- Email: [email protected]
Genetic associations of common deletion polymorphisms in families with Avellino corneal dystrophy
Biochemical and Biophysical Research Communications 387 (2009) 688–693
Contents lists available at ScienceDirect
Biochemical and Biophysical Research Communications journal homepage: www.elsevier.com/locate/ybbrc
Genetic associations of common deletion polymorphisms in families with Avellino corneal dystrophy Miey Park a, Dong-Joon Kim a, Kwang Joong Kim a,b, Chang Bum Hong a, Young Jin Kim a, Hyun Sub Cheong c, Hyoung Doo Shin d, Eun-Ju Lee e, Han-Na Kim e, Hye Won Chung f, Eung Kweon Kim g, Jong-Young Lee a,*, Hyung-Lae Kim a,* a
Center for Genome Sciences, National Institute of Health (NIH), 194 Tongil-Lo, Eunpyung-Gu, Seoul 122-701, Republic of Korea Department of Molecular Bioscience, Kangwon National University, Chunchon, Republic of Korea Department of Genetic Epidemiology, SNP Genetics, Inc., Seoul 153-801, Republic of Korea d Laboratory of Genomic Diversity, Department of Life Science, Sogang University, Seoul 121-742, Republic of Korea e Department of Biochemistry, College of Medicine, Ewha Women’s University, Seoul, Republic of Korea f Department of Obstetrics and Gynecology, College of Medicine, Ewha Women’s University, Seoul, Republic of Korea g Corneal Dystrophy Research Institute, Department of Ophthalmology, Yonsei University College of Medicine, Seoul, Republic of Korea b c
a r t i c l e
i n f o
Article history: Received 7 July 2009 Available online 19 July 2009 Keywords: Avellino corneal dystrophy (ACD) Copy number variation Deletion polymorphism HLA-F gene
a b s t r a c t Although the locations of many common deletion variants in the human genome are unknown, such deletions may be causative in rare disorders. Deletions can be mapped through the identification of Mendelian inconsistencies in pedigrees. Data for a total of 341,577 SNPs from an ACD family cohort (n = 551) and 341,039 SNPs from a Korean-Vietnamese family cohort (n = 554) were collected for a genome-wide association study using Illumina 370K-Duo BeadchipsÒ. In the present study, a Mendelian inconsistency analysis of genotype data identified 1029 deletion variants in Korean and Korean-Vietnam family cohorts of 404 trios comprising 1105 individuals. Small-deletion copy number variations adjacent to 10 deletion variants were then validated by the real-time quantitative polymerase chain reaction. The expected copy numbers of each deletion variant were directly matched to its genotype cluster image. Deletion variants were also in strong linkage disequilibrium with nearby SNPs. To determine the overall contribution of the 1029 deletion variants, we analyzed case–control trio associations with the risk for Avellino corneal dystrophy. One SNP marker (rs885945) neighboring the gene encoding major histocompatibility complex class I F (HLA-F) was significantly associated with the risk of Avellino corneal dystrophy (P = 0.0003). rs885945 showed high LD with SNPs within the HLA-F gene. Therefore, HLA-F may be a potential candidate gene for Avellino corneal dystrophy. Crown Copyright Ó 2009 Published by Elsevier Inc. All rights reserved.
Introduction Current SNP genotyping methods do not generally detect deletions. When a deletion is transmitted from parent to child, genotypes at SNPs often appear to violate the rules of Mendelian transmission [1]. Nevertheless, copy number variation (CNV) in particular deletions can be identified from genotype determinations by a variety of approaches. Conrad et al. used apparent errors in Mendelian inheritance to identify 586 potential deletions [1].
* Corresponding author. Fax: +82 2 354 1063. E-mail addresses: [email protected] (M. Park), [email protected] (D.-J. Kim), [email protected] (K.J. Kim), [email protected] (C.B. Hong), [email protected] (Y.J. Kim), [email protected] (H.S. Cheong), [email protected] (H.D. Shin), [email protected] (E.-J. Lee), [email protected] (H.-N. Kim), hyewon@ ewha.ac.kr (H.W. Chung), [email protected] (E.K. Kim), [email protected] (J.-Y. Lee), [email protected] (H.-L. Kim).
McCarroll et al. identified clusters of SNPs that were not in Hardy–Weinberg equilibrium and discovered 541 potential deletions in parent-offspring trios [2]. These approaches, however, are only capable of identifying deletions [3]. Deletions can also be causal in rare disorders. For example, in a study of autism, most mutations (12 out of 15) were identified as deletions and linked to a recurrent 16p11.2 microdeletion [4,5]. Deletion polymorphisms may also be important in the genetics of complex traits involving multiple factors, each of which has a relatively small effect [1]. Enhancement of, or qualitative change in, the activity of a protein because of an exon deletion is a particularly important consequence of some variants [6]. A number of deletions have been mapped by observing Mendelian inconsistencies in pedigrees, but that approach can only be applied to families [7]. A number of forms of CNV, including somatic chromosomal changes, duplications and deletions, have been implicated in the causation of several disorders [8,9]. CNVs can affect drug and
0006-291X/$ - see front matter Crown Copyright Ó 2009 Published by Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2009.07.084
689
M. Park et al. / Biochemical and Biophysical Research Communications 387 (2009) 688–693
immune responses as well as susceptibility to disease [10–12]. Recently, several studies have reported associations between CNVs and human disease, but CNV methods are still limited; in particular, there are no reliable genotyping methods for CNV. Accurate determination of CNV genotypes is also very important for association studies [4,7,13]. In this paper we describe a new approach to determining smalldeletion variants in CNV regions and confirming their copy numbers. We found several SNP markers with intensity differences indicating different copy numbers and identified deletions that are transmitted to offspring. To identify the SNP markers in common deletions, we manually inspected the genotype clusters in heritability errors (>1% Parent–Parent–Child error; P–P–C error). We identified 1029 deletion variants in 404 trios, 205 in Avellino corneal dystrophy (ACD) families and 199 in a Korean-Vietnamese family cohort. We also assessed the relationship between these 1029 deletion variants and an ACD case/control association (226 ACD and 325 controls) and experimentally validated 10 of them by real-time polymerase chain reaction (PCR).
