Human Molecular Genetics, 2000, Vol. 9, No. 20 2947-2957
© 2000 Oxford University Press
Major factors influencing linkage disequilibrium by analysis of different chromosome regions in distinct populations: demography, chromosome recombination frequency and selection
1Dipartimento di Scienze Biomediche e Biotecnologie, University of Cagliari, Via Jenner, Cagliari 09121, Italy, 2Servizio di Diabetologia Pediatrica, Ospedale G. Brotzu, Via Peretti, Cagliari 09121, Italy, 3Wellcome Trust Centre for Molecular Mechanisms in Disease, University of Cambridge, Addenbrooke’s Hospital, Cambridge CB2 2XY, UK and 4Ambulatorio di Medicina di base di Gavoi, ASL 3, Via Manno 2, Gavoi 08020, Italy
Received 8 August 2000; Revised and Accepted 24 October 2000.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Linkage disequilibrium (LD) mapping of disease genes is complicated by population- and chromosome-region-specific factors. We have analysed demographic factors by contrasting intermarker LD results obtained in a large cosmopolitan population (UK), a large genetic isolate (Sardinia) and a subisolate (village of Gavoi) for two regions of the X chromosome. A dramatic increase of LD was found in the subisolate. Demographic history of populations therefore influences LD. Chromosome-region-specific effects, namely the pattern and frequency of homologous recombination, were next delineated by the analysis of chromosome 6p21, including the HLA region. Patterns of global LD in this region were very similar in the UK and Sardinian populations despite their entirely distinct demographies, and correlate well with the pattern of recombinations. Nevertheless, haplotypes extend across recombination hot spots indicative of selection of certain haplotypes. Subisolate aside, chromosome-region-specific differences in LD patterns appear to be more important than the differences in intermarker LD between distinct populations.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Linkage disequilibrium (LD), the non-random association of alleles at closely linked loci, is being used to infer the location of unknown disease genes by virtue of their correlated appearance with surrounding markers. Although this approach has been successful for identifying genes responsible for rare Mendelian disorders in isolated populations, there is considerable debate concerning its application to common multifactorial diseases (1–4).
Under the common disease–common variant hypothesis, the aetiologic polymorphisms involved in the genetic predisposition to multifactorial traits are ancient and LD around them is expected to be low, at levels similar to those observed around neutral, non-disease-predisposing alleles. Consequently, identification of disease associations would require an extremely dense map of markers. However, the exact number and nature of the markers and even the populations of choice to be employed in LD mapping studies of complex traits are still open issues (2,5–9). Indeed, this approach depends on the pattern of LD between markers in the genome, which is difficult to predict owing to the numerous forces that influence LD and complicate the relationship between LD and physical distance.
Evidence exists for both significant chromosome-region-specific variation as well as population-specific differences in LD (8,10–12). A survey of microsatellite markers mapping to chromosome Xq13 provided empirical evidence that small and demographically stable populations like the Saami exhibit higher degrees of LD than large, rapidly expanding populations (11). More recently, a study using dense maps of single nucleotide polymorphisms (SNPs) in large regions of the X chromosome, namely Xq25 and Xq28, found areas of extensive LD punctuated by regions of equilibrium (8). Importantly, the LD patterns of the regions were remarkably similar even in populations that had largely different demographic histories such as the Finns and the Sardinians and a mixed US–European sample [Centre d’Etude des Polymorphisms Human (CEPH) males]. Consistently, a survey of microsatellite markers mapping to 6.5 cM on chromosome 18q21 suggested that large genetic isolates like Finland and Sardinia might not prove appreciably more valuable for the LD mapping of common disease genes than admixed populations from the UK and USA (7).
However, these observations raised several questions. How indicative were the patterns of LD observed in these regions of what is present across the genome? Also, with respect to the extrapolation of LD to an entire population, is it possible that subisolates would display substantially different patterns of LD compared with those detected in the general population? To answer these questions, we used dense maps of markers from three distinct regions of two different chromosomes and compared the results between the Sardinian and UK data sets. The maps included the following. (i) Seven microsatellites, spanning 
9–11.5 Mb on Xq13, were chosen because their pair-wise LD had been analysed previously in four northern European populations including the Finns and the Saami (11). (ii) Twenty-one microsatellite markers that map to 5.4 Mb on chromosome Xp21–p11 were genotyped as part of a programme to identify a gene for type 1 diabetes mellitus (T1DM) in this region. (iii) Twenty-one microsatellite markers that map to a 9.452 Mb interval on chromosome 6p21.31 were chosen. To measure the LD in a subisolate, 73 unrelated males from the Sardinian village of Gavoi were analysed by genotyping the two regions of the X chromosome. We chose Gavoi because it represents an example of a small and slowly expanding population that has undergone, at its foundation and during its history, many bottlenecks (see Materials and Methods).
In agreement with what was previously reported for chromosome 18 and on the X chromosome (7,8), we did not observe striking differences in the mean LD values detected between the general Sardinian and UK populations in the three different chromosome regions examined. However, a primary effect of demography was indicated by analysis of the subisolate, which showed a dramatic increase of the average LD at levels similar to those previously observed in the Saami.
