Skip Navigation


Rheumatology Advance Access originally published online on January 27, 2007
Rheumatology 2007 46(5):763-771; doi:10.1093/rheumatology/kel443
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow CME/CE:
Take the course for this article:
 Rheumatology First Quarter 2007 Quiz
Right arrow All Versions of this Article:
46/5/763    most recent
kel443v2
kel443v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (15)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Carter, K. W.
Right arrow Articles by Brown, M. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Carter, K. W.
Right arrow Articles by Brown, M. A.
Related Collections
Right arrow Spondylarthropathies
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author 2007. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

Combined analysis of three whole genome linkage scans for Ankylosing Spondylitis

K. W. Carter, A. Pluzhnikov1, A. E. Timms2, C. Miceli-Richard3, C. Bourgain4, B. P. Wordsworth2, H. Jean-Pierre5, N. J. Cox1, L. J. Palmer, M. Breban6, J. D. Reveille7 and M. A. Brown2,8

Laboratory for Genetic Epidemiology, Western Australian Institute for Medical Research, UWA Centre for Medical Research, University of Western Australia, Ground Floor, B Block, Hospital Avenue, Nedlands, Western Australia 6009, Australia, 1Department of Human Genetics, The University of Chicago, Chicago, USA, 2Botnar Research Centre, Institute of Musculoskeletal Sciences, University of Oxford, Nuffield Orthopaedic Centre, Windmill Road, Headington, OX3 7LD, UK, 3Laboratoire de Génétique des Maladies Inflammatoires de l’Intestin et Fondation Jean Dausset/CEPH, Paris, 4INSERM U535, Villejuif, 5INSERM AVENIR U458 and AP-HP, Hôpital Robert Debré, 48 Bd Sérurier, 75019 Paris, 6Département d’Immunologie, Institut Cochin, INSERM U567, CNRS UMR 8104, IFR116, and Hôpital Ambroise Paré, Assistance Publique-Hôpitaux de Paris, Université Paris, Ile de France Ouest Paris, France, 7The University of Texas Health Science Center at Houston, Houston, Texas, USA and 8Centre for Immunology and Cancer Research, Princess Alexandra Hospital, Woolloongabba, Queensland, Australia.

Correspondence to: K. W. Carter. E-mail: kcarter{at}cyllene.uwa.edu.au


   Abstract
Top
 Abstract
Introduction
 Patients and methods
 Results
 Discussion
 Acknowledgements
 References
 
Objective. Ankylosing spondylitis (AS) is a debilitating chronic inflammatory condition with a high degree of familiality ({lambda}s = 82) and heritability (>90%) that primarily affects spinal and sacroiliac joints. Whole genome scans for linkage to AS phenotypes have been conducted, although results have been inconsistent between studies and all have had modest sample sizes. One potential solution to these issues is to combine data from multiple studies in a retrospective meta-analysis.

Methods. The International Genetics of Ankylosing Spondylitis Consortium combined data from three whole genome linkage scans for AS (n = 3744 subjects) to determine chromosomal markers that show evidence of linkage with disease. Linkage markers typed in different centres were integrated into a consensus map to facilitate effective data pooling. We performed a weighted meta-analysis to combine the linkage results, and compared them with the three individual scans and a combined pooled scan.

Results. In addition to the expected region surrounding the HLA-B27 gene on chromosome 6, we determined that several marker regions showed significant evidence of linkage with disease status. Regions on chromosome 10q and 16q achieved ‘suggestive’ evidence of linkage, and regions on chromosomes 1q, 3q, 5q, 6q, 9q, 17q and 19q showed at least nominal linkage in two or more scans and in the weighted meta-analysis. Regions previously associated with AS on chromosome 2q (the IL-1 gene cluster) and 22q (CYP2D6) exhibited nominal linkage in the meta-analysis, providing further statistical support for their involvement in susceptibility to AS.

Conclusion. These findings provide a useful guide for future studies aiming to identify the genes involved in this highly heritable condition.

KEY WORDS: Ankylosing Spondylitis, genome scans, meta-analysis


   Introduction
Top
 Abstract
 Introduction
Patients and methods
 Results
 Discussion
 Acknowledgements
 References
 
Ankylosing spondylitis (AS) is a common, chronic inflammatory arthropathy that primarily affects the spinal joints and sacroiliac joints, causing symptoms of pain and stiffness, and progressive fusion of involved joints. Peripheral arthritis and/or uveitis (inflammation of the eye) complicates approximately 40–50% of cases; less common non-spinal manifestations include enthesitis, aortitis and pneumonitis. Associations between AS and psoriasis and inflammatory bowel disease are well described. AS has an estimated prevalence of 1/1000 to 4/1000 in white populations [1, 2].