Materials and methods Subjects and whole-genome SNP genotyping. Data for a total of 341,577 SNPs from the ACD families (n = 551) and 341,039 from the Korean-Vietnamese family cohort (n = 554) were collected for a genome-wide association study using Illumina 370K-Duo BeadchipsÒ. A total of 274 families (n = 1105) were also analyzed for Mendelian inconsistency within each family. Approximately 750 ng of genomic DNA was used to genotype each sample. Briefly, the entire genome was amplified, fragmented, precipitated and resuspended in an appropriate hybridization buffer. Denatured samples were hybridized on HumanCNV370-Duo BeadchipsÒ for a minimum of 16 h at 48 °C. After completion of the assay, the BeadChips were scanned with a two-color confocal BeadArrayTM reader. Image intensities were extracted and genotyped using Illumina’s BeadStudio 3.2Ò software. The overall call rate for all SNPs was 100%. Identification of common deletion variants. BeadStudio 3.2Ò software offers specialized genotype clustering plots consisting of normalized intensity values (y axis) and allelic intensity ratios (x axis). Two-color allele-specific fluorescence intensity measurements are displayed in clusters. This software was used to investigate genotype clusters according to SNP markers. The cluster image of a marker with Mendelian inconsistency is quite different from that of an autosomal marker without CNV [13]. To identify deletion variants, we deployed a heritability error check with
BeadStudio 3.2Ò software using 341,577 SNPs from the ACD families (205 trios) and 341,039 SNPs from the Korean-Vietnamese family cohort (199 trios); 1450 SNPs showed Mendelian inconsistency in more than 5 trios (>1%). We visually inspected all the genotype clusters of these 1450 SNPs (>1% P–P–C error) using BeadStudio 3.2Ò GenePlot. The 1029 SNP markers that showed clear intensity differences were identified as deletion variants [Supplementary Table 1]. Markers with unclear cluster separations were excluded to avoid possible genotyping errors. All 1029 SNPs (>1% P–P–C error) were also compared with the International HapMap database (http://www.hapmap.org) for possible overlap with known CNV regions. Statistical analysis (PBAT). In order to perform a family-based association test (FBAT), we transformed the deletion status of each variant into genotypes. Homozygous deletions, heterozygotes and homozygotes with normal copy numbers were coded ‘AA’, ‘AB’, and ‘BB’, respectively. After genotype coding with consideration of CNVs, each deletion variant was tested for association with the ACD family using FBAT, which is implemented in HelixTree with PBAT v5 (Golden Helix, Bozeman, MT) [14]. Association testing was done assuming an additive genetic model. CNV validation using TaqManÒ real-time PCR assay. To validate the existence of CNV regions around 10 SNPs (Table 1), Primer Express Software 3.0Ò was used to identify two sequences in each of these regions, which were then used to generate specific primers. The positions and lengths of hybridization probes and primers were optimized by the BLAST program and TaqMan MGB probes. Sequences for the TaqMan MGB probe and PCR primers are listed in [Supplementary Table 2]. Copy numbers were determined using the ABI Prism 7900 sequence detection system. The RNase P gene was co-amplified with the marker and served as an internal standard. Amplification reactions (10 ll) were carried out in duplicate with 10 ng of template DNA, 1 TaqManÒ Universal Master Mix buffer (Applied Biosystems, Foster City, CA), 900 nM of each primer, and 250 nM of each fluorogenic probe. Thermal cycling was initiated with 2 min incubation at 50 °C, followed by a first denaturation step of 10 min at 95 °C, and then 40 cycles of 15 s at 95 °C and 1 min at 60 °C. Three replicate reactions were performed for each primer pair, and a comparative CT method was used to calculate the copy number [15]; Applied Biosystems user bulletin #2 [p/n4303859]. DCT was calculated by subtracting the RNase P CT value from the sample CT value for each replicate. The average CT value for the three replicates was then calculated. The DCT values from all other samples were normalized to determine DDCT. The copy number was calculated by the formula 2 2DDC T . Assessment of LD. The LD between the deletion variants and SNPs was measured as described [4]. Briefly, we used the
Table 1 Family-based association test of common deletion variants with lowest P-value and the risk of Avellino corneal dystrophy among Korean family cohorts. Multi-allelic CNV marker
Chromosome (position)
Gene
Location (amino acid change)
Location relative to near gene (bp)
Alleles (%, frequency)
Nc
FBAT P-value
rs885945 rs11220 rs11810038 rs79072 rs13276950 rs10757330 rs11119097 rs693980 rs6451374 rs7551421
6 (29804831) 11 (73259191) 1 (182411618) 22 (47145850) 8 (93988068) 9 (22489958) 1 (206813021) 15 (43799987) 5 (38238788) 1 (109171438)
HLA-F CHCHD8 — — — DMRTA1 — SQRDL FLJ39155 C1orf62
Flanking 30 Flanking 30 Intergenica Intergenic Intergenic Flanking 30 Intergenic Flanking 30 Flanking 50 CDSb (Lys616Asn)
2569 2169 — — — 47,486 — 29,216 55,502 9431
C (29.4%), T (44.9%), (25.7%) C (62.7%), T (32.9%), (4.3%) G (9.0%), A (84.6%), (6.4%) G (50.4%), T (41.3%), (8.4%) G (26.7%), T (67.5%), (5.9%) A (76.2%), G (15.7%), (8.1%) A (58.3%), G (33.8%), (7.9%) G (47.3%), T (46.9%), (5.8%) G (49.6%), T (45.7%), (4.6%) G (21.0%), T (70.7%), (8.3%)
29 10 10 11 10 17 12 12 12 14
0.0003 0.004 0.006 0.008 0.013 0.013 0.014 0.015 0.016 0.017
, represents allele loss. a Location is >100 kb from the nearest gene. b Coding sequence. c Number of informative families.
690
M. Park et al. / Biochemical and Biophysical Research Communications 387 (2009) 688–693
gene-deletion genotypes (‘+/+’, ‘+/’, ‘/’) obtained by quantitative PCR (see above). We obtained all ACD family and KoreanVietnamese cohort SNP genotypes that extended 200 kb on either side of the deletion locus, removed SNPs that were covered by the deletion, and replaced them with the quantitative PCR-derived deletion genotypes from the same individuals. We then used the Haploview program to determine the phase of all SNP and deletion alleles to calculate the LD (r2) between the deletions and SNPs.
Results Identification of 1029 deletion variants We performed whole-genome genotyping on both the ACD families (n = 551) and Korean-Vietnamese cohort (n = 554) using Illumina 370K-Duo SNP chips (total n = 1105). The sample call rate was 100%, and the genotyping call rate was 99.82% (ACD families, 99.84%; Korean-Vietnamese family cohort, 99.79%). The minor allele frequency of the ACD families was 0.222, and that of the Korean-Vietnamese family cohort was 0.225. The ratio of SNPs with a Hardy–Weinberg equilibrium P-value of <0.01 was 1.89% for all samples, 2.08% for the ACD families, and 1.16% for the Korean-Vietnamese family cohort. These results indicate that few SNPs deviated from Hardy–Weinberg equilibrium, although the test samples were consistent with family. We analyzed all genotype clusters of 341,577 SNPs from the ACD families (n = 551) and 341,039 from the Korean-Vietnamese family cohort (n = 554) using Illumina BeadStudio 3.2 software. In addition, we performed a Mendelian inconsistency analysis on genotype data from 404 trios (199 ACD families and 205 Korean-Vietnamese family cohort) and identified 1029 common deletion variants in CNV regions. The
BeadStudio 3.2 software shows normal three genotype cluster images of SNP markers located in non-CNV regions. But, most SNP markers on the X chromosome have five genotyping clusters depending on gender and genotypes. Interestingly, similar genotype cluster images could represent the existence of CNVs around certain SNP markers on autosomal chromosomes. Annotation information for the deletion variants (SNP markers) is presented in [Supplementary Table 1]. The distributions in each chromosome and the locations of the deletion variants are shown in Fig. 1. Most of the markers were distributed in chromosome 6, with few in chromosome 21 (Fig. 1A). There were 13 common deletion variants (1.3%) in coding regions, 290 (28.2%) in intron regions, 388 (37.7%) in flanking-30 UTR regions, 329 (32.0%) in flanking-50 UTR regions, and 9 (0.9%) in UTRs (Fig. 1B). Novel regions To determine the proportion of variant regions discovered in this study that were novel, we compared our data with those in the Database of Genomic Variants (November 2008). Of our deletion variants, 449 (44%) were in the Database and 580 (56%) were not. Family-based association of deletion variants in the ACD families For each deletion variant, we performed FBAT using HelixTreeÒ software with the PBAT module and a case–control analysis of association with the risk for ACD. FBAT P-values and the number of informative families were computed for each deletion variant. The most significantly associated deletion variants (n = 10, P < 0.02) are shown in Table 1, which also presents gene information, allele frequencies, FBAT P-values and the number of families
Fig. 1. Number and distribution of the 1029 deletion variants across human autosomes. (A) The number of common CNV markers on chromosomes. Most of them were distributed in chromosome 6, with few in chromosome 21. (B) The locations of CNV markers. There were 13 common deletion variants (1.3%) in coding regions, 290 (28.2%) in intron regions, 388 (37.7%) in flanking-30 UTR regions, 329 (32.0%) in flanking-50 UTR regions and 9 (0.9%) in UTRs.