The contribution of other non-demographic forces in the generation and maintenance of the LD was evaluated by the analysis of the chromosome 6p21 region. We provide consistent evidence indicating that the pattern of LD observed in the human leukocyte antigen (HLA) region is the complex result of different forces, such as the tendency of the various subregions to recombine, and selection.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
LD on Xq13
We initially characterized in the Sardinian and UK populations the region on Xq13 that was analysed in different Scandinavian ethnic groups by Laan and Paabo (11). We considered sample sets from the UK and Sardinian general populations and a special Sardinian subisolate from a village (Gavoi) located in the most internal part of the island. Cross-population comparisons of these three new sample sets, each composed of 73 healthy males, with those previously analysed by Laan and Paabo (11) were made (Table 1). The UK and the general Sardinian samples showed that the majority of the loci pairs were in equilibrium, similarly to what was observed in the other European populations with the obvious exception of the Saami. Most strikingly, the results from the Sardinian subisolate are very similar to those reported for the Saami, both in terms of the number of marker-pairs in LD as well as the distances covered. In comparison with the general Sardinian samples, Gavoi showed more marker pairs in LD (19 of 21) versus (2 of 21), (respectively, 12 of 21 and 2 of 21 after correction); this is comparable to the value for the Saami (17 of 21 before correction). The differences between the Gavoi and the general Sardinian and UK samples were also underlined by the D' values stratified by the chromosome distances, with the subisolate consistently higher (Fig. 1). It should be noted that, although allelic variation at the various microsatellite markers was not restricted in the subisolate, the number of the observed haplotypes was substantially reduced (49 haplotypes) in comparison with the Sardinian and UK sample sets (72 and 70, respectively).
|
|
LD on Xp21–p11
Next, we wanted to replicate these observations in a different chromosome region. In order to fully use the Gavoi males data set while retaining the advantage of unequivocal haplotyping without parental data, we chose another region of the X chromosome, Xp21–p11. One possible reservation about the map used in Xq13 is that the marker density is rather low and the intervals are much longer than those used in practical LD searches. This raises the possibility that relevant differences over shorter distances between the large populations are masked. To provide an additional and presumably more applicable measure of LD, we covered the region with a much greater number (21 markers) and density (average interval of 260 kb) of markers.
We also found that in Xp21–p11, the subisolate showed much stronger degrees of LD than did the general Sardinian population, both in terms of the number of marker-pairs in LD and of the strength of LD evaluated by multiallelic D' (Fig. 2). The Gavoi samples had 65 of 210 marker-pairs in LD versus 18 of 210 in the general Sardinian population (respectively, 25 of 210 and 9 of 210 after correction). In these sets of 73 individuals, the UK sample showed similar degrees of LD to that in the general Sardinian data set (20 of 210 marker pairs in LD and 8 of 210 after correction).
|
We then analysed a much larger number of samples, 594 chromosomes in both the UK and the general Sardinian data sets, to explore in more detail the differences between the Sardinian and UK populations. Although a putative T1DM susceptibility locus mapped to this region, the pattern of LD in the T1DM samples was no different to that observed in the healthy individuals and in the unaffected chromosomes assembled from the T1DM families (unpublished data and Materials and Methods). Therefore, for this analysis we considered the total independent chromosomes available from the general UK and Sardinian sets, matched for sample number. In comparison with the UK, the Sardinians showed more marker-pairs in LD (51 of 210 versus 26 of 210; respectively, 29 of 210 and 13 of 210 after correction). Note 73 chromosomes in the Gavoi sample set provided more power to detect significant LD than
600 chromosomes from the Sardinian general population. The Sardinians also had some pairs that were in significant LD despite being separated by >500 kb, though few pairs in either population showed LD over an interval of >1 Mb. This trend is also reflected in the D' value in the two populations, with Sardinia somewhat higher only over intervals of <1 Mb (Fig. 3).
|
LD on 6p21.31
The HLA region provides a unique opportunity to study some of the other factors that are thought to influence LD. Specific regions that display relatively high levels of recombination have been mapped, so-called ‘hot spots’. Nevertheless, certain haplotypes extend across these hot spots, implicating the action of selection, presumably due to the resistance to infectious disease. Hence, analysis of LD in the HLA region in different populations might help in distinguishing the relative influences of population-specific and chromosome-specific effects, such as recombination patterns, on its generation and maintenance.
In Figure 4 we report LD between 21 microsatellite markers covering a region of 9.452 Mb on chromosome 6p21 for the Sardinian and UK data sets. The region examined encompasses the whole HLA, which accounts for 3.3 Mb of the total. Our samples of T1DM parents were enriched for extended diabetogenic haplotypes (DR3, in particular). Therefore, to provide LD data fully representative of the general populations, we established and analysed only the affected family-based control (AFBAC) haplotypes, 516 and 528 chromosomes in the Sardinian and UK populations, respectively (see Materials and Methods). The total number of pairs of markers showing significant LD was once again greater in the Sardinians than in the UK: 151 of 210 before correction, 104 of 210 after correction versus 106 of 210 before correction and 87 of 210 after correction, respectively.
|
There was a remarkable similarity of the LD patterns of the two populations, with the LD non-uniformly distributed, being stronger in the telomeric side, encompassing the HLA class III and class I regions, than in the centromeric part, including class II, of the map (Fig. 4). We can glean more information about LD by looking at the global D' matrix (lower part of Fig. 4). When we examine the pairs with higher, intermediate and lower D' values, a region of strong LD, telomeric to the marker D3A in the HLA class III subregion, becomes evident in both populations. This covers not just adjacent marker pairs, but also more distant pairs, including markers that are telomeric to the classically defined class I region. In contrast, the region centromeric of D3A showed only a few values of weak LD.
