January 2009
Volume 50, Issue 1
Free
Biochemistry and Molecular Biology  |   January 2009
Association of PAX6 Polymorphisms with High Myopia in Han Chinese Nuclear Families
Author Affiliations
  • Wei Han
    From the Department of Ophthalmology, The First Affiliated Hospital, Medical College, Zhejiang University, Hangzhou, China; the
    School of Optometry and the
    Department of Health Technology and Informatics, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong SAR, China.
  • Kim Hung Leung
    Department of Health Technology and Informatics, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong SAR, China.
  • Wai Yan Fung
    Department of Health Technology and Informatics, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong SAR, China.
  • Joey Y. Y. Mak
    Department of Health Technology and Informatics, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong SAR, China.
  • Yu Min Li
    From the Department of Ophthalmology, The First Affiliated Hospital, Medical College, Zhejiang University, Hangzhou, China; the
  • Maurice K. H. Yap
    School of Optometry and the
  • Shea Ping Yip
    Department of Health Technology and Informatics, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong SAR, China.
Investigative Ophthalmology & Visual Science January 2009, Vol.50, 47-56. doi:10.1167/iovs.07-0813
  • Views
  • PDF
  • Share
  • Tools
    • Alerts
      ×
      This feature is available to authenticated users only.
      Sign In or Create an Account ×
    • Get Citation

      Wei Han, Kim Hung Leung, Wai Yan Fung, Joey Y. Y. Mak, Yu Min Li, Maurice K. H. Yap, Shea Ping Yip; Association of PAX6 Polymorphisms with High Myopia in Han Chinese Nuclear Families. Invest. Ophthalmol. Vis. Sci. 2009;50(1):47-56. doi: 10.1167/iovs.07-0813.

      Download citation file:


      © ARVO (1962-2015); The Authors (2016-present)

      ×
  • Supplements
Abstract

purpose. The paired box 6 (PAX6) gene is critical to eye development. Based on prior linkage evidence, this study was conducted to investigate the association of PAX6 polymorphisms with high myopia in a Han Chinese population.

methods. Tag single nucleotide polymorphisms (tSNPs) in the PAX6 locus were selected based on HapMap data, and other polymorphisms in its functional regions were also identified. Both tSNPs and identified variants were genotyped in 164 nuclear families with 170 highly myopic (spherical equivalent < −6.0 D in both eyes) offspring. The linkage disequilibrium pattern of SNPs was established in the parental group (n = 328). Family-based association tests were performed using family-based association testing (FBAT) and genetic association computer analyses.

results. Single marker analysis of SNPs rs3026390 and rs3026393 showed significant association with high myopia as a qualitative trait in dominant and recessive models (P = 0.0014 and P = 0.0011, respectively). For rs3026393, the genotype relative risk was 2.57 for G/T and 2.22 for T/T with reference to G/G. Significantly increased transmission was demonstrated for the haplotypes carrying allele T of rs3026393 in the additive and dominant models (P < 0.0070), whereas significantly decreased transmission was found for haplotypes carrying allele G of rs3026393 in the recessive model (P = 0.0173). Preferential transmission of single alleles and haplotypes remained significant after correction for multiple comparisons.

conclusions. This study demonstrates the association of PAX6 variants with susceptibility to high myopia. The PAX6 locus may contain polymorphisms playing a role in high myopia in southern Han Chinese.