M. Park et al. / Biochemical and Biophysical Research Communications 387 (2009) 688–693
691
Fig. 2. Cluster image of rs885945 and CNV determinations identified by real-time quantitative PCR of rs885945. (A) Cluster image shows the six genotypes for deletion variants. BeadStudio 3.2Ò software was used for genotype clustering. (B) CNV was determined by real-time quantitative PCR around rs885945. The expression level of SNP rs885945 from individuals who were identified by quantitative PCR as having zero, one and two copies.
from which the information was obtained. Among the deletion variants, one near the gene encoding major histocompatibility complex (MHC) class I F (HLA-F) within the MHC region of chromo-
some 6 had the most significant association with ACD (P = 0.0003), suggesting that this deletion variant is a potential candidate gene for ACD (Table 1). The associated P-value of rs885945 could not
Fig. 3. Linkage disequilibrium (r2) of deletion polymorphisms with SNPs. For each deletion variant, strong linkage disequilibrium is observed with other SNPs.
692
M. Park et al. / Biochemical and Biophysical Research Communications 387 (2009) 688–693
retain significance when Bonferroni’s correction was strictly adopted, but considering the low sample size in this study the association result might be worth following up.
deletion polymorphism may be important for HLA-F function. The high LD among SNPs within HLA-F and the deletion marker further supports the hypothesis that the deletion tends to be inherited together with the SNPs closest to HLA-F (Fig. 4).
Validation of CNVs around deletion variants To validate the existence of CNVs around the 10 deletion variants in Table 1 directly, we used quantitative PCR to genotype each individual accurately as a carrier of zero, one or two copies. As shown in Fig. 2 (and Supplementary Fig. 1), which presents the quantitative copy number values for individuals, no CT value could be calculated for any sample with the genotype /, indicating homozygous deletion. Two-copy number combinations, namely A/A A/B, and B/B, proved to be two- or multi-fold higher than in individuals with a single copy of the allele (A/ or B/). Linkage disequilibrium of deletion polymorphisms with SNPs We investigated whether the copy number of deletion variants could be reliably predicted from nearby SNPs. We treated haplotype SNPs at deletion variants as a quantitative trait (Fig. 3). Four of the 10 deletion variants showed significant linkage disequilibrium (LD) with nearby SNPs in the ACD families and the KoreanVietnamese family cohort, suggesting that the deletions and SNPs have similar evolutionary histories and those deletion polymorphisms can be reliably predicted from neighboring SNPs. LD block The observed CNV of the SNP marker rs885945 was statistically significant so it could be a causal deletion marker for ACD. In the 500-kb window, HLA-F was the gene closest to rs885945. Thus, this
Discussion ACD is a common corneal dystrophy disease that shows allelic homogeneity with an R124H mutation in the transforming growth factor beta-induced gene [16–18]. The R124H mutation is associated with lattice cornea dystrophy in which the protein ßig-h3 is mislocated and degraded [19,20]. ßig-h3 is an extracellular matrix protein secreted by non-leukocytes that regulates cell proliferation, differentiation and adhesion. It has also been shown to be prominently up-regulated in immature dendritic cells. Abundant ßig-h3 mRNA is detected in lymphoid tissues, and secretion of the protein by immature dendritic cells strongly suggests a role in immune regulation [20,21]. Immature dendritic cells are present mainly in peripheral tissues, and they participate in antigen uptake mechanisms such as receptor-dependent endocytosis and receptor-independent macropinocytosis. Many microbial structures and inflammatory stimuli can activate immature dendritic cells to undergo a complex maturation process that increases the expression of surface MHC and other co-stimulatory molecules [20,22,23]. Lee and Geraghty (2003) reported that cell surface expression of HLA-F is induced by transformation of normal peripheral blood B cells by EB virus infection, indicating that HLA-F is expressed on the surfaces of activated cells. Lee and Geraghty also proposed that HLA-F expression signals to immunocytes that the cell is in an activated state. The function of HLA-F, however, has not been elucidated [24].
Fig. 4. Haploview display of linkage disequilibrium between rs885945 and other markers.
M. Park et al. / Biochemical and Biophysical Research Communications 387 (2009) 688–693
Recently, much effort has been devoted to identifying CNVs in SNP genotype data in both familial and unrelated samples [7,25,26]. With whole-genome SNP analysis, the SNP array has become a valuable method for measuring copy number changes on the basis of probe signal intensity [13,27]. In the case of deletions, CNVs as small as few kilobases can be detected from genotyping data by identifying errors in Mendelian inheritance [3]. In this study, we developed an approach for discovering deletions from SNP genotypes in CNV regions. However, the boundaries of the CNV regions were not discernible and we could not evaluate the copy gain for CNVs by comparing intensities owing to the unclear cluster separation. Despite these limitations in our results, the information about deletion variants may help in identifying CNV regions within which accurate genotypes of SNPs can be called and in CNV-related association studies. Our finding about the LD between SNPs and deletions indicates that SNPs serve as CNV markers by LD for structural variants throughout the genome. Whereas deletion polymorphisms in unique regions of the human genome appear to result from ancestral mutations and to segregate on ancestral haplotypes [28], little is known about the LD properties of deletions or duplications in repeat-rich, structurally complex genomic regions [29]. We observed strong LD between SNPs from both the ACD families and KoreanVietnamese family cohort. Four of the 10 deletion variants showed significant LD with nearby SNPs and were perfect SNP proxies (r2 = 1). The LD between SNPs within CNV regions indicates the utility of a single database that integrates data on SNP genotypes and structural polymorphisms. Also, the LD between SNPs and deletions was similar to that between SNPs and other SNPs [2]. Among 1029 deletion variants, rs885945 showed the most significant association with the risk of ACD. Although rs885945 is not within the HLA-F coding region, it showed a high LD with SNPs within HLA-F. The markers rs1362126, rs2076180, and rs2235383 within HLA-F as well as other markers in flanking regions were not apparently associated with deletion polymorphisms. They showed normal genotype cluster images of SNP markers located in non-CNV regions [Supplementary Fig. 2]. Therefore, we suggest that the deletion polymorphism may play an important role in HLA-F function and that this gene may be a potential candidate for ACD. Acknowledgment This work was supported by an intramural Grant (4845-301430-210-13) from the Korea National Institute of Health, Korea Center for Disease Control, Republic of Korea. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bbrc.2009.07.084. References [1] D.F. Conrad, T.D. Andrews, N.P. Carter, M.E. Hurles, J.K. Pritchard, A highresolution survey of deletion polymorphism in the human genome, Nat. Genet. 38 (2006) 75–81. [2] S.A. McCarroll, T.N. Hadnott, G.H. Perry, P.C. Sabeti, M.C. Zody, J.C. Barrett, S. Dallaire, S.B. Gabriel, C. Lee, M.J. Daly, D.M. Altshuler, Common deletion polymorphisms in the human genome, Nat. Genet. 38 (2006) 86–92. [3] N.P. Carter, Methods and strategies for analyzing copy number variation using DNA microarrays, Nat. Genet. 39 (2007) S16–21. [4] J. Sebat, B. Lakshmi, D. Malhotra, J. Troge, C. Lese-Martin, T. Walsh, B. Yamrom, S. Yoon, A. Krasnitz, J. Kendall, A. Leotta, D. Pai, R. Zhang, Y.H. Lee, J. Hicks, S.J. Spence, A.T. Lee, K. Puura, T. Lehtimaki, D. Ledbetter, P.K. Gregersen, J. Bregman, J.S. Sutcliffe, V. Jobanputra, W. Chung, D. Warburton, M.C. King, D. Skuse, D.H. Geschwind, T.C. Gilliam, K. Ye, M. Wigler, Strong association of de novo copy number mutations with autism, Science 316 (2007) 445–449.