These results indicate that, although there is more LD in the Sardinians, there is also a remarkable similarity between the patterns of global LD observed in the two populations. For instance, contrast in Figure 5 the inter-marker LD detected on the different 6p21 subregions with those observed on chromosome Xp21–p11. It is evident in both populations that the region centromeric of D3A shows a somewhat intermediate degree of LD to that observed on chromosome Xp21–p11 and that evinced by the region telomeric to D3A. Since the genetic and demographic histories of Sardinia and the UK are vastly different, these similarities are unlikely to be accounted for by population-specific factors. Possible explanations could invoke the presence of shared selective sweeps, such as those related to resistance to infectious diseases, or structural factors such as a non-uniform distribution of chromosome regions that are more prone or less prone to recombination.
|
We directly explored the latter issue by contrasting the global patterns of LD between pairs of markers and the observed meiotic cross-overs. In the Sardinian sample set we observed 35 recombination events (in 257 pairs of siblings), 18 of which were unequivocally localized in the 4.001 Mb region centromeric of D3A where inter-marker LD was lower, and only 13 of which were in a 5.451 Mb interval telomeric of D3A, where LD was stronger. Consistent with the pattern found in the Sardinians, 59 recombinations (in 385 pairs of siblings) were counted in the UK data set, of which 35 were centromeric of D3A and 20 telomeric. The ratio of genetic to physical distance was 1.10 and 1.37 in the region centromeric of D3A and 0.73 and 0.42 telomeric of D3A in the UK and Sardinian populations, respectively. Contrast this with the genome-wide ratio of 0.89 (13).
To more precisely locate the sites of recombination, 105 SNPs, mostly in the class II genes themselves, in nine expressed genes located in the class II subregion were also considered in the Sardinian samples. This enhanced resolution increased the number of observable recombinations to 59 and showed that the distribution of the crossover events in the class II region was consistent with two of the three established hot spots, namely between HLA-DNA and RING3 and between DQB3 and DQB1 (Fig. 6). On our map the corresponding intervals are defined by DPB1 and DMA and by DOB and DQB1, respectively. We did not have enough resolution to localize the hot spot located within the second intron of the TAP2 gene. The distribution of the meiotic cross-overs in both the Sardinian and UK sample sets also suggests the presence of an additional hot spot, located between D6S439 and D6S1629 (overall six unequivocal recombination events in a 200 kb interval observed in each sample set) (Fig. 6).
|
0.40 shaded black, 0.4 > D' In agreement with what was noted by others (14,15), examining the pair-wise LD of the markers flanking the hot spots it was evident that LD is not always an accurate indicator for the fine definition of those domains which are highly prone to recombination. In some cases the expected correlation between high number of recombination events and the equilibrium between flanking markers could be seen. For instance, markers D6S439 and D6S1629 defining the putative new hot spot shows no LD (Pc = 1; D' = 0.14). In other cases, no clear correlation was found. For example, within the 125 kb interval between DPB1 and DMA, which contains the HLA-DNA-RING3 hot spot, we found 13 recombinations, but there is also a significant indication of moderate LD (Pc < 1 x 10–7, D' = 0.38). A similar state of affairs was also noted for the 148 kb interval between DOB and DQB1, where 10 recombinations were detected, but again there was also indication of moderate, albeit not significant after correction, LD (P = 0.026, Pc = 1, D' = 0.35).
Furthermore, it was also evident that even the general correlation between meiotic cross-overs and LD is not absolute. For instance, in the Sardinians we observed a significant degree of moderate LD between DQB1 and DPB1 (Pc < 1 x 10–7, D' = 0.27) and between DRB1 and DPB1 (Pc < 1 x 10–7, D' = 0.33) despite the fact that these markers are separated by a relatively large physical distance that is highly prone to recombination.
Stability of the microsatellites and their use in LD studies
Overall, the levels of LD observed in this study using microsatellite markers extended further than predicted by simulations based on SNPs (1). Could the purported instability of microsatellites explain the relatively high levels of LD among the microsatellite markers? That is, could the observed LD be just a consequence of the fact that novel microsatellite alleles, and hence new haplotypes, are continually forming? A simple way to address this issue was to compare the microsatellite composition of extended haplotypes in distantly related populations.
We determined the microsatellite allelic composition of all the DR3-B18 haplotypes observed in the Sardinian and UK samples. DR3-B18 is a well-established example of an extended ancestral haplotype. As such, the interval between the DRB1 and B loci offers the advantage that it is less prone to other sources of variability such as recombination and was therefore suitable to test for microsatellite stability. We determined the alleles of nine microsatellite loci included within the interval between DRB1 and B in 300 independent DR3-B18 haplotypes, 273 and 27 in Sardinia and UK, respectively (Fig. 7). A conserved haplotype was observed in a region spanning 1.25 Mb. In this interval the two populations showed a common pattern of alleles with a remarkable degree of identity at nine distinct microsatellite loci. Note that the Sardinian and the northern European DR3-B18 haplotypes are divergent at the DPB1 locus, allowing us to exclude the possibility that these results are due to recent admixture between the different populations (F. Cucca and J.A. Todd, unpublished data). These results extend previous findings (16) and indicate that the microsatellite alleles are marking ancient haplotypes that predate the population divergence in Europe. They also demonstrate that the relatively high levels of LD and the similar patterns of LD in different populations are not exclusively related to recent microsatellite mutation.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
We have examined different populations and different chromosome regions to dissect the contribution of the various forces that are thought to influence LD in humans. There was a striking increase of the LD observed in the two regions of the X chromosome analysed in a Sardinian subisolate (Gavoi) compared with the general Sardinian sample. The small population of the village of Gavoi over the course of its history appears as the critical factor that created considerable drift, accounting for the impressive LD observed in this subisolate. The size of the general Sardinian population is unlikely to have generated comparable levels of LD by drift. These results are clear evidence that small and isolated populations exhibit higher levels of LD than the larger populations from which they are derived. Importantly, we can exclude that the differences between the general Sardinian sample and the subisolate are related to other factors, such as differences in the genetic structure and exposure to different selective agents. These forces are expected to be very similar in the two groups from the same population (17) but might have affected, even in a subtle way, the results previously obtained in the comparison of distinct populations (11).
