3.4 Conserved Extended Haplotypes in the Singapore Chinese Population With the earlier lower-resolution SNP map, four conserved extended haplotypes CEHs stretching across the MHC was see
Trang 13.4 Conserved Extended Haplotypes in the Singapore Chinese Population
With the earlier lower-resolution SNP map, four conserved extended haplotypes (CEHs) stretching across the MHC was seen in the Singaporean Chinese population (Section 3.1.3) These were: A*0207-C*0102-B*4601-DRB1*0901 (referred to as A2-B46-DR9), A*0203-C*0702-B*3802-DRB1*1602 (A2-B38-DR16), A*1101-C*0801-B*1502-DRB1*1202 (A11-B15-DR12) and A*3303-C*0302-B*5801-DRB1*0301 (A33-B58-DR3) The analysis of these CEHs is expanded in this section, making use of the higher resolution SNP map to describe these haplotypes in greater detail The 29 HLA-homozygous samples genotyped were also included in this analysis, providing unambiguous detail to the extent and conservation of these haplotypes
3.4.1 High-Resolution Single-SNP Homozygosity Plots
Homozygosity at a polymorphic marker can be thought of as the probability that two randomly selected haplotypes are identical at that marker, and is an indication of the conservation and linkage disequilibrium of underlying haplotypes (Sabatti and Risch 2002) To see if the MHC conserved haplotypes are also visible in this high-resolution SNP map, the single-SNP homozygosity plots were created for the 4 CEHs from position 29.5Mb to 33.5Mb of the map (Figure 3.25) The homozygosity plot for the A*1101-C*0702-B*4001-DRB1*0901 (A11-B40-DR9) haplotype, which is present
at high frequency in the population but was shown in the earlier section not to be conserved, was also included for comparison The homozygosity plots were
Trang 2The plots show large contrast between the 4 CEHs and the non-conserved DR9 haplotype For each CEH, average homozygosity is high in the region between the HLA-A and HLA-DRB1 loci, averaging at least 0.96, before decaying noticeably
A11-B40-at the telomeric and centromeric ends By contrast the average homozygosity for the A11-B40-DR9 haplotype was only 0.78, with heterogeneity in the genotypes seen across the entire 4Mb
When compared with the plots generated earlier using the sparser SNP map, some additional variation within each CEH is seen At first glance, the A11-B15-DR12 haplotypes show heterogeneity at 2 regions where homozygosity values dip below 0.9: a 500kb block centromeric of the HLA-A locus and another telomeric to the HLA-DRB1 locus However, the variation seen here is present in only one out of a possible 9 chromosomes In other CEHs, variation is seen at sporadic SNPs at different locations, possibly due to recurrent mutations or gene conversion events that had accumulated in the evolutionary history of the conserved haplotype The SNPs at positions 31,838,993 (dbSNP id: rs707937) and 32,447,625 (rs2050189) shows variability in at least 2 CEHs, with homozygosity values dipping in each haplotype The former falls within an intron of the MSH5 gene, while the latter lies in the 5´ UTR of the C6orf10 gene and is less than 300bp away from an Alu transposable-element The consistent variability at these loci in multiple CEHs could indicate fragile segments of the genome prone to repeat mutation, double-crossover or gene conversion events
Trang 3HLA-A HLA-C HLA-B HLA-DRB1
Trang 4The homozygosity value drops to 0.54 at position 30,683,582 (dbSNP id: rs1264420)
in A2-B46-DR9 haplotypes, with 65% of the haplotypes carrying the allele ‘G’ and the rest carrying the allele ‘A’ This SNP lies within an intron of the PPP1R10 gene, bordering an AluJb transposable element In A2-B38-DR16 haplotypes, besides positions 31,838,993 and 32,447,625 described above, low homozygosity is also seen
at positions 30,230,602, 31,199,971 and 31,569,068, but the variants at these sites are only seen in one chromosome