Genetic factors have long been implicated in AS, with recognition of association with the HLA-B27 gene over 30 yrs ago [3, 4]. Whilst over 90% of AS patients carry the gene, only approximately 1–6% of the general population HLA-B27 carriers develop AS [1, 2]. Twin and family studies suggest that this can be explained by the additional involvement of several other genes [5, 6]. There is evidence that other genes lying within the major histocompatibility complex (MHC) [7], interleukin-1 gene cluster [8–11], and the gene CYP2D6 influence susceptibility to AS [12, 13], but clearly a substantial proportion of the non-B27 heritability of AS is as yet unexplained.

The International Genetics of Ankylosing Spondylitis (GAS) Consortium was established in 2003 as an international effort to define the candidate regions and genes for AS with greater precision and reliability than can be achieved by smaller individual studies. We have gathered the results of three separate whole genome scans for AS and examined them for evidence of linkage with disease. These are the ‘Oxford’ [14], Groupe Français d’Etude Génétique des Spondyloarthropathies (GFEGS) [15], and North American Spondyloarthritis Consortium (NASC) datasets [16]. The ‘Oxford’ cohort consists of two subsets, the findings of a genomewide scan on the first of which was first reported in 1998 [17], and the second in 2001 [14]. The two sets were pooled for the current study. These datasets represent all currently published genome-wide linkage data in AS. Whilst a previous meta-analysis has been reported in AS, it did not use the original linkage data as we have here, but analysed only data found in figures and tables from previous linkage screens, did not take into account the differences in phenotype definition in the different datasets, nor was able to identify and correct marker position and order errors in the original datasets [18].

In addition to examining studies individually, we examined a combined dataset of the three studies pooled together giving each dataset equal weighting (referred to as the ‘pooled’ analysis), and compared this with a weighted meta-analysis of the three studies to explore the overall combined effect across the studies (referred to as the ‘weighted’ meta-analysis). This weighted meta-analysis takes into account the differences in size, marker density and linkage information content extracted by each screen. The use of meta-analysis techniques should improve on the power of the individual datasets, allowing us to examine in finer detail the genetic effects across three independent populations.


   Patients and methods
Top
 Abstract
 Introduction
 Patients and methods
Results
 Discussion
 Acknowledgements
 References
 
Study populations
There are significant differences in the disease definitions used, types of families studied, and marker sets genotyped between the three studies. Both the Oxford and NASC cohorts consisted entirely of families with AS defined by the modified New York criteria [19]. In contrast, the GFEGS dataset included a substantial proportion of cases with ‘spondyloarthritis’ (SpA), as defined by the ESSG diagnostic criteria [20]. These cases may have AS, or may have features of seronegative arthritis but may not meet the strict ‘AS’ radiographic criteria of the modified New York criteria, and are clinically and genetically related to AS. The GFEGS dataset was studied either as a whole (GFEGS-SpA, including all those meeting the ESSG Criteria and modified New York Criteria for AS), or considering only those cases with AS, as defined by the modified New York criteria (GFEGS-AS). As it is likely that in time many cases with SpA will progress and meet the modified New York Criteria for AS, the SpA group not currently meeting the AS criteria were not analysed separately.

The GFEGS dataset consisted of 431 individuals from 86 pedigrees with a total of 139 affected sibling pairs, genotyped with a total of 346 markers from the Applied Biosystems Prism Linkage Mapping Set Version 2.0 (LMSV2) (Applied Biosystems). The NASC dataset was comprised of 2203 individuals from 232 families with a total of 245 sibling pairs, genotyped using the ABI LMSV2 marker set. The Oxford dataset consisted of 1102 individuals from 198 pedigrees with a total of 251 affected sibling pairs. These were genotyped using both the ABI LMSV2 and the Medical Research Council (UK) microsatellite marker sets [21]. The family makeup of each screen is given in Table 1.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Family compositions of genome screens employed in meta-analysisa

 
All study populations were collected under appropriate institutional ethics approvals, as detailed in the original papers (referenced earlier).

Consensus map
A consensus map that incorporated all typed markers in each of the three datasets was constructed to pool raw genotype data across samples. This map provided a framework of markers upon which typed markers were integrated using PYGMALION [22]; markers were ordered using the deCODE recombination marker map [23].

Linkage analysis
Possible errors in the pedigree and genotype data were assessed using the RELCHECK [24], PEDCHECK [25] and MERLIN [26] programs. These analyses suggested that the maximal potential error rate in any one dataset was less than 1% of all genotypes. Our error-checking procedure resulted in the exclusion of very few families at a given marker from any of the datasets. Marker allele frequencies were estimated using the MERLIN program by maximum likelihood estimation [27].

Multipoint non-parametric linkage analysis was performed using the MERLIN v0.10.2 pedigree analysis software package [27]. P-values were calculated by simulation and then converted to a –log10 scale and plotted using JLGraph [28]. Because the dataset was simulated 10 000 times, the minimum reported P-value is <10–4, and maximum –log10 (P-value) is 4.