We have further delineated the influence of demography on LD by contrasting the differences observed between the UK and Sardinian populations. Overall LD appears moderately but consistently stronger in the general Sardinian population than in the UK data set. The critical factor explaining the subtle differences between the UK and the general Sardinian population is most likely the different demographic growth of the two groups in the last few thousand years, the Sardinians being relatively stable and the UK rapidly expanding.
Subisolate aside, cross-comparison of the results obtained in the UK and general Sardinian populations on different chromosome regions indicates that overall differences in the LD patterns between distinct portions of the human genome might be even stronger than the corresponding differences between LD in the populations (Fig. 7). For instance, in both the Sardinian and UK samples, LD was not uniformly distributed across 6p21, being consistently stronger in the telomeric side than in the centromeric part of the map. We provide empirical evidence that there is a good overall correlation between these LD patterns and the general distribution of meiotic crossovers detected in the region. However, as expected considering the cumulative and multigenerational nature of the forces that influence LD, this correlation is not absolute. For example, in the Sardinian data set we observed a significant, albeit moderate, LD between DQB1 and DPB1 and between DRB1 and DPB1 despite the fact that these markers are separated by relatively large distances that are highly prone to recombination. The 750 generations since the ancestral Sardinian population settled on the island should have been sufficient to create a complete equilibrium between these loci. Similar results were also found in the UK data set (F. Cucca and J.A. Todd, unpublished data) and in studies of the Cayapa and of CEPH families, while weak LD or equilibrium was observed in some European populations (18). The most obvious interpretation for these data is that specific combinations of HLA class II alleles were kept in phase by selective sweeps that operate or operated until recently on these loci in some populations. Resistance to infectious disease is likely to be a major factor of natural selection.
Overall, the various sets of data from different populations and chromosome regions confirm the complex nature of LD and allow a better understanding of the nature of the main forces that contribute to its appearance and maintain it. But what about the practical implications of these results for LD mapping projects of common disease genes? One of the most striking results of this study is the strong degree of background LD found in the Sardinian subisolate. This suggests that these types of subisolate might represent ideal populations for the initial detection of polygenes involved in the predisposition to complex traits. In addition to the stronger background LD, there are other potential advantages in the choice of these subisolates (2,9). As a result of a reduced biochemical complexity and of a smaller number of disease alleles involved, some of the loci are expected to have stronger genetic effects on the trait under analysis. Furthermore, small isolates might contain mutations that are rare or non-existent in the larger population, thus allowing the identification of loci that would have been missed in the larger population.
One drawback in using such small populations for the LD analysis of common disease is that, except for the most common diseases, there will be few cases. Our recommendation is that such populations could be ideal in the search for markers in LD with quantitative traits underlying common disease endpoints. In quantitative trait loci (QTLs) analyses all phenotyped individuals are included in the study. Once LD is obtained and replicated, the fine mapping of the region in LD could be carried out in the general Sardinian population.
Finally, although the Sardinian subisolate shows impressive LD over long distances in Xq13, this, although still remarkable in comparison with the other populations, was less marked in Xp21–p11. This suggests that the prior knowledge of the LD patterns observed in chromosome regions of interest is a valuable step in LD mapping projects, even in ‘special populations’.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
For the analysis of the two regions of the X chromosome, we used three data sets formed respectively of 73 healthy, unrelated male blood donors from the UK; 73 healthy unrelated male blood donors from the blood transfusion centre of the Broztu Hospital in Cagliari, Sardinia; and 73 healthy males from the village of Gavoi, which is located in the most internal and isolated part of Sardinia. All the individuals from Gavoi were unrelated at least up to second degree cousins. Gavoi was founded 800–1000 years (40–50 generations) ago. Census records indicate that its population in 1650 was 1700 and it is now 3030. Over this period, the census size in the village ranged from a low of 495 in 1688 to a maximum of 3701 in 1959. These demographic data derive from archive records kindly furnished by the registrar office of the community of Gavoi. Contrast this with demographic data from the general Sardinian population. Direct evidence indicates that humans were already present in Sardinia in the pre-Neolithic age,
750 generations ago [15 590 ± 140 before present (BP)] (19). Around 3500 years BP the local population was of a substantial size (300 000 inhabitants), and remained constant until 400 years ago. In 1627 AD a fiscal survey estimated the population to be 297 424. Later, the plague of 1650–1652 AD and the great famine of 1680–1681 AD are thought to have taken 80 000 lives. By 1728 AD the population had recovered to an estimated 310 000 and it is now 1 500 000 individuals (20). In order to increase the power for the cross-population comparison of Xp21–p11, we enlarged the number of individuals in the UK and general Sardinian data sets by genotyping: respectively, 154 and 105 additional non-affected male blood donors and 90 and 257 parents of T1DM patients. Although a putative type 1 susceptibility locus mapped to this region of the X chromosome (21), the pattern of LD in the chromosomes transmitted to the patients was no different to that observed in the controls and pseudo-controls (i.e. the non-affected chromosomes assembled from the T1DM families defined as follows: the chromosomes not transmitted from the unaffected mothers to the affected children and the chromosomes of the unaffected fathers having only affected male children). Therefore, for the analysis of the general UK and Sardinian population we considered the data available matched in the two sample sets for the same number of chromosomes (594 chromosomes). Because unequivocal phase attribution of the markers in the HLA region requires family data, only the T1DM parental data were considered. However, as the major T1DM locus IDDM1 maps to this region, our samples of T1DM parents were enriched for extended diabetogenic haplotypes, particularly DR3. To avoid a possible bias, we established and analysed only the AFBAC haplotypes, as described by Thomson (22) for single alleles. We selected a similar number of control haplotypes in the Sardinians (516 haplotypes) and the UK (528 haplotypes) data sets.