To briefly summarise the SNP homozygosity plots, very high homozygosity values are seen across each of the CEHs, but several sites of variation exists As the haplotypes used to construct these plots are from a mixture of HLA homozygous samples as well as phase-inferred unrelated individuals, we cannot rule out that some
of these variations within CEHs – especially those seen in only one chromosome – are due to incorrect phase calls The variations seen in multiple chromosomes within CEHs are however likely to represent recurrent mutation, double-crossover or gene conversion events that had accumulated within these haplotypes
3.4.2 Haplotypes of HLA Homozygous Samples
The inclusion of HLA homozygous individuals in the genotyped samples provides robustness to the description of the CEHs in the population Three of the four Chinese CEHs are represented in these homozygous samples; there are eight A2-B46-DR9-homozygous and five A33-B58-DR3-homozygous samples included in the set There
is also a single A11-B15-DR12 homozygous individual
Trang 5The A11-B15-DR12-homozygous individual is completely homozygous from positions 29,637,733 to 33,648,670, except for 2 SNP positions: 31,838,993 and 32,447,625 These 2 SNP positions correspond to the outliers in the homozygosity plots discussed in the above section, and confirm that these were not artefacts from the in-silico haplotype phasing, reiterating the value that homozygous samples provide
With 5 unrelated individuals homozygous for the A33-B58-DR3 haplotype, or 10 independent chromosomes, it is possible to examine the CEH unambiguously without phase-reconstructed haplotypes The SNP haplotype alignment of these 10 chromosomes is illustrated in Figure 3.26 All 10 chromosomes are almost identical across the length of the entire MHC The similarity at the centromeric end starts to fall apart at position 33,011,615, before the location of HLA-DMB, and very close to
a recombination hotspot (33,008,000 to 33,0010,000) identified in the HapMap project (International HapMap Consortium 2005) At the telomeric end, a single chromosome breaks at position 29,591,947; again this location is close to a HapMap recombination hotspot at position 29,594,500 Based on 9 of 10 chromosomes, the conservation of the A33-B58-DR3 haplotype stretches from RFP at the 29Mb mark to 33Mb at the very least, a full 4Mb of the MHC In the chromosome 6p LD map (Section 3.1.2), a long segment of strong LD is seen telomeric to the MHC and is therefore likely that the A33-B58-DR3 CEH is conserved way past the telomeric boundary of this SNP map
Trang 7The full-length SNP haplotypes for the 8 A2-B46-DR9 haplotypes were also aligned and analysed in a similar manner (Figure 3.27) The 16 chromosomes show very high similarity from positions 29,830,949 to 32,847,866 of the chromosome 6p, beyond which no obvious consensus haplotype is seen This puts the telomeric break between MOG and HLA-F, and the centromeric break after the HLA-DQB2 locus Two other chromosomes break at position 29,791,787, which coincides with the breakpoint seen
in HLA-A02 alleles discussed in the previous section
Unlike the A33-B58-DR3 haplotypes, there are recurrent heterozygous sites in the A2-B46-DR9-homozygous samples Two of these sites coincide with the ones identified earlier in the homozygosity plots; 4 out of the 8 samples are heterozygous
at position 30,683,582 (rs1264420), and 3 out of the 8 samples are heterozygous at position 32,447,625 (rs2050189), reconfirming the observation in the homozygosity plots The SNP at position 30,473,719 (dbSNP id: rs3130118) is seen to be heterozygous in 2 samples This SNP lies in a gene-desert between RPP21 and HLA-