Pooled analysis
The three individual AS-alone datasets were combined to create a single pooled AS-alone dataset comprising all markers and families. This pooled dataset was analysed for linkage using Merlin (with the common integrated map) and empirical P-values calculated as previously described. A pooled SpA dataset was created by combining the markers and families from the AS alone and GFEGS-SpA datasets (i.e. all cases from the three datasets that meet the ESSG Criteria, most of whom also meet the modified New York Criteria), and analysed similarly. Because of the strong linkage between AS and markers contained in the MHC on chromosome 6 and the limits of our simulations, six markers contained in the Oxford dataset and four markers contained in the NASC dataset produced a P-value that was effectively zero. As this result could not be plotted on a –log10 scale, these markers were plotted with the arbitrary value of 10 (relative to surrounding markers). Seven markers in this same location in the both the AS-alone and SpA pooled datasets were plotted in the same manner.

Meta-analysis
We analysed each of three datasets for linkage to the genotyped markers. In order to perform a meta-analysis of the whole genome linkage scan results (for AS only affection), we then pooled linkage statistics across the datasets using a weighted combination of z-score approach. Given the varying information content and population sizes across the three genome scans, we chose an approach based on weighting for both information content and study size as previously described by [29].

Using the weighted meta-analysis approach, the z-scores for each of the 338 common markers across the scans were combined. These weighted z-scores follow a standard normal distribution (mean = 0, S.D. = 1), from which P-values were determined. As with the empirical linkage results, we converted the resulting P-values to a –log10 scale for plotting using JLGraph. Three marker locations contained in chromosome 6 produced P-values of zero, and as described previously were assigned the value of 10 for plotting purposes.

IBD sharing by affected sibling pairs and {lambda}(LOCUS) values were determined using MapMaker/Sibs [30] weighted according to the number of affected sibling pairs available in each cohort. The contribution of each locus to the overall sibling recurrence risk was calculated assuming a sibling recurrence risk ratio of 82 [6], and multiplicative interaction between loci [31], the most likely model in AS [6].


   Results
Top
 Abstract
 Introduction
 Patients and methods
 Results
Discussion
 Acknowledgements
 References
 
The results of the three individual whole genome linkage scans for AS are presented in Fig. 1. The results of the pooled AS-alone analysis and weighted meta-analyses are compared in Fig. 2 and Fig. 3 presents the results of the pooled AS-alone and the pooled SpA analyses. Significant results are indicated in Figs 1–3GoGo by small P-values (large values on a –log10 scale) forming peaks over relevant markers. Note that the positioning of results within each of the figures reflect that most p-terminal markers used in each study are not located at 0 cM on the chromosomal genetic map.


Figure 1
View larger version (33K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIG. 1. Ankylosing spondylitis whole genome linkage scans. A circle indicates the marker has a P-value < 0.05, a square indicates P-values < 0.01, a triangle indicates P-values < 0.001. Linkage with AS affection status is shown for GFEGS, Oxford and NASC datasets in red, blue and green respectively. Please note, chromosome 6 has a different –log10 P-value scale.

 

Figure 2
View larger version (30K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIG. 2. Ankylosing spondylitis weighted meta-analysis vs pooled dataset. A circle indicates the marker has a P-value < 0.05, a square indicates P-values < 0.01, a triangle indicates P-values < 0.001. Weighted meta-analysis is shown in red, while pooled linkage scan is shown in blue. Please note, chromosome 6 has a different –log10 P-value scale.

 

Figure 3
View larger version (32K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		FIG. 3. Pooled SpA dataset vs pooled AS-alone dataset. A circle indicates the marker has a P-value < 0.05, a square indicates P-values < 0.01, a triangle indicates P-values < 0.001. Pooled SpA is shown in red, while pooled AS-alone scan is shown in blue. Please note, chromosome 6 has a different –log10 P-value scale.

 
Individual scans
Considering the individual scans (Fig. 1), each shows highly significant linkage to the region of chromosome 6 encoding the MHC (P < 10–5). ‘Nominal’ or greater evidence of linkage (P < 0.05) was observed for two or more scans on chromosomes 1, 3, 5, 7, 9, 10, 13, 16, 17, 19 and 22. Regions achieving at least moderate evidence of linkage (P < 0.05 in three or more individual datasets or the meta-analyses) are given in Table 2. These were observed in the GFEGS-SpA scan on chromosomes 9 (129 cM, P = 0.0003), 11 (53 cM, P = 0.003), 12 (31 cM, P = 0.002), and 17 (93 cM, P = 0.005). When considering cases of AS alone, of these linkages only the chromosome 12 linkage remained significant (P < 0.01). In the NASC dataset, moderate or greater linkage was observed on chromosome 3 (95 cM, P = 0.002), 4 (35 cM, P = 0.007), 5 (204 cM, P = 0.007), 6q (161 cM and 172 cM, P = 0.002), 16 (50 cM, P = 0.005), and 22 (56 cM, P = 0.009). In the Oxford families, moderate or greater linkage was observed on chromosomes 1 (59 cM, P = 0.003; 272 cM, P = 0.006), 2 (135 cM, P = 0.0003), 5 (166 cM, P = 0.009), 10 (127 cM, P = 0.003), 16 (101 cM, P = 3 x 10–5), and 19 (76 cM, P = 0.0003).