The markers used in Xq13 have been reported previously (11). The primer sequences for markers used in Xp21–p11 (sWXD2449, sWXD2450, sWXD2451, CYBB, DXS8090, DXS1068, DXS8014, DXS8113, DXS8025, DXS1058, DXS977, DXS556, DXS8015, DXS8042, DXS1368, DXS8012, DXS1201, DXS8085, DXS228, DXS7 and MAOB) were obtained from the Genome Database (GDB, http://www.gdb.org ). The HLA region was genotyped for 21 microsatellite markers. The primer sequences for D6S291, D6S439, D6S1629, D6S1560, D6S1568, D6S2445, D6S2444, D6S273, C1-2-A, D6S265, D6S258, D6S1683 and D6S306 were obtained from Foissac and Cambon-Thomsen (23). Sequences for TNFa, TNFc, TNFd and TNFe were obtained from Udalova et al. (24). Sequences for D3A, 82-1 and 82-2/9N-1 were established by Hsieh et al. (25). The primer sequence for D6S2223 was obtained from the GDB (http://www.gdb.org ).
The expressed HLA genes considered in this study were typed as follows. The polymorphic second exons of the HLA-DRB1, -DQB1 and -DPB1 genes were amplified and the amplified products were dot-blot analysed using primers and sequence specific oligonucleotide (SSO) probes (previously described) in order to directly genotype, respectively, 51, 25 and 18 SNPs within these genes (26–28). LMP2 (1 SNP), LMP7 (1 SNP), DMA (4 SNPs) and DMB (3 SNPs) polymorphisms were typed using primers and conditions previously described (29,30). DOB (1 SNP) was typed with amplification of the fourth exon and subsequent dot-blot analysis of the amplified products using 5'-GTGTCTAGTACAGATTCTG-3' and 5'-CACTCCTCACAGGCTCAT-3' as ampli-primers and 5'-GTGGGAATCATCATCCAG-3' and 5'-GTGGGAATCGTCATCCAG-3' as SSO probes. Finally, the tapasin gene (1 SNP) was typed with amplification of the fourth exon and subsequent dot-blot analysis of the amplified products using 5'-AAATGGGACCTTCTGGCTGC-3' and 5'-AAGCTCCAGGGTGACCTGTC-3' as ampli-primers and 5'-GGCTGCCTAGAGTTCAACCC-3' and 5'-GGCTGCCTACAGTTCAACCC-3' as SSO probes (J. Copeman, personal communication). B18 alleles were typed using primers and polymerase chain reaction (PCR) mix previously reported (31). Amplification conditions were established on an MJ Research PTC 100 machine (MJ Research, Watertown, MA). The cycling parameters were as follows: 1 min denaturation; 7 touch-down cycles [25 s at 96°C, 45 s annealing from 68 to 65°C (–0.5°C/cycle), 30 s at 72°C]; 20 cycles with annealing temperature at 65°C (25 s at 96°C, 45 s at 65°C, 30 s at 72°C); 5 min at 72°C.
Microsatellite genotyping was performed by separating fluorescently tagged PCR products on a polyacrylamide gel using ABI 377 DNA sequencers and the GeneScan 3.1 and Genotyper 2.0 software (Perkin-Elmer Applied Biosystems, Warrington, UK). Markers showing failures during the first attempt at amplification were PCR amplified and genotyped a second time. PCR product standards, consisting of the amplification product of two different standards for each marker, were loaded on each gel for correct allele assignment. The two standards consisted of the CEPH individual no. 1347.02 and of a pool of DNAs. The alleles at each microsatellite were given a numerical value (1, 2, 3 etc.) starting with the allele with the lowest number of base pairs. The physical map of the Xp21–p11 region with relative order, map position and distances between markers were obtained from a yeast artificial chromosome-sequence tagged site (YAC-STS) map (32,33). The physical map of the HLA region with relative order, map position and distances between markers were obtained from the Sanger Centre (http://www.sanger.ac.uk/HGP/Chr6/MHC.shtml ; A. Mungall, personal communication).
Recombinations were established using the sibmap function of the GAS software (version 2.0; © Alan Young, Oxford University, 1993–1995) in 209 Sardinian T1DM families, in which 257 pairs of siblings were collected and genotyped. The observed recombinations were verified by re-checking the molecular typing of the relevant markers.
The total normalized disequilibrium (total D') between DRB1-DQB1 and the various marker loci was calculated using a multiallelic extension of Lewontin’s standardized measure of disequilibrium (34,35) and ranges from 0 to 1, with 0 reflecting perfect independence between alleles at the two loci compared and 1 reflecting complete LD. The respective P values were calculated using the Markov-chain method described by Guo and Thompson (36) (available at http://anthropologie.unige.ch/arlequin ). In all cases, 100 000 tables were explored. This P value (or table ratio) is reported with each pair-wise comparison and was considered significant each time the ratio was < 0.05. In keeping with the description of LD reported in previous studies by Laan and Paabo (11), we present the non-corrected P values only as a useful measure for comparison with previously reported sample populations as well as between our own. Values reported as ‘corrected’ have been subjected to the step-down Holm–Sidak procedure described in Lautenberger et al. (37) and represent true measures of significance. Each P value is corrected by the formula Pc = 1 – [1 – P (uncorrected)]n where n is the number of uncorrected P values for that locus less than or equal to the value to be corrected. For example, if the eighth most significant P value was 0.01 before correction, the corrected value would be Pc = 1 – (1 – 0.01)8 = 0.0773.