Myopia is the most prevalent human eye refractive error in all human populations and has become a public health problem causing considerable social cost every year. 1 2 In general, a myopic eye with refractive error less than −6 D is termed as high myopia or pathologic myopia. 3 High myopia is associated with serious ocular morbidity and hence is a major cause of visual impairment and even blindness. 4  
Understanding the genetic network of myopia has been the subject of numerous studies for years. Myopia is thought to be multifactorial with environmental and genetic factors as well as their interactions being involved. 5 6 7 8 9 Asian populations generally have a higher prevalence and degree of myopia than Caucasian and African populations, 1 10 11 which suggests that specific hereditary backgrounds may account for this interesting phenomenon. Twin studies have consistently reported a higher concordance of myopia in monozygotic than dizygotic twins and high heritability for myopia. 12 13 All forms of myopia including low, moderate, and high myopia are likely to have a genetic involvement, 11 12 13 14 15 with the suggestion that high myopia may have a stronger genetic background. 16 17 18 19 20 21 Linkage studies have identified several chromosome loci for high myopia, and these have been reviewed in detail. 9 22 Nonsyndromic high myopia is thought to be polygenic, 20 and association studies are therefore likely to be a more powerful approach to identifying genetic variations for such complex traits. 23 Recent association studies have reported significant association of individual candidate genes with high myopia, although none of these have so far been replicated. 24 25 26 27 The use of the transmission disequilibrium test (TDT) 28 and its derivatives, such as family-based association tests (FBATs), 29 can avoid spurious association due to population stratification and thus provide an important means of analysis in complex trait gene mapping. In the current study, we used such an approach to investigate the relationship between PAX6 polymorphisms and high myopia. 
The paired box 6 (PAX6) gene is a highly conserved member of a family of transcription factors containing the paired and homeobox DNA-binding domains and is an essential pleiotropic transcriptional regulator for vertebrate tissue development such as eye, brain, and pancreas. 30 Mutations of the PAX6 gene have been extensively investigated and can result in different disease phenotype, such as human aniridia, keratitis, and microphthalmia. 31 32 33 34 Dosage of PAX6 expression is also critical to normal eye development, because overexpression can cause microphthalmia. 35 The significant changes in PAX6 expression level in an animal form-deprivation myopia model suggest that PAX6 may be involved in the occurrence of myopia. 36 In 221 randomly selected dizygotic twins, Hammond et al. 37 found strong linkage of refractive error with the PAX6 locus, but association of PAX6 polymorphisms could not be demonstrated in their study. Both linkage and functional evidence from previous studies suggest that PAX6 plays a role in the control of eye globe growth. Therefore, we hypothesized that the PAX6 gene is a potential candidate for susceptibility to high myopia. 
The PAX6 gene is located on chromosome 11 region p13 and has 13 exons spanning approximately 25 kilobases (kb). Using the FBAT approach, we investigated the genetic association between high myopia and PAX6 polymorphisms in a group of Han Chinese nuclear families with highly myopic siblings. The International Haplotype Genetic Map (HapMap) provides a genome-wide linkage disequilibrium (LD) pattern and greatly facilitates the association analysis for complex diseases. In this study, the tag single nucleotide polymorphism (tSNP) markers for family-based association analysis were selected according to the phase II HapMap data (http://www.hapmap.org/ provided in the public domain by a consortium of partners and sponsors worldwide). In addition, we also used 20 random healthy Han Chinese samples to screen variants in the exons and potential regulatory regions of the PAX6 gene and the identified SNPs or microsatellite (MS) were tested in subsequent association analyses as well. 
Materials and Methods
Subjects
Genomic DNA was extracted from venous blood of random healthy donors and all recruited high myopia nuclear family members as previously described. 24 38 The study was approved by the Human Subject Ethics Subcommittees of Hong Kong Polytechnic University and Zhejiang University and adhered to the tenets of the Declaration of Helsinki. Twenty samples from random healthy Han Chinese donors obtained from Hong Kong Red Cross Blood Transfusion Service were used for identifying potential variants. Each recruited nuclear family consisted of two parents and at least one affected sibling with high myopia. Clinical data including onset age of myopia, refractive error, intraocular pressure (IOP), corneal power (CP), ocular axial length (AXL), anterior chamber depth (ACD), and lens thickness (LT) were obtained as previously described. 24 The entry criterion for highly myopic siblings in this study was less stringent than that in our previous study 24 as we recruited some subjects with a spherical equivalent worse than −6.0 D for both eyes (n = 18). The entry age of these additional 18 siblings was younger than 16 years, and we thought that they were very likely to develop high myopia worse than −10.0 D. Mean spherical equivalent (MSE) of the two eyes for each sibling was used for analysis. Myopic siblings were excluded if they had an average corneal power (CP) of two meridians > 47.0 D in either eye. 39  
Selection of Genetic Markers for Association Analysis
Genetic markers were selected based on three approaches. First, tSNPs in the PAX6 locus were selected based on the tSNP data of phase II HapMap and tested in the genetic association study. Second, the PAX6 coding sequence and their immediate flanking intronic regions (within 250 bp) and noncoding sequences approximately 3.5 kb upstream of the start codon and 2.0 kb downstream of the stop codon were screened to identify potential functional polymorphisms. The identified SNPs or MS were also tested in association analysis. Third, according to the tSNP data of phase II HapMap, additional SNPs were tested if they were in tight LD with the SNPs, showing significant association in the initial phase. 
Polymerase Chain Reaction
Four tSNPs were identified from the HapMap database for PAX6, and four primer pairs were designed on computer (Oligo, ver. 6.57; Molecular Biology Insights, Cascade, WA) to amplify the DNA fragments containing tSNPs: rs3026354 (SNP2), rs3026393 (SNP6), rs1506 (SNP7), and rs662702 (SNP9). Note that the SNPs are also called SNP1, SNP2, and so on, in the order from the 5′ to the 3′ end of the locus, for the sake of easy discussion. The tSNPs were selected based on the phase II HapMap data with the criteria of r 2 > 0.8 and minor allele frequency (MAF) > 0.05. 40 Another primer pair was designed to amplify a GA dinucleotide MS marker (VNTR11_31793115; http://www.microsatellites.org) in the 5′ upstream region. Meanwhile, 28 additional primer pairs were designed to amplify the PAX6 coding and potential regulatory sequences. Primer sequences are available on request. Touchdown PCR was used to amplify the DNA samples, and the same DNA pooling strategy of identifying variants was applied again in the screening stage. 38  
SNP Identification and Genotyping for Random Healthy Blood Donors
In the screening stage, a DNA fragment analysis system (WAVE; Transgenomic, Omaha, NE) was used to analyze the PCR products. The details of denaturing high performance liquid chromatography (DHPLC) analysis for SNP identification are described in our previous report. 38  
Genotyping of SNP and MS Markers for Myopia Families
The tSNP rs3026393 (SNP6), located in the recognition sequence (5′CTCTTCN1↓3′/3′GAGAAGN4↑5′) of Eam1104I (or Ksp6321) was genotyped with restriction analysis. The forward primer was 5′-TTTTTAGGTTTGCCTCTCTTCTCACAGC-3′ with poly-T tail (in bold) and a modified base T (in italic) to create an internal control site for enzyme reaction. The reverse primer was 5′-CCATCCCCCAGGGACAAGGA-3′. A 10-μL reaction mix consisting of 5 μL PCR product, 1 μL 10× yellow buffer (Tango; MBI Fermentas, Vilnius, Lithuania), and 5 units of Eam1104I was prepared and incubated at 37°C for 16 hours. The restriction products were analyzed by agarose gel electrophoresis. The size of the PCR product is 306 bp. Cleavage of the internal control site produced two DNA fragments (285 and 21 bp). For samples carrying allele T, the 285-bp fragment was further cut into two fragments of 203 and 82 bp. 
The other 3 tSNPs—SNP2, -7, and -9—were genotyped with an SNP detection kit (AcycloPrime-FP; PerkinElmer, Boston, MA). The method was a modification of template-directed dye-terminator incorporation with fluorescence polarization (FP) detection. 41 The sequences of the SNP primers were 5′-CCATCCCCAAACTCTTTAACTG-3′ for SNP2, 5′-ATAAATCGGCACACATCTACACCC-3′ for SNP7, and 5′-TGTGTGGTGAGGGAACTGTCAGA-3′ for SNP9. The optimal cycles for FP reaction were 30 to 35 for these three tSNPs. FP was measured (Victor 3 V Multilabel Reader; PerkinElmer), and genotypes were called using the SNPscorer software (PerkinElmer). 
For all the remaining SNPs, including rs1894620 (SNP1), rs3026371 (SNP3), rs2239789 (SNP4), rs3026390 (SNP5), and rs12421026 (SNP8), the same DHPLC-based approach 38 was applied to genotype the high myopia samples because of the logistic arrangement in the use of instruments in our laboratory. 
The MS located in the 5′ upstream region was genotyped based on the size of the PCR-amplified fragments separated with a genetic analyzer (Prism 310; Applied Biosystems [ABI], Foster City, CA). The sequence was 5′-FAM-CCACTGGGCTGCTAGGAA-3′, for the 5′ labeled forward primer, and 5′-CCTTGGCACCTTCAGGCT-3′, for the reverse primer. The PCR product (0.5 μL) was mixed with 0.5-μL fluorescent size standard (GeneScan-500 ROX Standard; ABI) and 9 μL formamide (HiDi; ABI). The mixture was subjected to electrophoresis in a genetic analyzer (Prism 310; ABI), and the raw data were analyzed on computer (GeneScan 3.1 software; ABI). 
Statistical Analysis of Ocular Data
Commercial software (SPSS, ver. 11.0; SPSS Inc., Chicago, IL) was used to test the partial correlation between MSE of the affected siblings and other ocular components (AXL, ACD, LT, and CP). 
LD Analysis of the SNPs in the Parental Group
The SNP genotype data of the parents of high-myopia nuclear families were input into the software Haploview 42 (version 3.32; http://www.broad.mit.edu/mpg/haploview) which then performed the Hardy-Weinberg equilibrium (HWE) test and haplotype block analysis. Haploview also calculated the LD parameter r 2
Genetic Association Study
A genetic association study was performed using the Family-Based Association Test software package (FBAT, ver. 1.7.3; http://www.biostat.harvard.edu/fbat/default.html/). 29 43 Association tests for single loci and haplotypes under additive, dominant and recessive models were performed using FBAT with high myopia taken as a dichotomous trait: affected with high myopia (spherical equivalent < −6.0 D for both eyes) or unaffected. The null hypothesis was no association in the presence of linkage in view of the prior linkage evidence from Hammond’s study, 37 and the alternative hypothesis was that there was both linkage and association. 