View this table:
[in this window]
[in a new window]

  		TABLE 2. Markers achieving P ≤ 0.05 in three or more datasets/analysesa

 
Pooled analysis
In addition to the very significant evidence of linkage with the 30–70 cM region of chromosome 6 containing the MHC, the pooled analysis results show markers with moderate evidence of linkage (P < 0.01) in chromosomes 3, 4, 5, 6, 10, 11, 13, 16, 17 and 22 (Table 2). Clusters of significant marker results were seen on chromosome 1 at approximately 170 cM (P < 0.05), chromosome 3 at 88–130 cM (P < 0.01) and 198 cM (P < 0.05), and with several markers across chromosome 17.

Consistent with each of the individual scans and the AS-alone weighted meta-analysis, the pooled analysis exhibited significant evidence of linkage (P < 0.00001) with the 30–90 cM region of chromosome 6 in the AS-alone and AS-SpA datasets. Clusters of significant marker results are also seen on chromosomes 1 (P < 0.01), 3 (P < 0.01), 10, 11, 12, and 13 (P < 0.01), 17 (P < 0.001), 21 (P < 0.01) and 22 (P < 0.01).

Weighted meta-analysis
Considering the analyses of AS alone, highly significant linkage was observed with all three individual scans and with the combined analyses (P < 10–5) to the 37–75 cM region of chromosome 6, consistent for linkage with the location of the HLA-B27 marker. In the weighted AS-alone meta-analysis, the next strongest evidence of linkage was found at chromosome 16q (maximum linkage at 99 cM, P = 1.8 x 10–4). Evidence for linkage between markers in this this region with AS was seen in both the NASC and Oxford datasets (P = 0.02 at 99 cM and P = 3 x 10–5 at 101 cM, respectively). Other regions achieving moderate evidence of linkage (P < 0.01) were identified on chromosomes 3, 10 and 19. On chromosome 3, peak linkage was seen at 202 cM (P = 0.005), with support in the GFEGS, NASC and Oxford datasets (P = 0.02, 0.03 and 0.03, respectively, each at 202 cM). On chromosome 10, peak linkage was seen at 127 cM (P = 0.008), with support in the NASC and Oxford datasets (P = 0.02 at 113 cM and P = 0.003 at 127 cM, respectively). A broad region of linkage was observed on chromosome 19, the peak of which lay at 76 cM (P = 0.005), with support in the NASC and Oxford datasets, although the peaks of linkage in these studies were quite separated (P = 0.03 at 109 cM and P = 0.0003 at 76 cM, respectively).

The {lambda}(LOCUS) value for chromosome 6p was 4.5, representing 34% of the familiality of AS, assuming a multiplicative interaction between loci. For chromosome 16q, the {lambda}(LOCUS) value was 1.6, representing 11% of the familiality of AS.

Previously associated regions
Association has previously been reported between AS and the IL-1 gene cluster on chromosome 2q13 (at 126 cM), and the gene CYP2D6 on chromosome 22q13.2 (at 50 cM). Considering the IL-1 gene cluster, this region showed evidence of linkage in the Oxford dataset (135 cM, P = 0.0003) and in the weighted meta-analysis (135 cM, P = 0.01), but not in other individual datasets. The CYP2D6 region showed nominal evidence of linkage in the North American and Oxford datasets (56 cM, P = 0.009 and 37 cM, P = 0.04, respectively) and in the pooled analysis (56 cM, P = 0.003).

Examining heterogeneity of findings between the three studies (for AS-alone), using the 338 common markers, we found that 114 (approx 33%) of the marker locations had substantial heterogeneity (>50%) using the I2 measure [32], and 105 show substantial heterogeneity (P < 0.1) using Cochran's Q test. Groups of markers showing substantial heterogeneity are discussed further in the Discussion.


   Discussion
Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
Acknowledgements
 References
 
This study was designed to investigate the evidence for linkage to AS phenotypes in a retrospective analysis of combined genomewide data from three separate studies. Meta-analysis can potentially improve on the power of individual genome screens to identify linkages, and help clarify the significance of linkage results where there is inconsistent replication across screens (such as [29, 33–35]). We have performed meta-analysis by both pooling results and a weighted combination of linkage statistics across studies. These two approaches each have strengths and weaknesses. Weighted meta-analysis more correctly accounts for the difference in size and information content between the different scans, but only uses markers common to the three studies (n = 338). The pooled analysis uses all markers available potentially increasing its power, but weights each study equally. Where there is discordance between the studies this can lead to the findings from smaller studies having a disproportionate effect on pooled results.