This method was used because it has several desirable features compared with the Bonferroni correction. The fact that it is sensitive to the number of scores allows the treatment of data sets of different sizes, whereas Bonferroni only treats the number of comparisons (in this case, number of loci analysed), not the number of individuals sampled. Also, the Holm–Sidak procedure maps all P values onto the interval [0,1], whereas the Bonferroni often gives results of >1, which are not interpretable as a probability.
Parental haplotypes were reconstructed using software available at ftp://ftp-gene.cimr.cam.ac.uk/pub/software/ . Briefly, haplotypes were constructed only when phase could be unambigously determined, either using the genotype information of the offspring, or with the help of a tightly linked adjacent marker. Haplotypes with low informativity were discarded.
|
ACKNOWLEDGEMENTS |
|---|
We wish to thank Antonio Cao, Mario Silvetti, Stefano De Virgiliis, Efisio Angius, Mathias Herr, Gualtiero Colombo and PierGiorgio Satta for help, advice and support, Cesare Zavattari for writing a program that allows assignment of the allele sizes into their appropriate allele bins, James Copeman for information about tapasin, Margi Chessa, Paola Frongia and Rossella Ricciardi for help in collecting the DNA samples from Sardinian T1DM families and Andrew Mungall (Sanger Center) for the establishment of the 6p21 map. In particular, we are indebted to the population from Gavoi for their enthusiastic participation in this study and to Giovanni Maoddi-Costeri of the registrar office of Gavoi for his help in collecting unrelated individuals and for valuable demographic data on Gavoi. We are also grateful to the Italian Telethon, the Regione Autonoma Sardegna (L.R.11, 30-4-90) and the Wellcome Trust for financial support. F.C. and J.A.T. are recipients of a Wellcome Trust Biomedical Research Collaboration grant and J.A.T. was a Wellcome Trust Principal Research Fellow.
|
FOOTNOTES |
|---|
+ To whom correspondence should be addressed. Tel: +39 070 6095681; Fax: +39 070 6095558; Email: fcucca@mcweb.unica.it
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
1 Kruglyak, L. (1999) Prospects for whole-genome linkage disequilibrium mapping of common disease genes. Nature Genet., 22, 139–144.[Web of Science][Medline]
2 Terwilliger, J.D. and Weiss, K.M. (1998) Linkage disequilibrium mapping of complex disease: fantasy or reality? Curr. Opin. Biotechnol., 9, 578–594.[Web of Science][Medline]
3 Collins, A., Lonjou, C. and Morton, N.E. (1999) Genetic epidemiology of single-nucleotide polymorphisms. Proc. Natl Acad. Sci. USA, 96, 15173–15177.
4 Ott, J. (2000) Predicting the range of linkage disequilibrium. Proc. Natl Acad. Sci. USA, 97, 2–3.
5 Lonjou, C., Collins, A. and Morton, N.E. (1999) Allelic association between marker loci. Proc. Natl Acad. Sci. USA, 96, 1621–1626.
6 Kruglyak, L. (1999) Genetic isolates: separate but equal? Proc. Natl Acad. Sci. USA, 96, 1170–1172.
7 Eaves, I.A., Merriman, T.R., Barber, R.A., Nutland, S., Tuomilehto-Wolf, E., Tuomilehto, J., Cucca, F. and Todd, J.A. (2000) The genetically isolated populations of Finland and Sardinia may not be a panacea for linkage disequilibrium mapping of common disease genes. Nature Genet., 25, 320–323.[Web of Science][Medline]
8 Taillon-Miller, P., Bauer-Sardina, I., Saccone, N.L., Putzel, J., Laitinen, T., Cao, A., Kere, J., Pilia, G., Rice, J.P. and Kwok, P.Y. (2000) Juxtaposed regions of extensive and minimal linkage disequilibrium in human Xq25 and Xq28. Nature Genet., 25, 324–328.[Web of Science][Medline]
9 Wright, A.F., Carothers, A.D. and Pirastu, M. (1999) Population choice in mapping genes for complex diseases. Nature Genet., 23, 397–404.[Web of Science][Medline]
10 Huttley, G.A., Smith, M.W., Carrington, M. and O’Brien, S.J. (1999) A scan for linkage disequilibrium across the human genome. Genetics, 152, 1711–1722.
11 Laan, M. and Paabo, S. (1997) Demographic history and linkage disequilibrium in human populations. Nature Genet., 17, 435–438.[Web of Science][Medline]
12 Kidd, J.R., Pakstis, A.J., Zhao, H., Lu, R.B., Okonofua, F.E., Odunsi, A., Grigorenko, E., Tamir, B.B., Friedlaender, J., Schulz, L.O. et al. (2000) Haplotypes and linkage disequilibrium at the phenylalanine hydroxylase locus, PAH, in a global representation of populations. Am. J. Hum. Genet., 66, 1882–1899.[Web of Science][Medline]
13 Gyapay, G., Morissette, J., Vignal, A., Dib, C., Fizames, C., Millasseau, P., Marc, S., Bernardi, G., Lathrop, M. and Weissenbach, J. (1994) The 1993–94 Genethon human genetic linkage map. Nature Genet., 7, 246–339.[Web of Science][Medline]
14 Carrington, M. (1999) Recombination within the human MHC. Immunol. Rev., 167, 245–256.[Web of Science][Medline]
15 Jeffreys, A.J., Ritchie, A. and Neumann, R. (2000) High resolution analysis of haplotype diversity and meiotic crossover in the human TAP2 recombination hotspot. Hum. Mol. Genet., 9, 725–733.