E, and is close to an L1-Line repeat at position 30,474,008 Two samples are heterozygous at position 31,551,302 (rs12660382) This SNP lies within a dormant ribosomal RNA large-subunit, and within 50bp of a THE1A element (at position 31,551,334) that is associated with recombination hotspots (Myers et al 2005) Additionally, 3 samples are heterozygous at position 32,821,245 (rs2239800), a located within an intron of the HLA-DQA2 locus
Trang 93.4.3 Shared Ancestral Blocks Among Conserved Haplotypes
The MHC is associated with over 100 diseases, including most autoimmune and infection diseases (Lechler and Warren 2000, Horton et al 2004) Most of these disease associations are identified through segregation of disease phenotype with an HLA allele, several of which belong to known CEHs in various populations (Price et
al 1999, Lechler and Warren 2000, Stewart et al 2004) Very few of these diseases have been conclusively linked to an HLA allele; rather it is thought that non-HLA genes in strong LD with the HLA locus are major contributing factors, possibly transmitted along within the conserved haplotype A lack of detailed knowledge of the variation and allelic content of the CEHs beyond the HLA genes has hampered fine-mapping of disease loci To this end, the MHC haplotype project (Allcock et al 2002, Stewart et al 2004, Traherne et al 2006, Horton et al 2008) set out to completely sequence 8 Caucasian CEHs, providing a wealth of variation data of the MHC in the Caucasian population The high-resolution SNP map of the 4 Chinese CEHs in this study contributes to this objective and enables comparisons to be made between the Asian and European CEHs For example, the Caucasian A1-B8-DR3 and the Chinese A2-B46-DR9 haplotypes are both associated with myasthenia gravis (Chan et al
1993, Price et al 1999), knowing which segments are shared within both CEHs will help narrow down possible disease loci, a form of recombinant mapping between these conserved haplotypes
In order to compare the 8 full-length Caucasian CEH sequences with the Chinese
Trang 10accurately determined for each Caucasian CEH SNPs with flanking sequences that were poorly matched or had multiple mapping locations in a CEH were ignored and null alleles was assigned to those positions To identify segments that were shared between CEHs, the 203 haplotype-blocks identified earlier in Section 3.2.3 were used – each CEH was converted into a string of contiguous haplotype blocks based on the SNP alleles Haplotype blocks accurately identify regions of low diversity that are uninterrupted by recombination, with sizes and distribution that are robust across non-African populations (International HapMap Project 2005), and using these as the basic unit for comparisons will succinctly describe similarities and differences between CEHs For the local Chinese CEHs, the reference haplotypes for the A2-B46-DR9 and A33-B58-DR3 CEHs were obtained from the consensus of homozygous samples, while the reference haplotypes for the A11-B15-DR12 and A2-B38-DR1 CEHs were constructed using the HLA-homozygous samples supplemented
by consensus of phase-inferred chromosomes
The haplotypes of the 12 CEHs (8 Caucasian and 4 Chinese) were compared from 29.0Mb to 33.0Mb and illustrated as 4 separate 1Mb panels in Figure 3.28 Haplotype blocks are drawn linearly across each panel with different colours used to distinguish between different haplotypes within each block Gaps in the diagram signify blocks where the haplotype could not be identified for a particular CEH, either due to gaps in the sequence assembly (for the full-length MHC sequences) or an inability to identify
a consensus haplotype due to recombination (local CEHs) Large gaps are seen in 5 of the 8 Caucasian CEHs (APD, DBB, MANN, MCF and SSTO), as the sequence coverage of the MHC in these haplotypes is not complete (Horton et al 2008)