The study confirms on a genome-wide level the ‘highly significant’ linkage (P < 3 x 10–7) of the region of chromosome 6 encoding the MHC, and also identifies ‘suggestive’ linkage (P < 7.4 x 10–4) on chromosomes 6q, 10q and 16q. Several other regions, whilst not achieving ‘suggestive’ or greater evidence of linkage, found support in two or more screens such as to suggest that it is likely that they are true positive findings.

In addition to the highly significant linkage observed on chromosome 6p around the location of HLA-B27, there are a number of other interesting marker patterns on the chromosome 6 which illustrate the advantage of the meta-analysis approach compared with comparisons of individual screens. Between 140 cM and 175 cM in the NASC linkage results in Fig. 1, there are eight markers with significant (P < 0.01) linkage to AS. In contrast, the Oxford and GFEGS datasets return no significant markers for this region. The pooled analysis, as expected given the fact that the NASC dataset contributes 42% of the total studied, also returns significant results for the region. The weighted meta-analysis is therefore of particular interest, given the lack of replication across this region in the remaining two studies, identifying four significant markers for linkage in this region, the most significant result being at 160.64 cM (marker name AFM269ze1/D6S441, P = 0.007). Examining the three linkage scan results individually could have missed this linkage, as the results were not reproduced across studies, demonstrating the value of the meta-analysis method. These findings strongly support the existence of further non-MHC genes involved in susceptibility to AS on chromosome 6q.

Linkage to chromosome 16q in AS was first reported by the Oxford group in 1998 [17], with further support for this being a true positive finding coming from their further genome screen [14] and the NASC study [16]. However, no linkage was observed in this region in the GFEGS study. In the weighted meta-analysis this region is the most significant non-MHC region, achieving P = 0.0002, the peak lying at 99 cM. Results for chromosome 10 surrounding 124 cM were also significant in the NASC and Oxford but not the GFEGS dataset. In particular, the marker at location 127 cM (AFM331xa9/D10S597) was significant in the meta-analysis (P = 0.004) and the pooled analysis (P = 0.0005). Thus, despite the inconsistent replication of this finding in different genome screens, the meta-analysis findings suggest that these are probably true positive findings, warranting further study.

Other regions potentially containing AS genes may only fulfil less stringent statistical threshold but are worth pursuing, particularly if there is support from more than one screen. Other than the regions described earlier, nominal or greater evidence of linkage (P < 0.05) was observed for two or more scans on chromosomes 1, 3, 5, 9, 13, 17, 19 and 22. Although there is clearly less certainty about these findings, there is at least a moderate probability that they contain disease-susceptibility genes.

Substantial differences were observed between the findings of the GFEGS study and those from the NASC and Oxford cohorts, the main exception being the consistent linkage to the MHC in all three screens. Potential explanations for this include the lower sample size of the GFEGS study, differences in structure of the families studied (the GFEGS cohort included a higher proportion of large pedigrees rather than affected-sibling pair families than the other screens), or the diagnostic scheme applied. The GFEGS–AS dataset is significantly smaller than the NASC or Oxford datasets, which likely explains the difference in the findings for that study. These differences between the scans at least partly explain the differences in finding between the pooled analysis and weighted meta-analyses.

Examining the differences between the pooled AS-alone and SpA analyses, a number of interesting patterns occur. As shown in Fig. 3, the evidence for linkage of chromosome 16q with disease is evident in the AS-alone dataset and not in the pooled SpA dataset. However, as noted previously, the GFEGS dataset is smaller than the NASC and Oxford datasets, likely explaining the difference between the pooled analysis and weighted meta-analysis findings at this region. On chromosome 21 at ~40–56 cM the individual scans and the weighted meta-analysis reveal no evidence for significant linkage, while the pooled SpA analysis reveals a cluster of markers suggestive of linkage, peaking at marker D21S268 (49 cM, P = 0.0009). Similarly, a closer examination of markers in the 35–50 cM region of chromosome 10 also shows evidence for linkage (P < 0.01) with markers in the pooled SpA analysis, that do not appear in the weighted meta-analysis. These results warrant further investigation to determine whether this is a result of the combined pooled analysis composition, or reflects heterogeneity in disease definition between the pooled datasets, or some other confounding factor.