16 Lampis, R., Morelli, L., Congia, M., Macis, M.D., Mulargia, A., Loddo, M., De Virgiliis, S., Marrosu, M., Todd, J. and Cucca, F. (2000) The intraregional distribution of HLA class II haplotypes indicates the suitability of the Sardinian population for case–control association studies in complex diseases. Hum. Mol. Genet., 9, 2959–2965.
17 Crouau-Roy, B., Bouzekri, N., Carcassi, C., Clayton, J., Contu, L. and Cambon-Thomsen, A. (1996) Strong association between microsatellites and an HLA-B, DR haplotype (B18-DR3): implication for microsatellite evolution. Immunogenetics, 43, 255–260.[Web of Science][Medline]
18 Sanchez-Mazas, A., Djoulah, S., Busson, M., Le Monnier de Gouville, I., Poirier, J.C., Dehay, C., Charron, D., Excoffier, L., Schneider, S., Langaney, A. et al. (2000) A linkage disequilibrium map of the MHC region based on the analysis of 14 loci haplotypes in 50 French families. Eur. J. Hum. Genet., 8, 33–41.[Web of Science][Medline]
19 Sondaar, P.Y., Elburg, R., Klein Hofmeijer, G., Martini, F., Sanges, M., Spaan, A. and De Visser, J.A. (1995) The human colonization of Sardinia: a late-Pleistocene human fossil from Corbeddu Cave. C. R. Acad. Sci. Paris, 320, 145–150.
20 Gatti, A. (1995) Outlines on the Peopling and Demographic Evolution in Sardinia (Italy): History, Effects and Perspectives. University of Cagliari, Cagliari, Italy, pp. 7–12.
21 Cucca, F., Goy, J.V., Kawaguchi, Y., Esposito, L., Merriman, M.E., Wilson, A.J., Cordell, H.J., Bain, S.C. and Todd, J.A. (1998) A male-female bias in type 1 diabetes and linkage to chromosome Xp in MHC HLA-DR3-positive patients. Nature Genet., 19, 301–302.[Web of Science][Medline]
22 Thomson, G. (1995) Mapping disease genes: family-based association studies. Am. J. Hum. Genet., 57, 487–498.[Web of Science][Medline]
23 Foissac, A. and Cambon-Thomsen, A. (1998) Microsatellites in the HLA region: 1998 update. Tissue Antigens, 52, 318–352.[Web of Science][Medline]
24 Udalova, I.A., Nedospasov, S.A., Webb, G.C., Chaplin, D.D. and Turetskaya, R.L. (1993) Highly informative typing of the human TNF locus using six adjacent polymorphic markers. Genomics, 16, 180–186.[Web of Science][Medline]
25 Hsieh, S.L., March, R.E., Khanna, A., Cross, S.J. and Campbell, R.D. (1997) Mapping of 10 novel microsatellites in the MHC class III region: application to the study of autoimmune disease. J. Rheumatol., 24, 220–222.[Web of Science][Medline]
26 Horn, G.T., Bugawan, T.L., Long, C.M. and Erlich, H.A. (1988) Allelic sequence variation of the HLA-DQ loci: relationship to serology and to insulin-dependent diabetes susceptibility. Proc. Natl Acad. Sci. USA, 85, 6012–6016.
27 Wordsworth, B.P., Allsopp, C.E., Young, R.P. and Bell, J.I. (1990) HLA-DR typing using DNA amplification by the polymerase chain reaction and sequential hybridization to sequence-specific oligonucleotide probes. Immunogenetics, 32, 413–418.[Web of Science][Medline]
28 Bugawan, T.L., Begovich, A.B. and Erlich, H.A. (1991) Rapid HLA-DPB typing using enzymatically amplified DNA and nonradioactive sequence-specific oligonucleotide probes. Immunogenetics, 34, 413.[Web of Science][Medline]
29 Deng, G.Y., Muir, A., Maclaren, N.K. and She, J.X. (1995) Association of LMP2 and LMP7 genes within the major histocompatibility complex with insulin-dependent diabetes mellitus: population and family studies. Am. J. Hum. Genet., 56, 528–534.[Web of Science][Medline]
30 Djilali-Saiah, I., Benini, V., Schmitz, J., Timsit, J., Assan, R., Boitard, C., Bach, J.F. and Caillat-Zucman, S. (1996) Absence of primary association between DM gene polymorphism and insulin-dependent diabetes mellitus or celiac disease. Hum. Immunol., 49, 22–27.[Medline]
31 Tonks, S., Marsh, S.G., Bunce, M. and Bodmer, J.G. (1999) Molecular typing for HLA class I using ARMS-PCR: further developments following the 12th International Histocompatibility Workshop. Tissue Antigens, 53, 175–183.[Web of Science][Medline]
32 Nagaraja, R., MacMillan, S., Kere, J., Jones, C., Griffin, S., Schmatz, M., Terrell, J., Shomaker, M., Jermak, C., Hott, C. et al. (1997) X chromosome map at 75-kb STS resolution, revealing extremes of recombination and GC content. Genome Res., 7, 210–222.