693
[5] R.A. Kumar, S. KaraMohamed, J. Sudi, D.F. Conrad, C. Brune, J.A. Badner, T.C. Gilliam, N.J. Nowak, E.H. Cook Jr., W.B. Dobyns, S.L. Christian, Recurrent 16p11.2 microdeletions in autism, Hum. Mol. Genet. 17 (2008) 628–638. [6] B. Frank, J.L. Bermejo, K. Hemminki, C. Sutter, B. Wappenschmidt, A. Meindl, M. Kiechle-Bahat, P. Bugert, R.K. Schmutzler, C.R. Bartram, B. Burwinkel, Copy number variant in the candidate tumor suppressor gene MTUS1 and familial breast cancer risk, Carcinogenesis 28 (2007) 1442–1445. [7] L. Franke, C.G. deKovel, Y.S. Aulchenko, G. Trynka, A. Zhernakova, K.A. Hunt, H.M. Blauw, L.H. van den Berg, R. Ophoff, P. Deloukas, D.A. van Heel, C. Wijmenga, Detection, imputation, and association analysis of small deletions and null alleles on oligonucleotide arrays, Am. J. Hum. Genet. 82 (2008) 1316–1333. [8] K. Inoue, J.R. Lupski, Molecular mechanisms for genomic disorders, Annu. Rev. Genomics Hum. Genet. 3 (2002) 199–242. [9] J.A. Lee, J.R. Lupski, Genomic rearrangements and gene copy-number alterations as a cause of nervous system disorders, Neuron 52 (2006) 103–121. [10] E.J. Hollox, J.A. Armour, J.C. Barber, Extensive normal copy number variation of a beta-defensin antimicrobial-gene cluster, Am. J. Hum. Genet. 73 (2003) 591– 600. [11] E. Gonzalez, H. Kulkarni, H. Bolivar, A. Mangano, R. Sanchez, G. Catano, R.J. Nibbs, B.I. Freedman, M.P. Quinones, M.J. Bamshad, K.K. Murthy, B.H. Rovin, W. Bradley, R.A. Clark, S.A. Anderson, J. O’Connell, R.B.K. Agan, S.S. Ahuja, R. Bologna, L. Sen, M.J. Dolan, S.K. Ahuja, The influence of CCL3L1 gene-containing segmental duplications on HIV-1/AIDS susceptibility, Science 307 (2005) 1434–1440. [12] T.J. Aitman, R. Dong, T.J. Vyse, P.J. Norsworthy, M.D. Johnson, J. Smith, J. Mangion, C. Roberton-Lowe, A.J. Marshall, E. Petretto, M.D. Hodges, G. Bhangal, S.G. Patel, K. Sheehan-Rooney, M. Duda, P.R. Cook, D.J. Evans, J. Domin, J. Flint, J.J. Boyle, C.D. Pusey, H.T. Cook, Copy number polymorphism in Fcgr3 predisposes to glomerulonephritis in rats and humans, Nature 439 (2006) 851–855. [13] J.S. Bae, H.S. Cheong, J.O. Kim, S.O. Lee, E.M. Kim, H.W. Lee, S. Kim, J.W. Kim, T. Cui, I. Inoue, H.D. Shin, Identification of SNP markers for common CNV regions and association analysis of risk of subarachnoid aneurysmal hemorrhage in Japanese population, Biochem. Biophys. Res. Commun. 373 (2008) 593–596. [14] C. Lange, D. DeMeo, E.K. Silverman, S.T. Weiss, N.M. Laird, PBAT: tools for family-based association studies, Am. J. Hum. Genet. 74 (2004) 367–369. [15] L. Bodin, P.H. Beaune, M.A. Loriot, Determination of cytochrome P450 2D6 (CYP2D6) gene copy number by real-time quantitative PCR, J. Biomed. Biotechnol. 2005 (2005) 248–253. [16] R. Folberg, E. Alfonso, J.O. Croxatto, N.G. Driezen, N. Panjwani, P.R. Laibson, S.A. Boruchoff, J. Baum, E.S. Malbran, R. Fernandez-Meijide, et al., Clinically atypical granular corneal dystrophy with pathologic features of lattice-like amyloid deposits. A study of these families, Ophthalmology 95 (1988) 46–51. [17] G.O. Rosenwasser, B.M. Sucheski, N. Rosa, B. Pastena, A. Sebastiani, J.W. Sassani, H.D. Perry, Phenotypic variation in combined granular-lattice (Avellino) corneal dystrophy, Arch. Ophthalmol. 111 (1993) 1546–1552. [18] K. Tsujikawa, M. Tsujikawa, H. Watanabe, N. Maeda, Y. Inoue, T. Fujikado, Y. Tano, Allelic homogeneity in Avellino corneal dystrophy due to a founder effect, J. Hum. Genet. 52 (2007) 92–97. [19] K. Fujiki, Y. Hotta, K. Nakayasu, T. Yamaguchi, T. Kato, Y. Uesugi, N.T. Ha, S. Endo, N. Ishida, W.N. Lu, A. Kanai, Six different mutations of TGFBI (betaig-h3, keratoepithelin) gene found in Japanese corneal dystrophies, Cornea 19 (2000) 842–845. [20] W. Cao, P. Tan, C.H. Lee, H. Zhang, J. Lu, A transforming growth factor-betainduced protein stimulates endocytosis and is up-regulated in immature dendritic cells, Blood 107 (2006) 2777–2785. [21] J.E. Kim, R.W. Park, J.Y. Choi, Y.C. Bae, K.S. Kim, C.K. Joo, I.S. Kim, Molecular properties of wild-type and mutant betaIG-H3 proteins, Invest. Ophthalmol. Vis. Sci. 43 (2002) 656–661. [22] J. Banchereau, F. Briere, C. Caux, J. Davoust, S. Lebecque, Y.J. Liu, B. Pulendran, K. Palucka, Immunobiology of dendritic cells, Annu. Rev. Immunol. 18 (2000) 767–811. [23] I. Mellman, S.J. Turley, R.M. Steinman, Antigen processing for amateurs and professionals, Trends Cell Biol. 8 (1998) 231–237. [24] T. Shobu, N. Sageshima, H. Tokui, M. Omura, K. Saito, Y. Nagatsuka, M. Nakanishi, Y. Hayashi, K. Hatake, A. Ishitani, The surface expression of HLA-F on decidual trophoblasts increases from mid to term gestation, J. Reprod. Immunol. 72 (2006) 18–32. [25] J.R. Kohler, D.J. Cutler, Simultaneous discovery and testing of deletions for disease association in SNP genotyping studies, Am. J. Hum. Genet. 81 (2007) 684–699. [26] K. Kosta, I. Sabroe, J. Goke, R.J. Nibbs, J. Tsanakas, M.K. Whyte, M.D. Teare, A Bayesian approach to copy-number-polymorphism analysis in nuclear pedigrees, Am. J. Hum. Genet. 81 (2007) 808–812. [27] G.R. Bignell, J. Huang, J. Greshock, S. Watt, A. Butler, S. West, M. Grigorova, K.W. Jones, W. Wei, M.R. Stratton, P.A. Futreal, B. Weber, M.H. Shapero, R. Wooster, High-resolution analysis of DNA copy number using oligonucleotide microarrays, Genome Res. 14 (2004) 287–295. 1[28] D.A. Hinds, A.P. Kloek, M. Jen, X. Chen, K.A. Frazer, Common deletions and SNPs are in linkage disequilibrium in the human genome, Nat. Genet. 38 (2006) 82–85. [29] D.P. Locke, A.J. Sharp, S.A. McCarroll, S.D. McGrath, T.L. Newman, Z. Cheng, S. Schwartz, D.G. Albertson, D. Pinkel, D.M. Altshuler, E.E. Eichler, Linkage disequilibrium and heritability of copy-number polymorphisms within duplicated regions of the human genome, Am. J. Hum. Genet. 79 (2006) 275–290.