Comparison of pooled analysis and weighted meta-analysis results are helpful in the interpretation of chromosomal regions suggestive of linkage with AS-alone that are not replicated in the SpA datasets. Linkage observed with chromosome 10 at 127 cM in the weighted meta-analysis of AS-alone and SpA, and in the pooled AS-alone analysis, returned no significant evidence for linkage within the pooled SpA analysis. The broad region of chromosome 19 identified for linkage in the AS-alone meta-analysis was not identified at all within the SpA dataset (either in the meta-analysis or the pooled studies). Notably neither the GFEGS-SpA nor GFEGS-AS datasets provided evidence for linkage to this chromosome. This absence of linkage would least affect the AS-alone weighted meta-analysis. Thus, this disparity may be due to identification of marker locations specific to AS-alone rather that the more general condition of SpA, an observation warranting further study.

A closer examination of the population heterogeneity calculation across the studies reveals that the majority of markers showing substantial heterogeneity fall in regions where no significant evidence for linkage was discovered. The only major block of linkage results that overlay a large region of markers showing heterogeneity is on chromosome 12 at ~66–112 cM in the pooled SpA analysis. However, the weighted meta-analysis and pooled AS-alone analysis reveal no significant evidence for linkage in this region, suggesting that perhaps the linkage with SpA in this region may reflect population heterogeneity. Other large blocks of markers showing substantial heterogeneity are located at the q-terminal of chromosome 13 and across the entire chromosome 14, where no significant evidence for linkage was revealed in any of the analyses.

In comparison with a previously published meta-analysis [36], this study is more accurate and powerful. The previous study employed data extrapolated from published graphs and tables and did not accurately differentiate between AS and SpA, the AS families from the GFEGS screen being excluded from the AS-alone analysis. In contrast, we have used the original genotype data allowing for better genotype error checking, correction of marker maps to take into account the most recent genome maps, precise multipoint analysis rather than division into 30 cM bins, and pooled and weighted analyses. AS and SpA analysis have been presented separately. Lee et al.'s study report only three loci achieving P < 0.01 (on chromosomes 6, 16 and 19), which were each identified in this screen. Two other regions achieved nominal levels of significance with the two statistical scores used in this study (Psumrnk, Pord < 0.05, >0.01), lying on chromosome 17p and 9q. These areas overlap with regions identified in the current study. However, the level of significance we report is generally greater than reported by Lee et al., which also did not identify several regions we report, including on chromosomes 3, 5, 10 and 13 which were present in three or more datasets or analyses (Table 2).

In conclusion, this analysis provides an important step forward for narrowing down chromosomal locations that may hold the key to discovering the genetic linkages of AS. By performing a meta-analysis, we have identified several common markers that bear further investigation. In particular, we found ‘suggestive’ linkage on chromosomes 10q and 16q, and moderate evidence of linkage (P < 0.01) on chromosomes 3, 4, 5, 6, 10, 11, 16, 17 and 22. Further, regions on chromosome 2 (the IL-1 gene cluster) and 22 (the gene CYP2D6), previously reported to be associated with AS, showed at least nominal evidence of linkage (P < 0.05) with AS in the meta-analysis. These findings strongly support that this disease is at least oligogenic, and provide a valuable road-map for future studies aiming to identify the genes involved in this condition.

Formula


   Acknowledgements
Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 Acknowledgements
References
 
KWC was funded by the Australian Research Council. LJP and KWC were supported by the Wind Over Water Foundation (AUS). MAB and AET were funded by the Arthritis Research Campaign (UK). The work by GFEGS was supported by grants from Programme Hospitalier de Recherche Clinique (PHRC), Société Française de Rhumatologie (SFR), Association de Recherche sur la Polyarthrite (ARP), Institut National de la Santé et de la Recherche Médicale (INSERM), and the Pfizer-Pharmacia-Searle and Schering-Plough companies. The authors wish to acknowledge the support of members of the National Ankylosing Spondylitis Society (UK) and the Spondylitis Society of America and their colleagues in many countries who have allowed their families to contribute to these studies. We would like to acknowledge the assistance of Sylvie Pichelin (CICR, Brisbane, Australia) in manuscript preparation.

The authors have declared no conflicts of interest.


   References
Top
 Abstract
 Introduction
 Patients and methods
 Results
 Discussion
 Acknowledgements
 References
 