The multiple-comparison problems for the alleles of a given marker were solved by the global statistic under any given genetic model. The more powerful false-discovery rate (FDR) 44 was used to control for multiple-hypothesis testing, instead of the conventional Bonferroni adjustment. For bi-allelic SNP markers, dominant and recessive models give reciprocal results and thus are equivalent to one test for the purpose of accounting for multiple testing. For multiallelic MS markers, dominant and recessive models results are not reciprocal, and thus are two independent tests. Therefore, for the global statistics, there were nine SNPs, each tested under two different genetic models (additive and dominant/recessive), plus one MS tested under three genetic models and thus 21 tests of global association altogether. After adjustment for multiple comparisons and with an FDR level of 0.05, the cutoff for significant global association was 0.0048 calculated as P′ = 0.05i/21 when P i < 0.05i/21 for the first time, where P i is the i th probability in the list of 21 observed values sorted in an ascending order and the comparisons between P i and 0.05i/21 are made sequentially from the largest to the smallest values. 
Similarly, the multiple-comparison issues for the haplotypes of selected SNPs were solved by the global statistic under any given genetic model. For the global statistics of haplotypes consisting of four tSNPs, there were three tests of global association (see 1 Table 4 ). For the global statistics of haplotypes of SNP5, -6, and -8, there were four groups of haplotypes each tested under three different genetic models and thus 12 tests of global association 1 4 (Table 5) . Thus, there were 15 comparisons. After adjustment for multiple comparisons with an FDR of 0.05, the cutoff for significant global association was 0.0300 calculated as P′ = 0.05i/15 when P i < 0.05i/15 for the first time, where P i is the i th probability in the list of 15 observed probabilities sorted in ascending order, and the comparisons between P i and 0.05i/15 are made sequentially from the largest to the smallest values. 
A matched case–control dataset was generated with each myopic sibling matched to three possible pseudocontrol subjects created from the untransmitted parental allele. 45 46 Conditional logistic regression was used to analyze this case–pseudocontrol dataset and calculate the effect size of the marker genotype on the disease risk as the genotype relative risk (GRR) and the corresponding 95% confidence intervals (CIs). Analysis was performed with the GenAssoc package (http://www-gene.cimr.cam.ac.uk/clayton/software/stata/ developed by David Clayton, Cambridge University, Cambridge, UK) and executed within commercial software (Stata, ver. 8.2; Stata Corp., College Station, TX). 
Results
Analysis of Clinical Myopia Data
The detailed information of myopic offspring is listed in Table 1 . The MSE of highly myopic offspring was −11.99 ± 3.60 D with a single-eye spherical diopter ranging from −6.0 to −22.50 D and cylinder diopter from 0 to −4.25 D. Of the total 164 nuclear family families (328 parents and 170 highly myopic offspring) studied, 158 had one affected sibling with high myopia, and six had two highly myopic siblings. No significant sex association for the prevalence of high myopia was observed among the myopic siblings (P > 0.05). For all affected siblings, corneal curvature examination showed no corneal shape anomaly. Partial correlation analysis showed significant correlation between MSE and ocular refractive components of AXL, ACD, and CP, with the exception of LT. In line with our previous study, 24 AXL still had the strongest correlation (r = −0.60; P < 0.001) to refractive error. 
LD Analysis of the SNPs in the Parental Group
No SNP used in this study was found in the PAX6 coding sequence. All SNPs have been reported in the SNP database (http://www.ncbi.nlm.nih.gov/sites/entrez?db=snp&cmd=search&term=/ provided in the public domain by the National Center for Biotechnology Information, Bethesda, MD). In particular, three (SNP6, -7, and -9) of the four tSNPs were identified in our random Chinese samples, whereas SNP2 (also a tSNP) was not found in our random Chinese samples. SNP1 and -3 were identified in our study but not documented in the HapMap database, while SNP4 and -8 were identified in our study and also documented the HapMap database. SNP5 was also tested, as it belonged to the same LD block tagged with SNP6, which showed significant association results (described later). All the SNPs identified were common with MAF > 0.15 and in HWE (P= 0.198–0.965) among the parents (n = 328) of the highly myopic families under study. SNP1, which was not documented in the HapMap, was in strong LD with SNP2 (r 2 = 0.93), but not with any other SNPs (Table 2) . SNP4, -5, and -6 were in tight LD to each other with r 2 > 0.85, which could be defined as a haplotype block. Meanwhile, SNP8 showed evidence of LD with the SNPs in this block (r 2 > 0.75). The remaining SNPs (SNP3, -7, and -9) showed evidence of recombination with each other and all other SNPs (Table 2)
Genetic Association Study
No significant association of the global statistics was found for the MS marker located in the 5′ upstream region (P > 0.05). Meanwhile, all single alleles of the MS marker did not show significant association except allele (CA)17 (P = 0.0393), which obviously did not survive the correction for multiple comparisons. 
For single SNP marker analysis, no significant association of SNP1, -2, -3, -7, and -9 with high myopia was found under all the genetic models tested. For SNP4 and -8, significant association was observed (P = 0.0107 for SNP4 and P = 0.0202 for SNP8) in the dominant/recessive models only. SNP6 also showed significant association under additive model (P = 0.0460). However, these results became statistically insignificant after the correction for multiple comparisons. On the other hand, under the dominant/recessive models, SNP5 and -6 showed significant association with a global P = 0.0014 for SNP5 and P = 0.0011 for SNP6. The global statistics were still statistically significant even after FDR-based correction for multiple comparisons with the cutoff P = 0.0048 (number of comparisons, 21). For SNP6, the T allele showed significantly increased transmission under the dominant model (z = 3.598, P = 0.0003), whereas the G allele exhibited significantly reduced transmission under the recessive model (z = −3.598, P = 0.0003). Similar preferential transmission was observed in SNP5. It is interesting to note the reciprocal relationship for bi-allelic SNP markers (Table 3)
Analysis of SNP6 with GenAssoc gave a GRR of 2.57 (95% CI = 1.46–4.53, P = 0.0011) for genotype G/T and 2.22 (95% CI = 1.11–4.47, P = 0.025) for T/T with reference to G/G. That similar GRRs were obtained for G/T and T/T was consistent with the increased transmission of allele T in the dominant model. Meanwhile, SNP5 showed a GRR of 2.48 (95% CI = 1.42–4.33, P = 0.001) for genotype A/G and 1.91 (95% CI = 0.97–3.76, P = 0.063) for A/A with reference to G/G. This was also compatible with the results of increased transmission of allele A under the dominant model. 
For haplotype analysis of four tSNPs (SNP2, -6, -7, and -9), mild significant association was found under dominant and recessive genetic models (P = 0.0143 under the dominant and P = 0.0227 under the recessive model) and still survived the multiple comparisons correction of FDR (Table 4)
According to the tSNP data of phase II HapMap, SNP5, -6, and -8 was in tight LD and could be considered as a “proxy set.” 47 Single marker analysis of SNP5 and -6 showed the most significant association with high myopia trait (Table 3) . Therefore, haplotype analysis of these three SNPs was also performed. For the sake of easy discussion, a 0 is inserted when a particular SNP is not involved in the subhaplotype analysis. In the additive model, significant association was demonstrated for haplotypes of SNP5-SNP6-SNP8 (χ2 = 11.401, global P = 0.0033) and subhaplotype of 0-SNP6-SNP8 (χ2 = 12.428, global P = 0.0061). Both probabilities remained significant after correction for multiple testing based on FDR (Table 5) . When individual haplotypes were considered, significantly increased transmission was found for the most common haplotypes that carried the major alleles of the component SNPs involved: A-T-G (z = 2.151, P = 0.0315), A-T-0 (z = 2.136, P = 0.0327; Table 5 ). In comparison, in the dominant model, all analyses of haplotypes and subhaplotypes showed significant association: global P = 0.0004 for haplotypes of SNP5-SNP6-SNP8, P = 0.0012 for SNP5-SNP6–0, P = 0.0001 for 0-SNP6-SNP8, and P = 0.0070 for SNP5–0-SNP8. The association was still significant when the global statistics were adjusted for multiple comparisons based on FDR. When individual haplotypes were considered in the dominant model, significantly increased transmission was consistently found for the most common haplotypes that carried the major alleles of the component SNPs involved: A-T-G (z = 3.496, P = 0.0005), A-T-0 (z = 3.600, P = 0.0003), 0-T-G (z = 3.262, P = 0.0011), and A-0-G (Z = 2.674, P = 0.0075; Table 5 ). In the recessive model, no association of haplotypes was found except for the subhaplotypes of SNP5 and -6 (χ2 = 8.119, global P = 0.0173). Significantly reduced transmission (z = −2.661, P = 0.0078) was observed for the subhaplotype G-G-0 (Table 5)
Discussion
LD Analysis of the Selected SNP Markers in Parental Samples
Selecting tSNPs for association test on the basis of the appropriate LD pattern can greatly reduce the genotyping burden while maintaining sufficient power. 48 Using the phase II HapMap data, we selected four tSNPs (SNPs 2, 6, 7, and 9) in the PAX6 locus for association test with a threshold of r 2 > 0.8, which could give reasonable efficiency and power. 40 These four tSNPs can serve as proxies for other SNPs in the PAX6 gene. We also screened the variants in the potential functional regions of the PAX6 gene in order not to miss any variants of functional significance that may be directly disease causative. 23 49 The Han Chinese population used in the HapMap is the northern Chinese (from the Beijing area), who have more mixed genetic background and may be genetically different from the southern Han Chinese. 50 Therefore, LD pattern for the tSNPs suggested by the HapMap and the SNPs identified in our samples was estimated using the parental data of the nuclear families. As expected, the four tSNPs were not in LD with each other. SNP4, which is not documented in the HapMap data for the Chinese population, was in strong LD, with SNP5, -6, and -8 with SNP7 inserted as a “hole“ (Table 2) . 51 The allele and genotype frequencies of the SNPs in the parental group were similar to those from the Chinese HapMap data (P > 0.05; detailed data not shown). Meanwhile, the frequencies of haplotypes consisting of SNP5, -6, and -8 are also similar to those in the Chinese HapMap data calculated by Haploview (data not shown). Although our LD data obtained in the highly myopic nuclear families may be somewhat biased to myopia background rather than nonselected or randomized, the LD pattern established is generally comparable to that defined by the HapMap in Chinese and even in Caucasian samples, demonstrating the usefulness and transferability of the HapMap data across populations. 52 Nevertheless, SNP3 could have been missed if not tested individually, suggesting the complexity of the LD pattern in specific genome regions studied and the usefulness of our LD estimation. 
Genetic Association Analysis