Trang 15Shared regions between CEHs are clearly identifiable, for instance the HLA-A2 carrying haplotypes share a common pattern between 29.5Mb to 30.15Mb, replicating the fixed A2 region reported in the earlier section The shared segments between the local Chinese CEHs and the other haplotypes are also summarized in Table 3.10 In this list, consecutive runs of similar haplotype blocks are considered a contiguous segment, and only shared segments between 2 CEHs that are longer than 20kb are included The gene loci that are covered within each segment are also listed Several interesting observations are described in detail in the following paragraphs
A33-B58-DR3 CEH Comparisons
The A33-B58-DR3 haplotype remarkably shows more similarity to the Caucasian COX, PGF and QBL haplotypes than to the other 3 Chinese haplotypes Of these, A33-B58-DR3 shares the most number of similar segments, up to 1.15Mb across the MHC, with the QBL haplotype As this haplotype pair carries the same DR allele, they expectedly share up to 270kb long segments in the Class II region, covering the HLA-DR and DQB1 loci These 2 haplotypes also show high symmetry within the Class III region; in fact more than half of the entire Class III region is similar between these 2 haplotypes This includes 2 large 150kb and 170kb segments that carrying the
BF, C2 NOTCH4 and HSPA1A loci These 2 haplotypes are however not similar at the RCCX (RP-C4-CYP21-TNX) module that is sandwiched in between This set of observations confirms data from a recent publication that genotyped STR markers in a collection of CEHs at several locations across the MHC (Dorak et al 2006) The
Trang 16Start (Mb) End (Mb) Length (kb)
32.055 32.213 158.39 C4A C4B CREBL1 CYP21A2 FKBPL STK19 TNXB 31.478 31.534 55.89 MICA
30.432 30.472 40.32 32.806 32.839 33.16 HLA-DQA2 HLA-DQB2 32.418 32.447 28.30 C6orf10
32.519 32.547 28.20 HLA-DRA 30.432 30.472 40.32
32.222 32.253 30.02 AGPAT1 EGFL8 PPT2 PRRT1 31.588 31.613 24.19 BAT1
32.055 32.213 158.39 C4A C4B CREBL1 CYP21A2 FKBPL STK19 TNXB 29.772 29.839 66.28 HLA-F
30.432 30.472 40.32 32.767 32.790 22.49 31.538 31.683 145.35 ATP6V1G2 BAT1 HCG26 HCP5 LST1 LTA LTB MICB
NCR3 NFKBIL1 TNF 30.569 30.662 92.79 ABCF1 GNL1 HLA-E PRR3 32.931 33.007 76.28 PPP1R2P1 PSMB9 32.191 32.253 61.36 AGPAT1 CREBL1 EGFL8 FKBPL PPT2 PRRT1 32.734 32.790 55.33 HLA-DQB1
30.187 30.228 41.11 TRIM10 TRIM31 TRIM40 30.432 30.472 40.32
32.674 32.711 36.68 29.444 29.478 33.62 OR12D2 OR12D3 OR5V1 31.887 31.910 23.94 C6orf48 HSPA1A HSPA1B HSPA1L 30.879 30.901 22.54 C6orf214
A*3303-C*0302-B*5801-DRB1*0301
VS DBB: A*0201-C*0602-B*5701-DRB1*0701
31.973 32.038 65.89 BF C2 EHMT2 RDBP SKIV2L ZBTB12 32.418 32.487 68.71 BTNL2 C6orf10 HCG23
32.222 32.253 30.02 AGPAT1 EGFL8 PPT2 PRRT1 31.355 31.382 27.04
29.382 29.444 61.18 OR5U1 OR5V1 31.330 31.382 52.08 HLA-C 32.445 32.482 36.93 BTNL2 C6orf10 HCG23 29.453 29.478 24.73 OR12D2 OR12D3 OR5V1 30.996 31.020 24.32 SFTPG VARSL 29.245 29.389 143.52 OR2J2 OR5U1 32.115 32.213 98.11 CREBL1 CYP21A2 FKBPL TNXB 32.931 33.007 76.28 PPP1R2P1 PSMB9
29.449 29.478 28.67 OR12D2 OR12D3 OR5V1 32.418 32.447 28.30 C6orf10
31.681 31.784 102.81 AIF1 APOM BAT2 BAT3 BAT4 BAT5 C6orf47
CSNK2B LY6G5B LY6G5C LY6G6D 29.712 29.792 79.54 MOG ZFP57
31.304 31.338 33.74 32.445 32.472 27.08 BTNL2 C6orf10 HCG23
A*3303-C*0302-B*5801-DRB1*0301
VS MCF: A*0201-C*0304-B*1501-DRB1*0401
A*3303-C*0302-B*5801-DRB1*0301
VS PGF: A*0301-C*0702-B*0702-DRB1*1501
A*3303-C*0302-B*5801-DRB1*0301
VS SSTO: A*3201-C*0501-B*4402-DRB1*0403
A*3303-C*0302-B*5801-DRB1*0301
VS A*0207-C*0102-B*4601-DRB1*0901
A*3303-C*0302-B*5801-DRB1*0301
VS A*1101-C*0801-B*1502-DRB1*1202
A*3303-C*0302-B*5801-DRB1*0301
VS COX: A*0101-C*0701-B*0801-DRB1*0301
A*3303-C*0302-B*5801-DRB1*0301
VS MANN: A*2902-C*1601-B*4403-DRB1*0701
A*3303-C*0302-B*5801-DRB1*0301
VS A*0203-C*0702-B*3802-DRB1*1602
Table 3.10 Shared Segments Between CEH Pairs