33 Nagaraja, R., Jermak, C., Trusgnich, M., Yoon, J., MacMillan, S., McCauley, M.B., Brownstein, B. and Schlessinger, D. (1998) YAC/STS map of 15Mb of Xp21.3-p11.3, at 100kb resolution, with refined comparisons of genetic distances and DMD structure. Gene, 215, 259–267.[Web of Science][Medline]
34 Hedrick, P.W. (1987) Gametic disequilibrium measures: proceed with caution. Genetics, 117, 331–341.
35 Lewontin, R.C. (1988) On measures of gametic disequilibrium. Genetics, 120, 849–852.
36 Guo, S.W. and Thompson, E.A. (1992) Performing the exact test of Hardy–Weinberg for multiple alleles. Biometrics, 48, 361–372.[Web of Science][Medline]
37 Lautenberger, J.A., Stephens, J.C., O’Brien, S.J. and Smith, M.W. (2000) Significant admixture linkage disequilibrium across 30 cM around the FY locus in African Americans. Am. J. Hum. Genet., 66, 969–978.[Web of Science][Medline]
38 Martin, M., Mann, D. and Carrington, M. (1995) Recombination rates across the HLA complex: use of microsatellites as a rapid screen for recombinant chromosomes. Hum. Mol. Genet., 4, 423–428.
39 Cullen, M., Noble, J., Erlich, H., Thorpe, K., Beck, S., Klitz, W., Trowsdale, J. and Carrington, M. (1997) Characterization of recombination in the HLA Class II region. Am. J. Hum. Genet., 60, 397–407.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S. Cronin, S. Berger, J. Ding, J. C Schymick, N. Washecka, D. G. Hernandez, M. J. Greenway, D. G. Bradley, B. J. Traynor, and O. Hardiman A genome-wide association study of sporadic ALS in a homogenous Irish population Hum. Mol. Genet., March 1, 2008; 17(5): 768 - 774. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. W Lawrence, D. M Evans, and L. R Cardon Prospects and pitfalls in whole genome association studies Phil Trans R Soc B, August 29, 2005; 360(1460): 1589 - 1595. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. A. Rana, N. D. Ebenezer, A. R. Webster, A. R. Linares, D. B. Whitehouse, S. Povey, and A. J. Hardcastle Recombination hotspots and block structure of linkage disequilibrium in the human genome exemplified by detailed analysis of PGM1 on 1p31 Hum. Mol. Genet., December 15, 2004; 13(24): 3089 - 3102. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Tenesa, A.F. Wright, S.A. Knott, A.D. Carothers, C. Hayward, A. Angius, I. Persico, G. Maestrale, N.D. Hastie, M. Pirastu, et al. Extent of linkage disequilibrium in a Sardinian sub-isolate: sampling and methodological considerations Hum. Mol. Genet., January 1, 2004; 13(1): 25 - 33. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Rahman, A. Jones, J. Curtis, S. Bartlett, L. Peddle, B. A. Fernandez, and N. B. Freimer The Newfoundland population: a unique resource for genetic investigation of complex diseases Hum. Mol. Genet., October 15, 2003; 12(90002): R167 - 172. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Heutink and B. A. Oostra Gene finding in genetically isolated populations Hum. Mol. Genet., October 1, 2002; 11(20): 2507 - 2515. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Hall, E. M. Wijsman, J. L. Roos, J. A. Gogos, and M. Karayiorgou Extended Intermarker Linkage Disequilibrium in the Afrikaners Genome Res., June 1, 2002; 12(6): 956 - 961. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. J. Palmer, E. S. Silverman, S. T. Weiss, and J. M. Drazen Pharmacogenetics of Asthma Am. J. Respir. Crit. Care Med., April 1, 2002; 165(7): 861 - 866. [Full Text] [PDF]
|
|
|
|
|
|
L. Tiret, O. Poirier, V. Nicaud, S. Barbaux, S.-M. Herrmann, C. Perret, S. Raoux, C. Francomme, G. Lebard, D. Tregouet, et al. Heterogeneity of linkage disequilibrium in human genes has implications for association studies of common diseases Hum. Mol. Genet., February 1, 2002; 11(4): 419 - 429. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. G. Marrosu, R. Murru, M. R. Murru, G. Costa, P. Zavattari, M. Whalen, E. Cocco, C. Mancosu, L. Schirru, E. Solla, et al. Dissection of the HLA association with multiple sclerosis in the founder isolated population of Sardinia Hum. Mol. Genet., December 1, 2001; 10(25): 2907 - 2916. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Paunio, J. Ekelund, T. Varilo, A. Parker, I. Hovatta, J. A. Turunen, K. Rinard, A. Foti, J. D. Terwilliger, H. Juvonen, et al. Genome-wide scan in a nationwide study sample of schizophrenia families in Finland reveals susceptibility loci on chromosomes 2q and 5q Hum. Mol. Genet., December 1, 2001; 10(26): 3037 - 3048. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L.B. Jorde, W.S. Watkins, and M.J. Bamshad Population genomics: a bridge from evolutionary history to genetic medicine Hum. Mol. Genet., October 1, 2001; 10(20): 2199 - 2207. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Zavattari, R. Lampis, C. Motzo, M. Loddo, A. Mulargia, M. Whalen, M. Maioli, E. Angius, J. A. Todd, and F. Cucca Conditional linkage disequilibrium analysis of a complex disease superlocus, IDDM1 in the HLA region, reveals the presence of independent modifying gene effects influencing the type 1 diabetes risk encoded by the major HLA-DQB1, -DRB1 disease loci Hum. Mol. Genet., April 1, 2001; 10(8): 881 - 889. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. K. Service, R. A. Ophoff, and N. B. Freimer The genome-wide distribution of background linkage disequilibrium in a population isolate Hum. Mol. Genet., March 1, 2001; 10(5): 545 - 551. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Lampis, L. Morelli, M. Congia, M. D. Macis, A. Mulargia, M. Loddo, S. D. Virgiliis, M. G. Marrosu, J. A. Todd, and F. Cucca The inter-regional distribution of HLA class II haplotypes indicates the suitability of the Sardinian population for case-control association studies in complex diseases Hum. Mol. Genet., December 1, 2000; 9(20): 2959 - 2965. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||