  1. van der Linden S, Valkenburg H, Cats A. (1983) The risk of developing ankylosing spondylitis in HLA-B27 positive individuals: a family and population study. Br J Rheumatol 22:4 Suppl 2, 18–9.[Medline]
  2. Braun J, Bollow M, Remlinger G, Eggens U, Rudwaleit M, Distler A, Sieper J. (1998) Prevalence of spondylarthropathies in HLA-B27 positive and negative blood donors. Arthritis Rheum 41:58–67.[CrossRef][Web of Science][Medline]
  3. Brewerton DA, Hart FD, Nicholls A, Caffrey M, James DC, Sturrock RD. (1973) Ankylosing spondylitis and HL-A 27. Lancet 1:904–7.[Web of Science][Medline]
  4. Schlosstein L, Terasaki PI, Bluestone R, Pearson CM. (1973) High association of an HL-A antigen, W27, with ankylosing spondylitis. N Engl J Med 288:704–6.[Web of Science][Medline]
  5. Brown MA, Kennedy LG, MacGregor AJ, Darke C, Duncan E, Shatford JL, Taylor A, Calin A, Wordsworth P. (1997) Susceptibility to ankylosing spondylitis in twins: the role of genes, HLA, and the environment. Arthritis Rheum 40:1823–8.[Web of Science][Medline]
  6. Brown MA, Laval SH, Brophy S, Calin A. (2000) Recurrence risk modelling of the genetic susceptibility to ankylosing spondylitis. Ann Rheum Dis 59:883–6.[Abstract/Free Full Text]
  7. Brown MA, Kennedy LG, Darke C, Gibson K, Pile KD, Shatford JL, Taylor A, Calin A, Wordsworth BP. (1998) The effect of HLA-DR genes on susceptibility to and severity of ankylosing spondylitis. Arthritis Rheum 41:460–5.[CrossRef][Web of Science][Medline]
  8. McGarry F, Neilly J, Anderson N, Sturrock R, Field M. (2001) A polymorphism within the interleukin 1 receptor antagonist (IL-1Ra) gene is associated with ankylosing spondylitis. Rheumatology 40:1359–64.[Abstract/Free Full Text]
  9. van der Paardt M, Crusius JB, Garcia-Gonzalez MA, et al. (2002) Interleukin-1beta and interleukin-1 receptor antagonist gene polymorphisms in ankylosing spondylitis. Rheumatology 41:1419–23.[Abstract/Free Full Text]
  10. Maksymowych WP, Reeve JP, Reveille JD, et al. (2003) High-throughput single-nucleotide polymorphism analysis of the IL1RN locus in patients with ankylosing spondylitis by matrix-assisted laser desorption ionization-time-of-flight mass spectrometry. Arthritis Rheum 48:2011–8.[CrossRef][Web of Science][Medline]
  11. Timms AE, Crane AM, Sims AM, et al. (2004) The interleukin 1 gene cluster contains a major susceptibility locus for ankylosing spondylitis. Am J Hum Genet 75:587–95.[CrossRef][Web of Science][Medline]
  12. Brown MA, Edwards S, Hoyle E, et al. (2000) Polymorphisms of the CYP2D6 gene increase susceptibility to ankylosing spondylitis. Hum Mol Genet 9:1563–6.[Abstract/Free Full Text]
  13. Beyeler C, Armstrong M, Bird HA, Idle JR, Daly AK. (1996) Relationship between genotype for the cytochrome P450 CYP2D6 and susceptibility to ankylosing spondylitis and rheumatoid arthritis. Ann Rheum Dis 55:66–8.[Abstract/Free Full Text]
  14. Laval SH, Timms A, Edwards S, et al. (2001) Whole-genome screening in ankylosing spondylitis: evidence of non-MHC genetic-susceptibility loci. Am J Hum Genet 68:918–26.[CrossRef][Web of Science][Medline]
  15. Miceli-Richard C, Zouali H, Said-Nahal R, et al. (2004) Significant linkage to spondyloarthropathy on 9q31–34. Hum Mol Genet 13:1641–8.[Abstract/Free Full Text]
  16. Zhang G, Luo J, Bruckel J, et al. (2004) Genetic studies in familial ankylosing spondylitis susceptibility. Arthritis Rheum 50:2246–54.[CrossRef][Web of Science][Medline]
  17. Brown MA, Pile KD, Kennedy LG, et al. (1998) A genome-wide screen for susceptibility loci in ankylosing spondylitis. Arthritis Rheum 41:588–95.[CrossRef][Web of Science][Medline]
  18. Lee YH, Rho YH, Choi SJ, Ji JD, Song GG. (2005) Ankylosing spondylitis susceptiblity loci defined by genome-search meta-analysis. J Hum Genet 50:453–9.[CrossRef][Web of Science][Medline]
  19. van der Linden S, Valkenburg HA, Cats A. (1984) Evaluation of diagnostic criteria for ankylosing spondylitis. A proposal for modification of the New York criteria. Arthritis Rheum 27:361–8.[Web of Science][Medline]
  20. Dougados M, van der Linden S, Juhlin R, et al. (1991) The European Spondylarthropathy Study Group preliminary criteria for the classification of spondylarthropathy. Arthritis Rheum 34:1218–27.[Web of Science][Medline]