Genetic factors were found to play a role in the occurrence of myopia. 11 12 13 14 15 The profile of genetic and/or environmental components for myopic subjects is highly variable. Using typical high myopia defined as < −6 D, which can be taken as an extreme phenotype, could enhance the efficiency of association test and provide greater statistical power for association analysis of myopia candidate genes. 53 The age at onset in all affected subjects in our study was earlier than 12 years (Table 1) , suggesting the likely involvement of a hereditary background for myopia. Besides, AXL measures were also collected in subject recruitment, and the minimum value was 24.80 mm for the affected siblings in this study. Myopia is caused by delicate change in refractive error, with ocular components of AXL, ACD, CP, and LT all being involved. 39 The underlying profile of genetic and/or environmental components for myopic subjects is highly variable. 20 This could confound the association study for myopia. The most significant correlation between AXL and refractive error (Table 1)demonstrated that the posterior axial myopia was studied and the heterogeneity due to other ocular components was considerably decreased. This would give a more homogeneous phenotype. 39 54 Moreover, typical high myopia is mainly due to the elongation of ocular AXL, and inheritance has significant impact only on AXL but not on ACD or CP. 19 54 Therefore, strict clinical assessment of high myopia is crucial for minimizing the disease heterogeneity which may confound the association study. In other words, the enhanced homogeneity of the phenotype of our affected subjects enhanced the validity of association test. 55  
A recent study found linkage of myopia to a chromosomal region (11p13) where PAX6 was directly underneath the highest linkage peak, but no significant association could be demonstrated in the quantitative TDT analysis. The refractive error of the subjects used in Hammond et al. 37 study ranged from −12.12 to + 7. 25 D with MSE = +0.39 ± 2.38 D, and was taken as a continuous trait. In comparison, we studied highly myopic subjects with extreme phenotypes (MSE = −11.99 ± 3.60 D) and considered myopia as a qualitative trait. 
The single marker FBAT analysis of SNP6 showed the most significant association with high myopia, which remained significant after the FDR correction under the dominant/recessive models (Table 4) . GRR estimates also indicated that the genotypes G/T and T/T were risk factors for susceptibility to high myopia with reference to G/G (GRR = 2.57, P = 0.0011 for G/T and GRR = 2.22, P = 0.025 for T/T). 
Haplotype analysis of four tSNPs only showed moderate significant association (Table 4) . This was not unexpected because of the high haplotype diversity or linkage equilibrium of the four tSNPs covering the whole PAX6 gene region. 56 Haplotype association test for SNP5, -6, and -8 was performed to detect the subtle association of the potential variants that may exist in this haplotype background. The global haplotype analysis showed strong significant association, particularly under the dominant model (Table 5) . It is noticeable that haplotypes carrying the T allele of SNP6 always showed significantly increased transmission under the additive and dominant models, whereas haplotypes carrying the G allele of SNP6 showed reduced transmission under the recessive model. This is in line with the single marker FBAT results for SNP6, indicating a consistently increased transmission of allele T to highly myopic siblings. Haplotype analysis of the three tightly linked SNPs is powerful because of the low haplotype diversity and thus low degree of freedom (Table 5) . 57  
That 56 families had at least one highly myopic parent and 88 families had family history of myopia demonstrates an obvious hereditary component of high myopia. However, the other 76 families with nonmyopic parents also highlight the high complexity of the genetic network underlying nonsyndromic high myopia. 20 The complicated genetic context of high myopia may also explain the mild significance of association detected. On the other hand, misspecification of a genetic model for a disease could substantially decrease the power of association tests. 58 We used all three genetic models with corrections for multiple comparisons because the exact inheritance mode of high myopia is still not fully understood. Nevertheless, previous studies suggested that the inheritance of nonsyndromic high myopia might be autosomal dominant and highly penetrant. 9 22 Therefore, it seems reasonable that the most significant association results were always observed under the dominant model no matter whether single or multimarker FBAT analysis was performed (Tables 3 5)
It is intriguing to note that SNP5, -6, and -8 were all not investigated in the study by Hammond et al. 37 SNP4 was indeed tested as one of the tSNPs and also noted to be in a low-LD region. 37 Our LD data indicated that SNP4, SNP5, -6, and -8 formed an LD block and were not in LD with any other SNPs (Table 2) . Testing SNP4 could provide the information about the potential SNPs in LD with it. Nevertheless, the power of the association test with SNP4 for the variants nearby was not perfect (r 2 < 1), and SNP5 and -6, given the most significant results in our study, were not directly tested in the previous work. 37 Our results suggest that the variants of interest for myopia susceptibility may exist in proximity to SNP5 and SNP6 or their implicated haplotype. Meanwhile, our data showed no evidence of association for the 5′ region of the PAX6 gene. Other flanking genes, such as ELP4 (elongation protein 4 homologue) located in the 3′ downstream region, may be worthy of further investigation because the causal variants may be located within adjacent genes. 59 On the other hand, our data also demonstrated the great usefulness of the HapMap data for genetic marker selection. 47  
The PAX6 protein is a critical and sequence-conserved factor in human development, and the PAX6-mutant phenotypes are usually catastrophic. 33 PAX6 expression dosage and splicing variation are also critical in eye or brain development. 35 60 The potential causal variants of the PAX6 gene underlying nonsyndromic high myopia, if any, are more likely located in the regulatory regions, imposing a “fine-tuning“ mechanism on the PAX6 expression. Several regulatory binding sites have been identified in the PAX6 gene region including intronic or even downstream regions. 59 61 62 However, no common SNPs in the promoter or regulatory regions were identified in our random Chinese subjects (http://www.ensembl.org). 63 we used genome annotation (ElDorado; Genomatix GmbH, Munich, Germany) to search the PAX6 locus for potential transcription factor binding sites of the PAX6 gene. Results suggested that only SNP6 was located in the regulatory binding site of STAT6, a member of the STAT family of transcription factors, 64 which, however, was not experimentally verified. The potential SNPs of interest might be located in the regulatory binding sites in vicinity of the 3′ region. Further understanding of the complicated regulatory mechanisms of the PAX6 gene will also help in finding the potential causative polymorphisms. 
It is tempting to test the case–control association in parental subjects since there were highly myopic (spherical equivalent [SE] ≤ −6.0 D OU, n = 64) and emmetropic parents (SE ≥ −0.75 D OU, n = 208) in our study. A χ2 test showed weakly significant results for SNP6 (P = 0.0478, odds ratio = 1.32, 95% CI = 1.12–1.53 with reference to the G allele) and SNP8 (P = 0.0349, odds ratio = 1.53, 95% CI =1.12–1.93 with reference to the A allele). Haplotype A-T-G of SNP5-SNP6-SNP8 also showed borderline association (χ2 = 3.683, P = 0.0550), whereas subhaplotype 0-T-G showed a moderate significant association (X 2 = 4.762, P = 0.0291). Although the significance level is not strong enough to draw a clear conclusion and the sample size was also small, which caused a loss of power in the association test, the results of the case–control test for the parental group provide support for the findings of FBAT. The phenotype match between case and control groups may not be stringent as only refraction status was measured for the parents. However, to avoid the genetic background overlap between the case and control groups, parents with moderate myopia (−6.0 D < SE < −0.75 D) were not involved and thus could help to enhance the power of association test. 
It is intriguing to note several very recent studies reporting conflicting relationships between PAX6 and myopia. Two studies (one American and one British) reported negative association of PAX6 with common myopia, 65 66 whereas studies in Australian and Taiwan Chinese populations suggested association for PAX6 with high myopia. 67 68 It is not surprising that the conclusions of these recent reports are equivocal due to the complex etiology of myopia. In particular, the researchers in Australia reported association of some specific mutations in the PAX6 gene with highly myopic subjects. 67 More interesting, the researchers in Taiwan reported strong association of extremely high myopia with PAX6 in a small case–control sample, whereas no association was found if the analysis also included subjects with less extreme myopia. 68 Of interest, the positive SNP (rs667773) reported in their study was also located in the 3′ region of the PAX6 gene (intron 9), where potential variants responsible for susceptibility to high myopia might exist as discussed earlier. Therefore, we also tested SNP rs667773 in our high-myopia samples. The MAF was 0.198 (T) and the genotypes were in HWE (P = 0.4274) for the parental group. This SNP showed the evidence of LD with SNP3 and -9 (r 2 = 0.67 and r 2 = 0.86, respectively), but was not in LD with any other SNPs. FBAT analysis showed no association for SNP rs667773 in the additive model (P = 0.1582) and only weak association (P = 0.0458) in the dominant/recessive models. The results were not surprising, as rs667773 did not belong to the LD block formed by SNP4, -5, -6, and SNP8, which were found to be significant in our study. Our data did not support the association of rs667773 with high myopia. A similar conclusion was reached when a meta-analysis was performed for rs667773 to pool the results of the case–control study conducted in Taiwan and our family-based study with the method recently proposed by Kazeem and Farrall. 69 The combined odds ratio for the risk allele T was 1.16 (95% CI = 0.87–1.54; P = 0.3226) if the cutoff threshold refractive error was −6.0 D and was 1.002 (95% CI = 0.71–1.41; P = 0.9889) if the cutoff refractive error was −10.0 D. A preliminary attempt was also made to conduct meta-analysis for the few PAX6 SNPs that were reported in two or more studies, but failed. The data could not be pooled for meta-analysis because of the data type reported (myopia as a qualitative trait or a quantitative trait) and inability to extract the essential information from the reports. 
In summary, based on previous linkage evidence, 37 this study investigated the association between PAX6 polymorphisms and high myopia in a group of southern Han Chinese high-myopia families. The strategy of marker selection based on the HapMap data for association test is logical and is further confirmed by characterizing the LD patterns in our samples. We also tested the potential functional variations identified in the Han Chinese population. The results of this work suggest that the polymorphisms in the PAX6 gene may be associated with typical early-onset high myopia in the southern Han Chinese population and the 3′ region of the PAX6 gene may contain sequence variations affecting susceptibility to high myopia. Further investigation with larger sample size and more markers will be instructive in elucidating the functional polymorphisms responsible for high myopia development. Replication with independent sample sets from different populations is also needed to validate the results of this study. 
 
