture is that associations will be obscured. It is of interest to examine the use of haplotypes in genetic studies. A haplotype is the linear combination of polymorphisms on a single chromosomal segment. Meiotic recombination between parental chromosomes occurs with a frequency of 1 to 2 per chromosome per generation. Even after hundreds of generations, SNPs in close proximity are not randomly assorted in the population and are said to be in linkage disequilibrium (LD). In regions of strong LD, dominant common haplotypes are observable within the population.
Haplotypes are helpful only in three circumstances. They sometimes become important when the side-by-side combination of SNPs on a particular chromosomal segment is required to alter gene function, such as the combination of SNPs in IL4-R To minimize genotyping for association studies, "tag" SNPs can be chosen from regions of high LD (known as blocks). Tag SNPs are intended to capture most of the information within the block (8). However, blocks are often fragmentary, and 30% of SNPs are not contained within block structures. As a consequence, SNP tagging is of most use in the early mapping of disease associations. Haplotypes usually have to be estimated from the genotype data, with random assignation whenever the haplotype cannot be inferred unambiguously. A significant proportion of haplotypes in any study will consequently be inaccurately assigned. The number of potential haplotypes doubles for each SNP studied. Even when LD reduces this number, statistical power is lost because of the extra degrees of freedom introduced by studying additional associations. Failure to include these extra degrees of freedom in analysis leads to multiple uncorrected tests of significance, with resulting confusion in interpreting and replicating results (9, 10). It is therefore not completely surprising that concentrating on SNPs in a standard logistic regression framework familiar to epidemiologist is more efficient statistically than blocking, tagging, and inferring haplotypes (11). The results from Melén and colleagues (5) agree with this observation, as there was no increase in the level of significance when haplotypes were tested for association compared with SNPs alone. The original discovery of GPR154 was made in founder populations (1). Because such populations descended recently from few individuals, minimal recombination appears within ancestral haplotypes and LD can extend very long distances. Our group has also been mapping asthma susceptibility genes on the short arm of chromosome 7 (12), and has recently shown that the bacterial recognition molecule NOD1 (which is four megabases away from GPR154) influences asthma and allergy in outbred populations (13). The possibility that NOD1 is contributing to linkage or association in the families used to identify GPR154 should be tested. These complications should not induce pessimism about GPR154 and asthma. The original genetic findings were strong, and the suggestion from the study of Melén and coworkers (5) is that the full information from the locus has not yet been brought out. Given the cost and the complexity of both large-scale epidemiology and large-scale genetics, this is not surprising. This paper is also an important testimonial to the high degree of cooperation required between groups to build a genuine understanding of the genetic susceptibility to asthma. FOOTNOTES Conflict of Interest Statement: M.M. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; P.H. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript; W.C. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. REFERENCES
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
(