Contents lists available at ScienceDirect
Biochemical and Biophysical Research Communications journal homepage: www.elsevier.com/locate/ybbrc
Genetic associations of common deletion polymorphisms in families with Avellino corneal dystrophy Miey Park a, Dong-Joon Kim a, Kwang Joong Kim a,b, Chang Bum Hong a, Young Jin Kim a, Hyun Sub Cheong c, Hyoung Doo Shin d, Eun-Ju Lee e, Han-Na Kim e, Hye Won Chung f, Eung Kweon Kim g, Jong-Young Lee a,*, Hyung-Lae Kim a,* a
Center for Genome Sciences, National Institute of Health (NIH), 194 Tongil-Lo, Eunpyung-Gu, Seoul 122-701, Republic of Korea Department of Molecular Bioscience, Kangwon National University, Chunchon, Republic of Korea Department of Genetic Epidemiology, SNP Genetics, Inc., Seoul 153-801, Republic of Korea d Laboratory of Genomic Diversity, Department of Life Science, Sogang University, Seoul 121-742, Republic of Korea e Department of Biochemistry, College of Medicine, Ewha Women’s University, Seoul, Republic of Korea f Department of Obstetrics and Gynecology, College of Medicine, Ewha Women’s University, Seoul, Republic of Korea g Corneal Dystrophy Research Institute, Department of Ophthalmology, Yonsei University College of Medicine, Seoul, Republic of Korea b c
a r t i c l e
i n f o
Article history: Received 7 July 2009 Available online 19 July 2009 Keywords: Avellino corneal dystrophy (ACD) Copy number variation Deletion polymorphism HLA-F gene
a b s t r a c t Although the locations of many common deletion variants in the human genome are unknown, such deletions may be causative in rare disorders. Deletions can be mapped through the identification of Mendelian inconsistencies in pedigrees. Data for a total of 341,577 SNPs from an ACD family cohort (n = 551) and 341,039 SNPs from a Korean-Vietnamese family cohort (n = 554) were collected for a genome-wide association study using Illumina 370K-Duo BeadchipsÒ. In the present study, a Mendelian inconsistency analysis of genotype data identified 1029 deletion variants in Korean and Korean-Vietnam family cohorts of 404 trios comprising 1105 individuals. Small-deletion copy number variations adjacent to 10 deletion variants were then validated by the real-time quantitative polymerase chain reaction. The expected copy numbers of each deletion variant were directly matched to its genotype cluster image. Deletion variants were also in strong linkage disequilibrium with nearby SNPs. To determine the overall contribution of the 1029 deletion variants, we analyzed case–control trio associations with the risk for Avellino corneal dystrophy. One SNP marker (rs885945) neighboring the gene encoding major histocompatibility complex class I F (HLA-F) was significantly associated with the risk of Avellino corneal dystrophy (P = 0.0003). rs885945 showed high LD with SNPs within the HLA-F gene. Therefore, HLA-F may be a potential candidate gene for Avellino corneal dystrophy. Crown Copyright Ó 2009 Published by Elsevier Inc. All rights reserved.
Introduction Current SNP genotyping methods do not generally detect deletions. When a deletion is transmitted from parent to child, genotypes at SNPs often appear to violate the rules of Mendelian transmission [1]. Nevertheless, copy number variation (CNV) in particular deletions can be identified from genotype determinations by a variety of approaches. Conrad et al. used apparent errors in Mendelian inheritance to identify 586 potential deletions [1].
* Corresponding author. Fax: +82 2 354 1063. E-mail addresses: [email protected] (M. Park), [email protected] (D.-J. Kim), [email protected] (K.J. Kim), [email protected] (C.B. Hong), [email protected] (Y.J. Kim), [email protected] (H.S. Cheong), [email protected] (H.D. Shin), [email protected] (E.-J. Lee), [email protected] (H.-N. Kim), hyewon@ ewha.ac.kr (H.W. Chung), [email protected] (E.K. Kim), [email protected] (J.-Y. Lee), [email protected] (H.-L. Kim).
McCarroll et al. identified clusters of SNPs that were not in Hardy–Weinberg equilibrium and discovered 541 potential deletions in parent-offspring trios [2]. These approaches, however, are only capable of identifying deletions [3]. Deletions can also be causal in rare disorders. For example, in a study of autism, most mutations (12 out of 15) were identified as deletions and linked to a recurrent 16p11.2 microdeletion [4,5]. Deletion polymorphisms may also be important in the genetics of complex traits involving multiple factors, each of which has a relatively small effect [1]. Enhancement of, or qualitative change in, the activity of a protein because of an exon deletion is a particularly important consequence of some variants [6]. A number of deletions have been mapped by observing Mendelian inconsistencies in pedigrees, but that approach can only be applied to families [7]. A number of forms of CNV, including somatic chromosomal changes, duplications and deletions, have been implicated in the causation of several disorders [8,9]. CNVs can affect drug and
0006-291X/$ - see front matter Crown Copyright Ó 2009 Published by Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2009.07.084
689
M. Park et al. / Biochemical and Biophysical Research Communications 387 (2009) 688–693
immune responses as well as susceptibility to disease [10–12]. Recently, several studies have reported associations between CNVs and human disease, but CNV methods are still limited; in particular, there are no reliable genotyping methods for CNV. Accurate determination of CNV genotypes is also very important for association studies [4,7,13]. In this paper we describe a new approach to determining smalldeletion variants in CNV regions and confirming their copy numbers. We found several SNP markers with intensity differences indicating different copy numbers and identified deletions that are transmitted to offspring. To identify the SNP markers in common deletions, we manually inspected the genotype clusters in heritability errors (>1% Parent–Parent–Child error; P–P–C error). We identified 1029 deletion variants in 404 trios, 205 in Avellino corneal dystrophy (ACD) families and 199 in a Korean-Vietnamese family cohort. We also assessed the relationship between these 1029 deletion variants and an ACD case/control association (226 ACD and 325 controls) and experimentally validated 10 of them by real-time polymerase chain reaction (PCR).