Table 4.
 
Haplotype Association Tests of the Four tSNPs for the PAX6 Gene by FBAT Analysis with High Myopia as a Qualitative Trait
Table 4.
 
Haplotype Association Tests of the Four tSNPs for the PAX6 Gene by FBAT Analysis with High Myopia as a Qualitative Trait
Haplotype HF Additive Model* Dominant Model* Recessive Model*
SNP2 rs3026354 SNP6 rs3026393 SNP7 rs1506 SNP9 rs662702 n z Score P Global Statistic n z Score P Global Statistic n z Score P Global Statistic
T G A G 0.427 100 −1.379 0.1680 df = 6, χ2 = 12.279, P = 0.0560 69 0.071 0.9430 df = 6, χ2 = 15.904, P = 0.0143 , † 52 2.445 0.0145 df = 2, χ2 = 7.571, P = 0.0227 , †
C T T G 0.224 76 −0.192 0.8478 68 0.253 0.8004 16 −1.000 0.3173
T T T A 0.163 61 1.269 0.2043 57 1.336 0.1817 10 0.513 0.6080
Table 5.
 
Haplotype Association Tests for SNPs rs3026390, rs3026393, and rs12421026 by FBAT Analysis with High Myopia as a Qualitative Trait
Table 5.
 
Haplotype Association Tests for SNPs rs3026390, rs3026393, and rs12421026 by FBAT Analysis with High Myopia as a Qualitative Trait
Haplotype HF Additive Model* Dominant Model* Recessive Model*
SNP5 rs3026390 SNP6 rs3026393 SNP8 rs12421026 n z Score P Global Statistic n z Score P Global Statistic n z Score P Global Statistic
A T G 0.478 90 2.151 0.0315 df = 2, χ2 = 11.401, P = 0.0033 , † 54 3.496 0.0005 df = 2, χ2 = 15.423, P = 0.0004 , † 56 0.067 0.9464 df = 2, χ2 = 6.285, P = 0.0432
G G A 0.427 92 −1.089 0.2763 59 0.532 0.5948 51 2.38 0.0173
A T 0 0.498 92 2.136 0.0327 df = 3, χ2 = 8.687, P = 0.0338 55 3.600 0.0003 df = 3, χ2 = 15.849, P = 0.0012 , † 58 −0.067 0.9468 df = 2, χ2 = 8.119, P = 0.0173 , †
G G 0 0.457 92 −1.592 0.1114 59 −0.005 0.9960 51 2.661 0.0078
0 T G 0.487 94 1.943 0.0521 df = 3, χ2 = 12.428, P = 0.0061 , † 56 3.262 0.0011 df = 3, χ2 = 20.875, P = 0.0001 , † 58 0.000 1.0000 df = 2, χ2 = 5.435, P = 0.0660
0 G A 0.438 97 −0.972 0.3311 63 0.509 0.6106 54 2.177 0.0295
A 0 G 0.498 95 1.557 0.1194 df = 2, χ2 = 6.035, P = 0.0489 56 2.674 0.0075 df = 2, χ2 = 9.913, P = 0.0070 , † 59 −0.066 0.9478 df = 2, χ2 = 6.481, P = 0.0391
G 0 A 0.432 94 −0.993 0.3205 61 0.651 0.5148 51 2.380 0.0173
Table 1.
 
Clinical and Demographic Information of Highly Myopic Siblings (n = 170)*
Table 1.
 
Clinical and Demographic Information of Highly Myopic Siblings (n = 170)*
Age at entry ± SD (y) 20.34 ± 11.49
Sex ratio (male/female), † 88/82 (P > 0.05)
Age at onset of myopia ± SD (y) 7.09 ± 2.87
Families with no myopic parents 65
Families with 1 myopic parent 64
Families with 1 highly myopic parent, ‡ 46
Families with 2 myopic parents 35
Families with 2 highly myopic parents, ‡ 9
MSE ± SD (D) −11.99 ± 3.60
 Range of spherical diopter −6.00 to −22.50
 Range of cylinder diopter 0 to −4.25
AXL ± SD (mm) 27.61 ± 1.63, (r = −0.60, P < 0.001)
CP ± SD (D) 43.37 ± 1.42, (r = −0.21, P = 0.015)
ACD ± SD (mm) 3.70 ± 0.29, (r = −0.38, P < 0.001)
LT ± SD (mm) 3.63 ± 0.25, (r = −0.10, P = 0.253)
Figure 1.
 
LD Analysis for PAX6 SNPs in Parents of the High-Myopia Nuclear Families (n = 328)*
 
* The number in each cell is the r 2 value for the pair-wise LD measure calculated with the Haploview software.
 
The r 2 values in bold type denote evidence of LD with r 2 > 0.75.
Figure 1.
 
LD Analysis for PAX6 SNPs in Parents of the High-Myopia Nuclear Families (n = 328)*
 
* The number in each cell is the r 2 value for the pair-wise LD measure calculated with the Haploview software.
 
The r 2 values in bold type denote evidence of LD with r 2 > 0.75.
Table 3.
 
Single-Locus Association Tests for the PAX6 Gene by FBAT Analysis with High Myopia as a Qualitative Trait
Table 3.
 