  21. Reed PW, Davies JL, Copeman JB, et al. (1994) Chromosome-specific microsatellite sets for fluorescence-based, semi-automated genome mapping. Nat Genet 7:390–5.[CrossRef][Web of Science][Medline]
  22. Pluzhnikov A, Wen W, Luban S, Chen Y, Cox NJ. (2003) Efficient construction of genetic maps based on publicly available physical and genetic map data. Am J Hum Genet : Suppl A438.
  23. Kong A, Gudbjartsson DF, Sainz J, et al. ( 2002) A high-resolution recombination map of the human genome. Nat Genet 31:241–7.[CrossRef][Web of Science][Medline]
  24. Broman KW and Weber JL. (1998) Estimation of pairwise relationships in the presence of genotyping errors. Am J Hum Genet 63:1563–4.[CrossRef][Web of Science][Medline]
  25. O’Connell JR and Weeks DE. (1998) PedCheck: a program for identification of genotype incompatibilities in linkage analysis. Am J Hum Genet 63:259–66.[CrossRef][Web of Science][Medline]
  26. Abecasis GR, Cherny SS, Cookson WO, Cardon LR. (2002) Merlin–rapid analysis of dense genetic maps using sparse gene flow trees. Nat Genet 30:97–101.[CrossRef][Web of Science][Medline]
  27. Abecasis GR, Cherny SS, Cookson WO, Cardon LR. (2002) Merlin–rapid analysis of dense genetic maps using sparse gene flow trees. Nat Genet 30:97–101.[CrossRef][Web of Science][Medline]
  28. Carter KW and Palmer LJ. (2004) Java linkage graph (online). In. http:/www.genepi.com.au/projects/jlgraph.
  29. Dempfle A and Loesgen S. (2004) Meta-analysis of linkage studies for complex diseases: an overview of methods and a simulation study. Ann Hum Genet 68:69–83.[CrossRef][Web of Science][Medline]
  30. Kruglyak L, Daly MJ, Reeve-Daly MP, Lander ES. (1996) Parametric and nonparametric linkage analysis: a unified multipoint approach. Am J Hum Genet 58:1347–63.[Web of Science][Medline]
  31. Risch N. (1987) Assessing the role of HLA-linked and unlinked determinants of disease. Am J Hum Genet 40:1–14.[Web of Science][Medline]
  32. Higgins J and Thompson S. (2002) Quantifying heterogeneity in a meta-analysis. Statistics in medicine 21:1539–58.[CrossRef][Web of Science][Medline]
  33. Fisher SA, Abecasis GR, Yashar BM, et al. (2005) Meta-analysis of genome scans of age-related macular degeneration. Hum Mol Genet 14:2257–64.[Abstract/Free Full Text]
  34. Koivukoski L, Fisher SA, Kanninen T, et al. (2004) Meta-analysis of genome-wide scans for hyptertension and blood pressure in Caucasians shows evidence of susceptibility regions on chromosomes 2 and 3. Hum Mol Genet 13:2325–32.[Abstract/Free Full Text]
  35. van Heel DA, Fisher SA, Kirby A, Daly MJ, Rioux JD, Lewis CM. (2004) Consortium GSM-AGotIIG. Inflammatory bowel disease susceptibility loci defined by genome scan meta-analysis of 1952 affected relative pairs. Hum Mol Genet 13:763–70.[Abstract/Free Full Text]
  36. Lee YH, Rho YH, Choi SJ, Ji JD, Song GG. (2005) Ankylosing spondylitis susceptibility loci defined by genome-search meta-analysis. J Hum Genet 50:453–9.[CrossRef][Web of Science][Medline]
Submitted 26 June 2006; revised version accepted 12 December 2006.
Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Ann Rheum DisHome page
E Dernis, R Said-Nahal, M-A D'Agostino, P Aegerter, M Dougados, and M Breban
Recurrence of spondylarthropathy among first-degree relatives of patients: a systematic cross-sectional study
Ann Rheum Dis, April 1, 2009; 68(4): 502 - 507.
[Abstract] [Full Text] [PDF]

Home page
 Ann Rheum DisHome page
D Harvey, J J Pointon, C Sleator, A Meenagh, C Farrar, J Y Sun, D Senitzer, D Middleton, M A Brown, and B P Wordsworth
Analysis of killer immunoglobulin-like receptor genes in ankylosing spondylitis
Ann Rheum Dis, April 1, 2009; 68(4): 595 - 598.
[Abstract] [Full Text] [PDF]

Home page
 Rheumatology (Oxford)Home page
M. A. Brown
Breakthroughs in genetic studies of ankylosing spondylitis
Rheumatology, February 1, 2008; 47(2): 132 - 137.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow CME/CE:
Take the course for this article:
 Rheumatology First Quarter 2007 Quiz
Right arrow All Versions of this Article:
46/5/763    most recent
kel443v2
kel443v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (15)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Carter, K. W.
Right arrow Articles by Brown, M. A.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Carter, K. W.
Right arrow Articles by Brown, M. A.
Related Collections
Right arrow Spondylarthropathies
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?