Materials and methods Subjects and whole-genome SNP genotyping. Data for a total of 341,577 SNPs from the ACD families (n = 551) and 341,039 from the Korean-Vietnamese family cohort (n = 554) were collected for a genome-wide association study using Illumina 370K-Duo BeadchipsÒ. A total of 274 families (n = 1105) were also analyzed for Mendelian inconsistency within each family. Approximately 750 ng of genomic DNA was used to genotype each sample. Briefly, the entire genome was amplified, fragmented, precipitated and resuspended in an appropriate hybridization buffer. Denatured samples were hybridized on HumanCNV370-Duo BeadchipsÒ for a minimum of 16 h at 48 °C. After completion of the assay, the BeadChips were scanned with a two-color confocal BeadArrayTM reader. Image intensities were extracted and genotyped using Illumina’s BeadStudio 3.2Ò software. The overall call rate for all SNPs was 100%. Identification of common deletion variants. BeadStudio 3.2Ò software offers specialized genotype clustering plots consisting of normalized intensity values (y axis) and allelic intensity ratios (x axis). Two-color allele-specific fluorescence intensity measurements are displayed in clusters. This software was used to investigate genotype clusters according to SNP markers. The cluster image of a marker with Mendelian inconsistency is quite different from that of an autosomal marker without CNV [13]. To identify deletion variants, we deployed a heritability error check with
BeadStudio 3.2Ò software using 341,577 SNPs from the ACD families (205 trios) and 341,039 SNPs from the Korean-Vietnamese family cohort (199 trios); 1450 SNPs showed Mendelian inconsistency in more than 5 trios (>1%). We visually inspected all the genotype clusters of these 1450 SNPs (>1% P–P–C error) using BeadStudio 3.2Ò GenePlot. The 1029 SNP markers that showed clear intensity differences were identified as deletion variants [Supplementary Table 1]. Markers with unclear cluster separations were excluded to avoid possible genotyping errors. All 1029 SNPs (>1% P–P–C error) were also compared with the International HapMap database (http://www.hapmap.org) for possible overlap with known CNV regions. Statistical analysis (PBAT). In order to perform a family-based association test (FBAT), we transformed the deletion status of each variant into genotypes. Homozygous deletions, heterozygotes and homozygotes with normal copy numbers were coded ‘AA’, ‘AB’, and ‘BB’, respectively. After genotype coding with consideration of CNVs, each deletion variant was tested for association with the ACD family using FBAT, which is implemented in HelixTree with PBAT v5 (Golden Helix, Bozeman, MT) [14]. Association testing was done assuming an additive genetic model. CNV validation using TaqManÒ real-time PCR assay. To validate the existence of CNV regions around 10 SNPs (Table 1), Primer Express Software 3.0Ò was used to identify two sequences in each of these regions, which were then used to generate specific primers. The positions and lengths of hybridization probes and primers were optimized by the BLAST program and TaqMan MGB probes. Sequences for the TaqMan MGB probe and PCR primers are listed in [Supplementary Table 2]. Copy numbers were determined using the ABI Prism 7900 sequence detection system. The RNase P gene was co-amplified with the marker and served as an internal standard. Amplification reactions (10 ll) were carried out in duplicate with 10 ng of template DNA, 1 TaqManÒ Universal Master Mix buffer (Applied Biosystems, Foster City, CA), 900 nM of each primer, and 250 nM of each fluorogenic probe. Thermal cycling was initiated with 2 min incubation at 50 °C, followed by a first denaturation step of 10 min at 95 °C, and then 40 cycles of 15 s at 95 °C and 1 min at 60 °C. Three replicate reactions were performed for each primer pair, and a comparative CT method was used to calculate the copy number [15]; Applied Biosystems user bulletin #2 [p/n4303859]. DCT was calculated by subtracting the RNase P CT value from the sample CT value for each replicate. The average CT value for the three replicates was then calculated. The DCT values from all other samples were normalized to determine DDCT. The copy number was calculated by the formula 2 2DDC T . Assessment of LD. The LD between the deletion variants and SNPs was measured as described [4]. Briefly, we used the
Table 1 Family-based association test of common deletion variants with lowest P-value and the risk of Avellino corneal dystrophy among Korean family cohorts. Multi-allelic CNV marker
Chromosome (position)
Gene
Location (amino acid change)
Location relative to near gene (bp)
Alleles (%, frequency)
Nc
FBAT P-value
rs885945 rs11220 rs11810038 rs79072 rs13276950 rs10757330 rs11119097 rs693980 rs6451374 rs7551421
6 (29804831) 11 (73259191) 1 (182411618) 22 (47145850) 8 (93988068) 9 (22489958) 1 (206813021) 15 (43799987) 5 (38238788) 1 (109171438)
HLA-F CHCHD8 — — — DMRTA1 — SQRDL FLJ39155 C1orf62
Flanking 30 Flanking 30 Intergenica Intergenic Intergenic Flanking 30 Intergenic Flanking 30 Flanking 50 CDSb (Lys616Asn)
2569 2169 — — — 47,486 — 29,216 55,502 9431
C (29.4%), T (44.9%), (25.7%) C (62.7%), T (32.9%), (4.3%) G (9.0%), A (84.6%), (6.4%) G (50.4%), T (41.3%), (8.4%) G (26.7%), T (67.5%), (5.9%) A (76.2%), G (15.7%), (8.1%) A (58.3%), G (33.8%), (7.9%) G (47.3%), T (46.9%), (5.8%) G (49.6%), T (45.7%), (4.6%) G (21.0%), T (70.7%), (8.3%)
29 10 10 11 10 17 12 12 12 14
0.0003 0.004 0.006 0.008 0.013 0.013 0.014 0.015 0.016 0.017
, represents allele loss. a Location is >100 kb from the nearest gene. b Coding sequence. c Number of informative families.
690
M. Park et al. / Biochemical and Biophysical Research Communications 387 (2009) 688–693
gene-deletion genotypes (‘+/+’, ‘+/’, ‘/’) obtained by quantitative PCR (see above). We obtained all ACD family and KoreanVietnamese cohort SNP genotypes that extended 200 kb on either side of the deletion locus, removed SNPs that were covered by the deletion, and replaced them with the quantitative PCR-derived deletion genotypes from the same individuals. We then used the Haploview program to determine the phase of all SNP and deletion alleles to calculate the LD (r2) between the deletions and SNPs.
Results Identification of 1029 deletion variants We performed whole-genome genotyping on both the ACD families (n = 551) and Korean-Vietnamese cohort (n = 554) using Illumina 370K-Duo SNP chips (total n = 1105). The sample call rate was 100%, and the genotyping call rate was 99.82% (ACD families, 99.84%; Korean-Vietnamese family cohort, 99.79%). The minor allele frequency of the ACD families was 0.222, and that of the Korean-Vietnamese family cohort was 0.225. The ratio of SNPs with a Hardy–Weinberg equilibrium P-value of <0.01 was 1.89% for all samples, 2.08% for the ACD families, and 1.16% for the Korean-Vietnamese family cohort. These results indicate that few SNPs deviated from Hardy–Weinberg equilibrium, although the test samples were consistent with family. We analyzed all genotype clusters of 341,577 SNPs from the ACD families (n = 551) and 341,039 from the Korean-Vietnamese family cohort (n = 554) using Illumina BeadStudio 3.2 software. In addition, we performed a Mendelian inconsistency analysis on genotype data from 404 trios (199 ACD families and 205 Korean-Vietnamese family cohort) and identified 1029 common deletion variants in CNV regions. The
BeadStudio 3.2 software shows normal three genotype cluster images of SNP markers located in non-CNV regions. But, most SNP markers on the X chromosome have five genotyping clusters depending on gender and genotypes. Interestingly, similar genotype cluster images could represent the existence of CNVs around certain SNP markers on autosomal chromosomes. Annotation information for the deletion variants (SNP markers) is presented in [Supplementary Table 1]. The distributions in each chromosome and the locations of the deletion variants are shown in Fig. 1. Most of the markers were distributed in chromosome 6, with few in chromosome 21 (Fig. 1A). There were 13 common deletion variants (1.3%) in coding regions, 290 (28.2%) in intron regions, 388 (37.7%) in flanking-30 UTR regions, 329 (32.0%) in flanking-50 UTR regions, and 9 (0.9%) in UTRs (Fig. 1B). Novel regions To determine the proportion of variant regions discovered in this study that were novel, we compared our data with those in the Database of Genomic Variants (November 2008). Of our deletion variants, 449 (44%) were in the Database and 580 (56%) were not. Family-based association of deletion variants in the ACD families For each deletion variant, we performed FBAT using HelixTreeÒ software with the PBAT module and a case–control analysis of association with the risk for ACD. FBAT P-values and the number of informative families were computed for each deletion variant. The most significantly associated deletion variants (n = 10, P < 0.02) are shown in Table 1, which also presents gene information, allele frequencies, FBAT P-values and the number of families
Fig. 1. Number and distribution of the 1029 deletion variants across human autosomes. (A) The number of common CNV markers on chromosomes. Most of them were distributed in chromosome 6, with few in chromosome 21. (B) The locations of CNV markers. There were 13 common deletion variants (1.3%) in coding regions, 290 (28.2%) in intron regions, 388 (37.7%) in flanking-30 UTR regions, 329 (32.0%) in flanking-50 UTR regions and 9 (0.9%) in UTRs.