Single-Locus Association Tests for the PAX6 Gene by FBAT Analysis with High Myopia as a Qualitative Trait
Marker MS SNP1 rs1894620 SNP2 rs3026354 SNP3 rs3026371 SNP4 rs2239789 SNP5 rs3026390 SNP6 rs3026393 SNP7 rs1506 SNP8 rs1241026 SNP9 rs662702
Allele (GA)11–34 G C T C G A A T A G T G A T G A G A
AF* 0.728 0.272 0.731 0.269 0.790 0.210 0.525 0.475 0.528 0.472 0.507 0.493 0.537 0.463 0.534 0.466 0.804 0.196
FBAT: Additive Model , †
N 1–84 83 83 99 99 96 96 98 98 95 95 99 99 98 98 96 96 73 73
z Score −1.800–1.706 0.572 −0.572 0.583 −0.583 −0.914 0.914 1.432 −1.432 1.622 −1.622 1.995 −1.995 −1.074 1.074 1.031 −1.031 −1.572 1.572
P 0.0719–1.0000 0.5673 0.5673 0.560 0.5600 0.3608 0.3608 0.1521 0.1521 0.1048 0.1048 0.0460 0.0460 0.2829 0.2829 0.3026 0.3026 0.1158 0.1158
Global statistic df = 17, χ2 = 25.509, P = 0.0839 df = 1, χ2 = 0.327, P = 0.5672 df = 1, χ2 = 0.340, P = 0.5600 df = 1, χ2 = 0.835, P = 0.3608 df = 1, χ2 = 2.051, P = 0.1521 df = 1, χ2 = 2.263, P = 0.1048 df = 1, χ2 = 3.981, P = 0.0460 df = 1, χ2 = 1.153, P = 0.2829 df = 1, χ2 = 1.063, P = 0.3026 df = 1, χ2 = 2.473, P = 0.1159
FBAT: Dominant Model , †
N 1–87 41 92 41 88 30 86 85 91 83 91 84 87 86 91 82 90 24 81
z Score −1.584–2.061 1.763 0.274 1.348 0.056 1.778 1.783 2.899 0.496 3.472 0.726 3.598 0.397 0.169 1.843 2.582 0.768 0.788 2.100
P 0.0393–1.0000 0.0779 0.7841 0.1776 0.9552 0.0754 0.0745 0.0037 0.6198 0.0005 0.4681 0.0003 0.6914 0.8658 0.0654 0.0098 0.4423 0.4308 0.0358
Global statistic df = 17, χ2 = 20.092, P = 0.2696 df = 2, χ2 = 3.226, P = 0.1993 df = 2, χ2 = 1.830, P = 0.4006 df = 2, χ2 = 6.432, P = 0.0401 df = 2, χ2 = 9.075, P = 0.0107 df = 2, χ2 = 13.208, P = 0.0014 , ‡ df = 2, χ2 = 13.629, P = 0.0011 , ‡ df = 2, χ2 = 3.577, P = 0.1672 df = 2, χ2 = 7.807, P = 0.0202 df = 2, χ2 = 5.192, P = 0.0746
FBAT: Recessive Model , †
N 0–28 91 41 87 41 86 30 91 85 91 83 87 84 91 86 90 82 81 24
z Score −1.414–1.095 −0.274 −1.763 −0.056 −1.348 −1.783 −1.778 −0.496 2.899 −0.726 3.472 −0.397 3.598 −1.843 −0.169 −0.768 2.582 −2.100 −0.788
P 0.1573–0.2737 0.7871 0.0779 0.9552 0.1776 0.0745 0.0754 0.6198 0.0037 0.4681 0.0005 0.6914 0.0003 0.0654 0.8658 0.4423 0.0098 0.0358 0.4308
Global statistic df = 2, χ2 = 2.995, P = 0.2237 df = 2, χ2 = 3.226, P = 0.1993 df = 2, χ2 = 1.830, P = 0.4006 df = 2, χ2 = 6.432, P = 0.0401 df = 2, χ2 = 9.075, P = 0.0107 df = 2, χ2 = 13.208, P = 0.0014 , ‡ df = 2, χ2 = 13.629, P = 0.0011 , ‡ df = 2, χ2 = 3.577, P = 0.1672 df = 2, χ2 = 7.807, P = 0.0202 df = 2, χ2 = 5.192, P = 0.0746
WongTY, FosterPJ, HeeJ, et al. Prevalence and risk factors for refractive errors in adult Chinese in Singapore. Invest Ophthalmol Vis Sci. 2000;41:2486–2494. [PubMed]
DandonaR, DandonaL. Refractive error blindness. Bull World Health Organ. 2001;79:237–243. [PubMed]
CurtinBJ. The Myopias: Basic Science and Clinical Management. 1985;Harper & Row Philadelphia.
SawSM, GazzardG, Shih-YenEC, ChuaWH. Myopia and associated pathological complications. Ophthalmic Physiol Opt. 2005;25:381–391. [CrossRef] [PubMed]
YapM, WuM, LiuZM, LeeFL, WangSH. Role of heredity in the genesis of myopia. Ophthalmic Physiol Opt. 1993;13:316–319. [CrossRef] [PubMed]
SawSM, ChuaWH, WuHM, YapE, ChiaKS, StoneRA. Myopia: gene-environment interaction. Ann Acad Med Singapore. 2000;29:290–297. [PubMed]
LyhneN, SjølieAK, KyvikKO, GreenA. The importance of genes and environment for ocular refraction and its determiners: a population based study among 20–45 year old twins. Br J Ophthalmol. 2001;85:1470–1476. [CrossRef] [PubMed]
MuttiDO, MitchellGL, MoeschbergerML, JonesLA, ZadnikK. Parental myopia, near work, school achievement, and children’s refractive error. Invest Ophthalmol Vis Sci. 2002;43:3633–3640. [PubMed]
TangWC, YapMK, YipSP. A review of current approaches to identifying human genes involved in myopia. Clin Exp Optom. 2008;91:4–22. [CrossRef] [PubMed]
ZhaoJL, PanXJ, SuiRF, MunozS, SperdutoRD, EllweinLB. Refractive error study in children: results from Shunyi District, China. Am J Ophthalmol. 2000;129:427–435. [CrossRef] [PubMed]
KleinsteinRN, JonesLA, HullettS, et al. Collaborative Longitudinal Evaluation of Ethnicity and Refractive Error Study Group. Refractive error and ethnicity in children. Arch Ophthalmol. 2003;121:1141–1147. [CrossRef] [PubMed]
HammondCJ, SniederH, GilbertCE, SpectorTD. Genes and environment in refractive error: the twin eye study. Invest Ophthalmol Vis Sci. 2001;42:1232–1236. [PubMed]
TeikariJM, O'DonnellJ, KaprioJ, KoskenvuoM. Impact of heredity in myopia. Hum Hered. 1991;41:151–156. [CrossRef] [PubMed]
ChenCY, StankovichJ, ScurrahKJ, et al. Linkage replication of the MYP12 locus in common myopia. Invest Ophthalmol Vis Sci. 2007;48:4433–4439. [CrossRef] [PubMed]
DiraniM, ChamberlainM, ShekarSN, et al. Heritability of refractive error and ocular biometrics: the Genes in Myopia (GEM) twin study. Invest Ophthalmol Vis Sci. 2006;47:4756–4761. [CrossRef] [PubMed]
MuttiDO, SeminaE, MarazitaM, CooperM, MurrayJC, ZadnikK. Genetic loci for pathological myopia are not associated with juvenile myopia. Am J Med Genet. 2002;112:355–360. [CrossRef] [PubMed]
IbayG, DoanB, ReiderL, et al. Candidate high myopia loci on chromosomes 18p and 12q do not play a major role in susceptibility to common myopia. BMC Med Genet. 2004;5:20–30. [PubMed]
MorganI, RoseK. How genetic is school myopia?. Prog Retin Eye Res. 2005;24:1–38. [CrossRef] [PubMed]
LiangCL, YenE, SuJY, et al. Impact of family history of high myopia on level and onset of myopia. Invest Ophthalmol Vis Sci. 2004;45:3446–3452. [CrossRef] [PubMed]
FarbrotherJE, KirovG, OwenMJ, GuggenheimJA. Family aggregation of high myopia: estimation of the sibling recurrence risk ratio. Invest Ophthalmol Vis Sci. 2004;45:2873–2878. [CrossRef] [PubMed]
NaiglinL, ClaytonJ, GazagneC, DallongevilleF, MalecazeF, CalvasP. Familial high myopia: evidence of an autosomal dominant mode of inheritance and genetic heterogeneity. Ann Genet. 1999;42:140–146. [PubMed]
YoungTL, MetlapallyR, ShayAE. Complex trait genetics of refractive error. Arch Ophthalmol. 2007;125:38–48. [CrossRef] [PubMed]
RischN. Searching for genetic determinants in the new millennium. Nature. 2000;405:847–856. [CrossRef] [PubMed]
HanW, YapMK, WangJ, YipSP. Family-based association analysis of hepatocyte growth factor (HGF) gene polymorphisms in high myopia. Invest Ophthalmol Vis Sci. 2006;47:2291–2299. [CrossRef] [PubMed]
LinHJ, WanL, TsaiY, et al. The TGFbeta1 gene codon 10 polymorphism contributes to the genetic predisposition to high myopia. Mol Vis. 2006;12:698–703. [PubMed]
WangIJ, ChiangTH, ShihYF, et al. The association of single nucleotide polymorphisms in the 5′-regulatory region of the lumican gene with susceptibility to high myopia in Taiwan. Mol Vis. 2006;12:852–857. [PubMed]
TangWC, YipSP, LoKK, et al. Linkage and association of myocilin (MYOC) polymorphisms with high myopia in a Chinese population. Mol Vis. 2007;3:534–544.
SpielmanRS, McGinnisRE, EwensWJ. Transmission test for linkage disequilibrium: the insulin gene region and insulin-dependent diabetes mellitus (IDDM). Am J Hum Genet. 1993;3:506–516.
LairdNM, HorvathS, XuX. Implementing a unified approach to family-based tests of association. Genet Epidemiol. 2000;19:S36–S42. [CrossRef] [PubMed]
SimpsonTI, PriceDJ. Pax6: a pleiotropic player in development. Bioessays. 2002;24:1041–1051. [CrossRef] [PubMed]
HillRE, FavorJ, HoganBL, et al. Mouse small eye results from mutations in a paired-like homeobox-containing gene. Nature. 1991;354:522–525. [CrossRef] [PubMed]
JordanT, HansonI, ZaletayevD, et al. The human PAX6 gene is mutated in two patients with aniridia. Nat Genet. 1992;1:328–332. [CrossRef] [PubMed]
HansonIM. PAX6 and congenital eye malformations. Pediatr Res. 2003;54:791–796. [CrossRef] [PubMed]
HeverAM, WilliamsonKA, van HeyningenV. Developmental malformations of the eye: the role of PAX6, SOX2 and OTX2. Clin Genet. 2006;69:459–470. [CrossRef] [PubMed]
SchedlA, RossA, LeeM, et al. Influence of PAX6 gene dosage on development: overexpression causes severe eye abnormalities. Cell. 1996;86:71–82. [CrossRef] [PubMed]
BhatSP, RaynerSA, ChauSC, AriyasuRG. Pax-6 expression in posthatch chick retina during and recovery from form-deprivation myopia. Dev Neurosci. 2004;26:328–335. [CrossRef] [PubMed]
HammondCJ, AndrewT, MakYT, SpectorTD. A susceptibility locus for myopia in the normal population is linked to the PAX6 gene region on chromosome 11: a genomewide scan of dizygotic twins. Am J Hum Genet. 2004;75:294–304. [CrossRef] [PubMed]
HanW, YipSP, WangJ, YapMKH. Using denaturing HPLC for SNP discovery and genotyping, and establishing the linkage disequilibrium pattern for the all- trans-retinol dehydrogenase (RDH8) gene. J Hum Genet. 2004;49:16–23. [CrossRef] [PubMed]
SorsbyA, LearyGA, RichardMJ. Correlation ammetropia and component ammetropia. Vision Res. 1962;2:309–313. [CrossRef]
de BakkerPI, YelenskyR, Pe'erI, GabrielSB, DalyMJ, AltshulerD. Efficiency and power in genetic association studies. Nat Genet. 2005;37:1217–1223. [CrossRef] [PubMed]
ChenX, LevineL, KwokPY. Fluorescence polarization in homogeneous nucleic acid analysis. Genome Res. 1999;9:492–498. [PubMed]
BarrettJC, FryB, MallerJ, DalyMJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics. 2005;21:263–265. [CrossRef] [PubMed]
RabinowitzD, LairdN. A unified approach to adjusting association tests for population admixture with arbitrary pedigree structure and arbitrary missing marker information. Hum Hered. 2000;50:211–223. [CrossRef] [PubMed]
BenjaminiY, HochbergY. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Statist Soc B. 1995;57:289–300.
CordellHJ, ClaytonDG. A unified stepwise regression procedure for evaluating the relative effects of polymorphisms within a gene using case/control or family data: application to HLA in type 1 diabetes. Am J Hum Genet. 2002;70:124–141. [CrossRef] [PubMed]
CordellHJ, BarrattBJ, ClaytonDG. Case/pseudocontrol analysis in genetic association studies: a unified framework for detection of genotype and haplotype associations, gene-gene and gene-environment interactions, and parent-of-origin effects. Genet Epidemiol. 2004;26:167–185. [CrossRef] [PubMed]
International HapMap Consortium. A haplotype map of the human genome. Nature. 2005;437:1299–1320. [CrossRef] [PubMed]
ServiceS, International Collaborative Group on Isolated PopulationsSabattiC, FreimerN. Tag SNPs chosen from HapMap perform well in several population isolates. Genet Epidemiol. 2007;31:189–194. [CrossRef] [PubMed]
CheungVG, SpielmanRS, EwensKG, WeberTM, MorleyM, BurdickJT. Mapping determinants of human gene expression by regional and genome-wide association. Nature. 2005;437:1365–1369. [CrossRef] [PubMed]
YaoYG, KongQP, BandeltHJ, KivisildT, ZhangYP. Phylogeographic differentiation of mitochondrial DNA in Han Chinese. Am J Hum Genet. 2002;70:635–651. [CrossRef] [PubMed]
WallJD, PritchardJK. Assessing the performance of the haplotype block model of linkage disequilibrium. Am J Hum Genet. 2003;73:502–515. [CrossRef] [PubMed]
HuangW, HeY, WangH, et al. Linkage disequilibrium sharing and haplotype-tagged SNP portability between populations. Proc Natl Acad Sci USA. 2006;31:1418–1421.
HsuFC, LiangKY, BeatyTH, BarnesKC. Unified sampling approach for multipoint linkage disequilibrium mapping of qualitative and quantitative traits. Genet Epidemiol. 2002;22:298–312. [CrossRef] [PubMed]
SawSM, CarkeetA, ChiaKS, StoneRA, TanTDH. Component dependent risk factors for ocular parameters in Singapore Chinese children. Ophthalmology. 2002;109:2065–2071. [CrossRef] [PubMed]
HealyDG. Case-control studies in the genomic era: a clinician’s guide. Lancet Neurol. 2006;5:701–707. [CrossRef] [PubMed]
SlagerSL, HuangJ, VielandVJ. Effect of allelic heterogeneity on the power of the transmission disequilibrium test. Genet Epidemiol. 2000;18:143–156. [CrossRef] [PubMed]
SchaidDJ. Evaluating associations of haplotypes with traits. Genet Epidemiol. 2004;27:348–364. [CrossRef] [PubMed]
LiZ, GastwirthJL, GailMH. Power and related statistical properties of conditional likelihood score tests for association studies in nuclear families with parental genotypes. Ann Hum Genet. 2005;69:296–314. [CrossRef] [PubMed]
KleinjanDA, SeawrightA, MellaS, et al. Long-range downstream enhancers are essential for Pax6 expression. Dev Biol. 2006;299:563–581. [CrossRef] [PubMed]
SinghS, MishraR, ArangoNA, DengJM, BehringerRR, SaundersGF. Iris hypoplasia in mice that lack the alternatively spliced Pax6 (5a) isoform. Proc Natl Acad Sci USA. 2002;99:6812–6815. [CrossRef] [PubMed]
MorganR. Conservation of sequence and function in the Pax6 regulatory elements. Trends Genet. 2004;20:283–287. [CrossRef] [PubMed]
KleinjanDA, SeawrightA, ChildsAJ, van HeyningenV. Conserved elements in Pax6 intron 7 involved in (auto)regulation and alternative transcription. Dev Biol. 2004;265:462–477. [CrossRef] [PubMed]
HubbardTJP, AkenBL, BealK, et al. Ensembl 2007. Nucleic Acids Res. 2007;35:D610–D617. [CrossRef] [PubMed]
HebenstreitD, WirnsbergerG, Horejs-HoeckJ, DuschlA. Signaling mechanisms, interaction partners, and target genes of STAT6. Cytokine Growth Factor Rev. 2006;17:173–188. [CrossRef] [PubMed]
MuttiDO, CooperME, O'BrienS, et al. Candidate gene and locus analysis of myopia. Mol Vis. 2007;13:1012–1019. [PubMed]
SimpsonCL, HysiP, BhattacharyaSS, et al. The Roles of PAX6 and SOX2 in Myopia: Lessons from the 1958 British Birth Cohort. Invest Ophthalmol Vis Sci. 2007;48:4421–4425. [CrossRef] [PubMed]
HewittAW, KearnsLS, JamiesonRV, WilliamsonKA, van HeyningenV, MackeyDA. PAX6 mutations may be associated with high myopia. Ophthalmic Genet. 2007;28:179–182. [CrossRef] [PubMed]
TsaiYY, ChiangCC, LinHJ, LinJM, WanL, TsaiFJ. A PAX6 gene polymorphism is associated with genetic predisposition to extreme myopia. Eye. 2008;22:598–599. [CrossRef] [PubMed]
KazeemGR, FarrallM. Integrating case-control and TDT studies. Ann Hum Genet. 2005;69:329–335. [CrossRef] [PubMed]
Figure 1.
 