M. Park et al. / Biochemical and Biophysical Research Communications 387 (2009) 688–693
691
Fig. 2. Cluster image of rs885945 and CNV determinations identified by real-time quantitative PCR of rs885945. (A) Cluster image shows the six genotypes for deletion variants. BeadStudio 3.2Ò software was used for genotype clustering. (B) CNV was determined by real-time quantitative PCR around rs885945. The expression level of SNP rs885945 from individuals who were identified by quantitative PCR as having zero, one and two copies.
from which the information was obtained. Among the deletion variants, one near the gene encoding major histocompatibility complex (MHC) class I F (HLA-F) within the MHC region of chromo-
some 6 had the most significant association with ACD (P = 0.0003), suggesting that this deletion variant is a potential candidate gene for ACD (Table 1). The associated P-value of rs885945 could not
Fig. 3. Linkage disequilibrium (r2) of deletion polymorphisms with SNPs. For each deletion variant, strong linkage disequilibrium is observed with other SNPs.
692
M. Park et al. / Biochemical and Biophysical Research Communications 387 (2009) 688–693
retain significance when Bonferroni’s correction was strictly adopted, but considering the low sample size in this study the association result might be worth following up.
deletion polymorphism may be important for HLA-F function. The high LD among SNPs within HLA-F and the deletion marker further supports the hypothesis that the deletion tends to be inherited together with the SNPs closest to HLA-F (Fig. 4).
Validation of CNVs around deletion variants To validate the existence of CNVs around the 10 deletion variants in Table 1 directly, we used quantitative PCR to genotype each individual accurately as a carrier of zero, one or two copies. As shown in Fig. 2 (and Supplementary Fig. 1), which presents the quantitative copy number values for individuals, no CT value could be calculated for any sample with the genotype /, indicating homozygous deletion. Two-copy number combinations, namely A/A A/B, and B/B, proved to be two- or multi-fold higher than in individuals with a single copy of the allele (A/ or B/). Linkage disequilibrium of deletion polymorphisms with SNPs We investigated whether the copy number of deletion variants could be reliably predicted from nearby SNPs. We treated haplotype SNPs at deletion variants as a quantitative trait (Fig. 3). Four of the 10 deletion variants showed significant linkage disequilibrium (LD) with nearby SNPs in the ACD families and the KoreanVietnamese family cohort, suggesting that the deletions and SNPs have similar evolutionary histories and those deletion polymorphisms can be reliably predicted from neighboring SNPs. LD block The observed CNV of the SNP marker rs885945 was statistically significant so it could be a causal deletion marker for ACD. In the 500-kb window, HLA-F was the gene closest to rs885945. Thus, this
Discussion ACD is a common corneal dystrophy disease that shows allelic homogeneity with an R124H mutation in the transforming growth factor beta-induced gene [16–18]. The R124H mutation is associated with lattice cornea dystrophy in which the protein ßig-h3 is mislocated and degraded [19,20]. ßig-h3 is an extracellular matrix protein secreted by non-leukocytes that regulates cell proliferation, differentiation and adhesion. It has also been shown to be prominently up-regulated in immature dendritic cells. Abundant ßig-h3 mRNA is detected in lymphoid tissues, and secretion of the protein by immature dendritic cells strongly suggests a role in immune regulation [20,21]. Immature dendritic cells are present mainly in peripheral tissues, and they participate in antigen uptake mechanisms such as receptor-dependent endocytosis and receptor-independent macropinocytosis. Many microbial structures and inflammatory stimuli can activate immature dendritic cells to undergo a complex maturation process that increases the expression of surface MHC and other co-stimulatory molecules [20,22,23]. Lee and Geraghty (2003) reported that cell surface expression of HLA-F is induced by transformation of normal peripheral blood B cells by EB virus infection, indicating that HLA-F is expressed on the surfaces of activated cells. Lee and Geraghty also proposed that HLA-F expression signals to immunocytes that the cell is in an activated state. The function of HLA-F, however, has not been elucidated [24].
Fig. 4. Haploview display of linkage disequilibrium between rs885945 and other markers.
M. Park et al. / Biochemical and Biophysical Research Communications 387 (2009) 688–693
Recently, much effort has been devoted to identifying CNVs in SNP genotype data in both familial and unrelated samples [7,25,26]. With whole-genome SNP analysis, the SNP array has become a valuable method for measuring copy number changes on the basis of probe signal intensity [13,27]. In the case of deletions, CNVs as small as few kilobases can be detected from genotyping data by identifying errors in Mendelian inheritance [3]. In this study, we developed an approach for discovering deletions from SNP genotypes in CNV regions. However, the boundaries of the CNV regions were not discernible and we could not evaluate the copy gain for CNVs by comparing intensities owing to the unclear cluster separation. Despite these limitations in our results, the information about deletion variants may help in identifying CNV regions within which accurate genotypes of SNPs can be called and in CNV-related association studies. Our finding about the LD between SNPs and deletions indicates that SNPs serve as CNV markers by LD for structural variants throughout the genome. Whereas deletion polymorphisms in unique regions of the human genome appear to result from ancestral mutations and to segregate on ancestral haplotypes [28], little is known about the LD properties of deletions or duplications in repeat-rich, structurally complex genomic regions [29]. We observed strong LD between SNPs from both the ACD families and KoreanVietnamese family cohort. Four of the 10 deletion variants showed significant LD with nearby SNPs and were perfect SNP proxies (r2 = 1). The LD between SNPs within CNV regions indicates the utility of a single database that integrates data on SNP genotypes and structural polymorphisms. Also, the LD between SNPs and deletions was similar to that between SNPs and other SNPs [2]. Among 1029 deletion variants, rs885945 showed the most significant association with the risk of ACD. Although rs885945 is not within the HLA-F coding region, it showed a high LD with SNPs within HLA-F. The markers rs1362126, rs2076180, and rs2235383 within HLA-F as well as other markers in flanking regions were not apparently associated with deletion polymorphisms. They showed normal genotype cluster images of SNP markers located in non-CNV regions [Supplementary Fig. 2]. Therefore, we suggest that the deletion polymorphism may play an important role in HLA-F function and that this gene may be a potential candidate for ACD. Acknowledgment This work was supported by an intramural Grant (4845-301430-210-13) from the Korea National Institute of Health, Korea Center for Disease Control, Republic of Korea. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bbrc.2009.07.084. References [1] D.F. Conrad, T.D. Andrews, N.P. Carter, M.E. Hurles, J.K. Pritchard, A highresolution survey of deletion polymorphism in the human genome, Nat. Genet. 38 (2006) 75–81. [2] S.A. McCarroll, T.N. Hadnott, G.H. Perry, P.C. Sabeti, M.C. Zody, J.C. Barrett, S. Dallaire, S.B. Gabriel, C. Lee, M.J. Daly, D.M. Altshuler, Common deletion polymorphisms in the human genome, Nat. Genet. 38 (2006) 86–92. [3] N.P. Carter, Methods and strategies for analyzing copy number variation using DNA microarrays, Nat. Genet. 39 (2007) S16–21. [4] J. Sebat, B. Lakshmi, D. Malhotra, J. Troge, C. Lese-Martin, T. Walsh, B. Yamrom, S. Yoon, A. Krasnitz, J. Kendall, A. Leotta, D. Pai, R. Zhang, Y.H. Lee, J. Hicks, S.J. Spence, A.T. Lee, K. Puura, T. Lehtimaki, D. Ledbetter, P.K. Gregersen, J. Bregman, J.S. Sutcliffe, V. Jobanputra, W. Chung, D. Warburton, M.C. King, D. Skuse, D.H. Geschwind, T.C. Gilliam, K. Ye, M. Wigler, Strong association of de novo copy number mutations with autism, Science 316 (2007) 445–449.
693
[5] R.A. Kumar, S. KaraMohamed, J. Sudi, D.F. Conrad, C. Brune, J.A. Badner, T.C. Gilliam, N.J. Nowak, E.H. Cook Jr., W.B. Dobyns, S.L. Christian, Recurrent 16p11.2 microdeletions in autism, Hum. Mol. Genet. 17 (2008) 628–638. [6] B. Frank, J.L. Bermejo, K. Hemminki, C. Sutter, B. Wappenschmidt, A. Meindl, M. Kiechle-Bahat, P. Bugert, R.K. Schmutzler, C.R. Bartram, B. Burwinkel, Copy number variant in the candidate tumor suppressor gene MTUS1 and familial breast cancer risk, Carcinogenesis 28 (2007) 1442–1445. [7] L. Franke, C.G. deKovel, Y.S. Aulchenko, G. Trynka, A. Zhernakova, K.A. Hunt, H.M. Blauw, L.H. van den Berg, R. Ophoff, P. Deloukas, D.A. van Heel, C. Wijmenga, Detection, imputation, and association analysis of small deletions and null alleles on oligonucleotide arrays, Am. J. Hum. Genet. 82 (2008) 1316–1333. [8] K. Inoue, J.R. Lupski, Molecular mechanisms for genomic disorders, Annu. Rev. Genomics Hum. Genet. 3 (2002) 199–242. [9] J.A. Lee, J.R. Lupski, Genomic rearrangements and gene copy-number alterations as a cause of nervous system disorders, Neuron 52 (2006) 103–121. [10] E.J. Hollox, J.A. Armour, J.C. Barber, Extensive normal copy number variation of a beta-defensin antimicrobial-gene cluster, Am. J. Hum. Genet. 73 (2003) 591– 600. [11] E. Gonzalez, H. Kulkarni, H. Bolivar, A. Mangano, R. Sanchez, G. Catano, R.J. Nibbs, B.I. Freedman, M.P. Quinones, M.J. Bamshad, K.K. Murthy, B.H. Rovin, W. Bradley, R.A. Clark, S.A. Anderson, J. O’Connell, R.B.K. Agan, S.S. Ahuja, R. Bologna, L. Sen, M.J. Dolan, S.K. Ahuja, The influence of CCL3L1 gene-containing segmental duplications on HIV-1/AIDS susceptibility, Science 307 (2005) 1434–1440. [12] T.J. Aitman, R. Dong, T.J. Vyse, P.J. Norsworthy, M.D. Johnson, J. Smith, J. Mangion, C. Roberton-Lowe, A.J. Marshall, E. Petretto, M.D. Hodges, G. Bhangal, S.G. Patel, K. Sheehan-Rooney, M. Duda, P.R. Cook, D.J. Evans, J. Domin, J. Flint, J.J. Boyle, C.D. Pusey, H.T. Cook, Copy number polymorphism in Fcgr3 predisposes to glomerulonephritis in rats and humans, Nature 439 (2006) 851–855. [13] J.S. Bae, H.S. Cheong, J.O. Kim, S.O. Lee, E.M. Kim, H.W. Lee, S. Kim, J.W. Kim, T. Cui, I. Inoue, H.D. Shin, Identification of SNP markers for common CNV regions and association analysis of risk of subarachnoid aneurysmal hemorrhage in Japanese population, Biochem. Biophys. Res. Commun. 373 (2008) 593–596. [14] C. Lange, D. DeMeo, E.K. Silverman, S.T. Weiss, N.M. Laird, PBAT: tools for family-based association studies, Am. J. Hum. Genet. 74 (2004) 367–369. [15] L. Bodin, P.H. Beaune, M.A. Loriot, Determination of cytochrome P450 2D6 (CYP2D6) gene copy number by real-time quantitative PCR, J. Biomed. Biotechnol. 2005 (2005) 248–253. [16] R. Folberg, E. Alfonso, J.O. Croxatto, N.G. Driezen, N. Panjwani, P.R. Laibson, S.A. Boruchoff, J. Baum, E.S. Malbran, R. Fernandez-Meijide, et al., Clinically atypical granular corneal dystrophy with pathologic features of lattice-like amyloid deposits. A study of these families, Ophthalmology 95 (1988) 46–51. [17] G.O. Rosenwasser, B.M. Sucheski, N. Rosa, B. Pastena, A. Sebastiani, J.W. Sassani, H.D. Perry, Phenotypic variation in combined granular-lattice (Avellino) corneal dystrophy, Arch. Ophthalmol. 111 (1993) 1546–1552. [18] K. Tsujikawa, M. Tsujikawa, H. Watanabe, N. Maeda, Y. Inoue, T. Fujikado, Y. Tano, Allelic homogeneity in Avellino corneal dystrophy due to a founder effect, J. Hum. Genet. 52 (2007) 92–97. [19] K. Fujiki, Y. Hotta, K. Nakayasu, T. Yamaguchi, T. Kato, Y. Uesugi, N.T. Ha, S. Endo, N. Ishida, W.N. Lu, A. Kanai, Six different mutations of TGFBI (betaig-h3, keratoepithelin) gene found in Japanese corneal dystrophies, Cornea 19 (2000) 842–845. [20] W. Cao, P. Tan, C.H. Lee, H. Zhang, J. Lu, A transforming growth factor-betainduced protein stimulates endocytosis and is up-regulated in immature dendritic cells, Blood 107 (2006) 2777–2785. [21] J.E. Kim, R.W. Park, J.Y. Choi, Y.C. Bae, K.S. Kim, C.K. Joo, I.S. Kim, Molecular properties of wild-type and mutant betaIG-H3 proteins, Invest. Ophthalmol. Vis. Sci. 43 (2002) 656–661. [22] J. Banchereau, F. Briere, C. Caux, J. Davoust, S. Lebecque, Y.J. Liu, B. Pulendran, K. Palucka, Immunobiology of dendritic cells, Annu. Rev. Immunol. 18 (2000) 767–811. [23] I. Mellman, S.J. Turley, R.M. Steinman, Antigen processing for amateurs and professionals, Trends Cell Biol. 8 (1998) 231–237. [24] T. Shobu, N. Sageshima, H. Tokui, M. Omura, K. Saito, Y. Nagatsuka, M. Nakanishi, Y. Hayashi, K. Hatake, A. Ishitani, The surface expression of HLA-F on decidual trophoblasts increases from mid to term gestation, J. Reprod. Immunol. 72 (2006) 18–32. [25] J.R. Kohler, D.J. Cutler, Simultaneous discovery and testing of deletions for disease association in SNP genotyping studies, Am. J. Hum. Genet. 81 (2007) 684–699. [26] K. Kosta, I. Sabroe, J. Goke, R.J. Nibbs, J. Tsanakas, M.K. Whyte, M.D. Teare, A Bayesian approach to copy-number-polymorphism analysis in nuclear pedigrees, Am. J. Hum. Genet. 81 (2007) 808–812. [27] G.R. Bignell, J. Huang, J. Greshock, S. Watt, A. Butler, S. West, M. Grigorova, K.W. Jones, W. Wei, M.R. Stratton, P.A. Futreal, B. Weber, M.H. Shapero, R. Wooster, High-resolution analysis of DNA copy number using oligonucleotide microarrays, Genome Res. 14 (2004) 287–295. 1[28] D.A. Hinds, A.P. Kloek, M. Jen, X. Chen, K.A. Frazer, Common deletions and SNPs are in linkage disequilibrium in the human genome, Nat. Genet. 38 (2006) 82–85. [29] D.P. Locke, A.J. Sharp, S.A. McCarroll, S.D. McGrath, T.L. Newman, Z. Cheng, S. Schwartz, D.G. Albertson, D. Pinkel, D.M. Altshuler, E.E. Eichler, Linkage disequilibrium and heritability of copy-number polymorphisms within duplicated regions of the human genome, Am. J. Hum. Genet. 79 (2006) 275–290.