LD Analysis for PAX6 SNPs in Parents of the High-Myopia Nuclear Families (n = 328)*
 
* The number in each cell is the r 2 value for the pair-wise LD measure calculated with the Haploview software.
 
The r 2 values in bold type denote evidence of LD with r 2 > 0.75.
Figure 1.
 
LD Analysis for PAX6 SNPs in Parents of the High-Myopia Nuclear Families (n = 328)*
 
* The number in each cell is the r 2 value for the pair-wise LD measure calculated with the Haploview software.
 
The r 2 values in bold type denote evidence of LD with r 2 > 0.75.
Table 4.
 
Haplotype Association Tests of the Four tSNPs for the PAX6 Gene by FBAT Analysis with High Myopia as a Qualitative Trait
Table 4.
 
Haplotype Association Tests of the Four tSNPs for the PAX6 Gene by FBAT Analysis with High Myopia as a Qualitative Trait
Haplotype HF Additive Model* Dominant Model* Recessive Model*
SNP2 rs3026354 SNP6 rs3026393 SNP7 rs1506 SNP9 rs662702 n z Score P Global Statistic n z Score P Global Statistic n z Score P Global Statistic
T G A G 0.427 100 −1.379 0.1680 df = 6, χ2 = 12.279, P = 0.0560 69 0.071 0.9430 df = 6, χ2 = 15.904, P = 0.0143 , † 52 2.445 0.0145 df = 2, χ2 = 7.571, P = 0.0227 , †
C T T G 0.224 76 −0.192 0.8478 68 0.253 0.8004 16 −1.000 0.3173
T T T A 0.163 61 1.269 0.2043 57 1.336 0.1817 10 0.513 0.6080
Table 5.
 
Haplotype Association Tests for SNPs rs3026390, rs3026393, and rs12421026 by FBAT Analysis with High Myopia as a Qualitative Trait
Table 5.
 
Haplotype Association Tests for SNPs rs3026390, rs3026393, and rs12421026 by FBAT Analysis with High Myopia as a Qualitative Trait
Haplotype HF Additive Model* Dominant Model* Recessive Model*
SNP5 rs3026390 SNP6 rs3026393 SNP8 rs12421026 n z Score P Global Statistic n z Score P Global Statistic n z Score P Global Statistic
A T G 0.478 90 2.151 0.0315 df = 2, χ2 = 11.401, P = 0.0033 , † 54 3.496 0.0005 df = 2, χ2 = 15.423, P = 0.0004 , † 56 0.067 0.9464 df = 2, χ2 = 6.285, P = 0.0432
G G A 0.427 92 −1.089 0.2763 59 0.532 0.5948 51 2.38 0.0173
A T 0 0.498 92 2.136 0.0327 df = 3, χ2 = 8.687, P = 0.0338 55 3.600 0.0003 df = 3, χ2 = 15.849, P = 0.0012 , † 58 −0.067 0.9468 df = 2, χ2 = 8.119, P = 0.0173 , †
G G 0 0.457 92 −1.592 0.1114 59 −0.005 0.9960 51 2.661 0.0078
0 T G 0.487 94 1.943 0.0521 df = 3, χ2 = 12.428, P = 0.0061 , † 56 3.262 0.0011 df = 3, χ2 = 20.875, P = 0.0001 , † 58 0.000 1.0000 df = 2, χ2 = 5.435, P = 0.0660
0 G A 0.438 97 −0.972 0.3311 63 0.509 0.6106 54 2.177 0.0295
A 0 G 0.498 95 1.557 0.1194 df = 2, χ2 = 6.035, P = 0.0489 56 2.674 0.0075 df = 2, χ2 = 9.913, P = 0.0070 , † 59 −0.066 0.9478 df = 2, χ2 = 6.481, P = 0.0391
G 0 A 0.432 94 −0.993 0.3205 61 0.651 0.5148 51 2.380 0.0173
Table 1.
 
Clinical and Demographic Information of Highly Myopic Siblings (n = 170)*
Table 1.
 
Clinical and Demographic Information of Highly Myopic Siblings (n = 170)*
Age at entry ± SD (y) 20.34 ± 11.49
Sex ratio (male/female), † 88/82 (P > 0.05)
Age at onset of myopia ± SD (y) 7.09 ± 2.87
Families with no myopic parents 65
Families with 1 myopic parent 64
Families with 1 highly myopic parent, ‡ 46
Families with 2 myopic parents 35
Families with 2 highly myopic parents, ‡ 9
MSE ± SD (D) −11.99 ± 3.60
 Range of spherical diopter −6.00 to −22.50
 Range of cylinder diopter 0 to −4.25
AXL ± SD (mm) 27.61 ± 1.63, (r = −0.60, P < 0.001)
CP ± SD (D) 43.37 ± 1.42, (r = −0.21, P = 0.015)
ACD ± SD (mm) 3.70 ± 0.29, (r = −0.38, P < 0.001)
LT ± SD (mm) 3.63 ± 0.25, (r = −0.10, P = 0.253)
Table 3.
 
Single-Locus Association Tests for the PAX6 Gene by FBAT Analysis with High Myopia as a Qualitative Trait
Table 3.
 
Single-Locus Association Tests for the PAX6 Gene by FBAT Analysis with High Myopia as a Qualitative Trait
Marker MS SNP1 rs1894620 SNP2 rs3026354 SNP3 rs3026371 SNP4 rs2239789 SNP5 rs3026390 SNP6 rs3026393 SNP7 rs1506 SNP8 rs1241026 SNP9 rs662702
Allele (GA)11–34 G C T C G A A T A G T G A T G A G A
AF* 0.728 0.272 0.731 0.269 0.790 0.210 0.525 0.475 0.528 0.472 0.507 0.493 0.537 0.463 0.534 0.466 0.804 0.196
FBAT: Additive Model , †
N 1–84 83 83 99 99 96 96 98 98 95 95 99 99 98 98 96 96 73 73
z Score −1.800–1.706 0.572 −0.572 0.583 −0.583 −0.914 0.914 1.432 −1.432 1.622 −1.622 1.995 −1.995 −1.074 1.074 1.031 −1.031 −1.572 1.572
P 0.0719–1.0000 0.5673 0.5673 0.560 0.5600 0.3608 0.3608 0.1521 0.1521 0.1048 0.1048 0.0460 0.0460 0.2829 0.2829 0.3026 0.3026 0.1158 0.1158
Global statistic df = 17, χ2 = 25.509, P = 0.0839 df = 1, χ2 = 0.327, P = 0.5672 df = 1, χ2 = 0.340, P = 0.5600 df = 1, χ2 = 0.835, P = 0.3608 df = 1, χ2 = 2.051, P = 0.1521 df = 1, χ2 = 2.263, P = 0.1048 df = 1, χ2 = 3.981, P = 0.0460 df = 1, χ2 = 1.153, P = 0.2829 df = 1, χ2 = 1.063, P = 0.3026 df = 1, χ2 = 2.473, P = 0.1159
FBAT: Dominant Model , †
N 1–87 41 92 41 88 30 86 85 91 83 91 84 87 86 91 82 90 24 81
z Score −1.584–2.061 1.763 0.274 1.348 0.056 1.778 1.783 2.899 0.496 3.472 0.726 3.598 0.397 0.169 1.843 2.582 0.768 0.788 2.100
P 0.0393–1.0000 0.0779 0.7841 0.1776 0.9552 0.0754 0.0745 0.0037 0.6198 0.0005 0.4681 0.0003 0.6914 0.8658 0.0654 0.0098 0.4423 0.4308 0.0358
Global statistic df = 17, χ2 = 20.092, P = 0.2696 df = 2, χ2 = 3.226, P = 0.1993 df = 2, χ2 = 1.830, P = 0.4006 df = 2, χ2 = 6.432, P = 0.0401 df = 2, χ2 = 9.075, P = 0.0107 df = 2, χ2 = 13.208, P = 0.0014 , ‡ df = 2, χ2 = 13.629, P = 0.0011 , ‡ df = 2, χ2 = 3.577, P = 0.1672 df = 2, χ2 = 7.807, P = 0.0202 df = 2, χ2 = 5.192, P = 0.0746
FBAT: Recessive Model , †
N 0–28 91 41 87 41 86 30 91 85 91 83 87 84 91 86 90 82 81 24
z Score −1.414–1.095 −0.274 −1.763 −0.056 −1.348 −1.783 −1.778 −0.496 2.899 −0.726 3.472 −0.397 3.598 −1.843 −0.169 −0.768 2.582 −2.100 −0.788
P 0.1573–0.2737 0.7871 0.0779 0.9552 0.1776 0.0745 0.0754 0.6198 0.0037 0.4681 0.0005 0.6914 0.0003 0.0654 0.8658 0.4423 0.0098 0.0358 0.4308
Global statistic df = 2, χ2 = 2.995, P = 0.2237 df = 2, χ2 = 3.226, P = 0.1993 df = 2, χ2 = 1.830, P = 0.4006 df = 2, χ2 = 6.432, P = 0.0401 df = 2, χ2 = 9.075, P = 0.0107 df = 2, χ2 = 13.208, P = 0.0014 , ‡ df = 2, χ2 = 13.629, P = 0.0011 , ‡ df = 2, χ2 = 3.577, P = 0.1672 df = 2, χ2 = 7.807, P = 0.0202 df = 2, χ2 = 5.192, P = 0.0746
×
×

This PDF is available to Subscribers Only

Sign in or purchase a subscription to access this content. ×

You must be signed into an individual account to use this feature.

×