Absence of Association between COL1A1 Polymorphisms and High Myopia in the Japanese Population

Abstract

purpose. The collagen type I alpha 1 (COL1A1) gene was recently reported to be associated with high myopia in the Japanese population. To validate this positive association, the tag single-nucleotide polymorphism (tSNP) approach was used.

methods. Eight tSNPs, including rs2075555 and rs2269336 (previously reported to be high myopia–susceptible SNPs in the Japanese), were selected to tag the linkage disequilibrium blocks harboring the COL1A1. These tSNPs were genotyped by using an SNP assay. A total of 427 unrelated Japanese cases with high myopia (axial length, ≥26.50 mm in both eyes; the refraction of the 644 phakic eyes ranged from −5.0 to −36.0 D, with a mean ± SD of −13.61 ± 4.20 D) and 420 Japanese control subjects were recruited. Genotype and allele distributions were compared between the cases and controls by using the χ² test, with multiple testing corrections performed by the permutation test.

results. There was no association noted between high myopia and rs2075555 (P = 0.47, P _c > 0.99) and rs2269336 (P = 0.40, P _c > 0.99). Meta-analysis of a previous Japanese study and new data obtained in a fixed-effect model indicated a mild significant association of high myopia with rs2075555 (odds ratio [OR], 1.19; 95% confidence interval [CI], 1.03–1.38, P = 0.022) and rs2269336 (OR, 1.18; 95% CI, 1.02–1.36, P = 0.026). No significant associations were seen with further tSNPs tests.

conclusions. This study did not replicate the previously reported positive association between COL1A1 and high myopia in the Japanese population, and thus the genetic risk associated with this gene, if any, is weaker than originally reported.

Myopia is a common ocular disorder that is found worldwide. The most important contributor to myopic refraction is the axial length of the eyeball (i.e., longer eyes are more myopic), ^{1 2 3} and when the elongation of the eyeball is excessive, the condition is called high myopia. It is well known that high myopia is associated with many ocular complications ⁴ and is one of the major causes of blindness in many developed countries. ^{5 6 7 8 9 10} Thus, the economic and social burden of high myopia is an important public health problem. 

Recent population-based studies have estimated the prevalence of high myopia in the elderly population to be approximately 1% to 5%, ^{2 11 12 13 14 15 16} and this prevalence has been increasing worldwide, especially in the younger East Asian population. ^{17 18 19} One possible explanation for the increase in high myopia in developed countries is a change in lifestyle. It has been reported that environmental factors such as near work and higher education can contribute to the development of high myopia. However, genetic factors also have been reported to be responsible for the development of high myopia ²⁰ (for detailed review, see Refs. ²¹ , ²² ). For example, several twin studies have shown that there is a high heritability of refraction and axial length. ^{23 24 25 26 27 28} There have been many studies in which investigators have attempted to use a genetic approach to identify the susceptible locus or genes for high myopia (for detailed review, see Refs. ²² , ²⁹ , ³⁰ ), with several genes now reported to have an association. ^{31 32 33 34 35 36 37 38 39} However, there are other studies in which the original findings for these genes were not replicated. ^{25 38 40 41 42 43 44 45 46 47 48}  

Many animal studies on myopia have indicated that there is a local control mechanism of eye growth; hyperopic defocus produces signals from the retina through the retinal pigment epithelium and choroid to cause remodeling of the scleral tissue, and the secondary scleral remodeling results in axial elongation (for detailed review, see Refs. ²¹ , ²² , ⁴⁹ , ⁵⁰ ). In mammals, the scleral tissue contains approximately 90% collagen by weight, predominantly type I ⁵¹ (Zorn M, et al. IOVS 1992;33:ARVO Abstract 1811; Norton TT, et al. IOVS 1995;36:ARVO Abstract 3517). Several animal studies have reported that mRNA expression of type I collagen in the sclera is reduced during the development of myopia. ^{52 53} The COL1A1 (collagen type I, alpha 1) gene encodes the pro-α1 chains of type I collagen. This COL1A1 is located on 17q21.33, where a myopia susceptibility locus (MYP5, 17q21-22) has been reported. ⁵⁴ These pathologic, expression, and genetic studies indicated that COL1A1 is a good candidate gene for myopia. In 2007, Inamori et al. ³⁶ reported that the single-nucleotide polymorphisms (SNPs) rs2075555 and rs2269336 in COL1A1 are significantly associated with high myopia in the Japanese population. However, Liang et al. ⁴⁴ reported that the polymorphisms of COL1A1 are not significantly associated with high myopia in the Taiwanese population. 

In the present study, we conducted a systematic case–control study to validate the association between the polymorphisms of the COL1A1 gene (including previously reported susceptible SNPs) and high myopia in the Japanese population. 

Methods

All investigations in this study adhered to the tenets of the Declaration of Helsinki. The Institutional Review Board and the Ethics Committee of the each institute approved the protocols of this study. All the patients were fully informed of the purpose and procedures of this study, and written consent was received from each patient. 

Study Population

A total of 427 unrelated Japanese patients with high myopia (mean age ± SD, 57.6 ± 14.1 years; men/women, 31.4% vs. 68.6%) were recruited from the Center for Macular Diseases of Kyoto University Hospital, Fukushima Medical University Hospital, and the high myopia clinic of Tokyo Medical and Dental University Hospital. All underwent comprehensive ophthalmic examinations, including dilated indirect and contact lens slit lamp biomicroscopy, automatic objective refraction evaluation, and measurement of the axial length by applanation A-scan ultrasound (UD-6000; Tomey, Nagoya, Japan) or partial coherence interferometry (IOLMaster; Carl Zeiss Meditec, Dublin, CA). To be enrolled in the study, the patients with high myopia were required to have an axial length of ≥26.50 mm in both eyes. The axial lengths of the 854 eyes ranged from 26.50 to 36.32 mm (mean ± SD, 29.18 ± 1.68). Among the 854 eyes enrolled, 644 (75.4%) were phakic, 185 (21.7%) were pseudophakic, and 25 (2.9%) were aphakic. The mean refraction of the 644 phakic eyes ranged from −5.00 to −36.00 D (mean ± SD, −13.61 ± 4.20). To check the results in another axial length-based definition of high myopia, a subset with longer axial lengths was also defined. The inclusion criterion for this subset was axial length ≥ 28.00 mm in both eyes. A total of 278 patients were enrolled in this subset. The axial length of the 556 eyes in this subset was 29.95 ± 1.43 mm. There were 394 phakic eyes in this subset, with the refraction ranging from −7.25 to −36.00 D (−15.03 ± 4.14). If subjects had preexisting ocular diseases or a history of ocular surgery, with the exception of cataract surgery, they were excluded from the study. 

As a population-based control, DNA samples from 420 subjects (mean age ± SD, 44.3 ± 12.1 years; men/women, 46.2% vs. 53.8%) were randomly selected from the Pharma SNP Consortium. The cohort had been recruited for previous genomic studies and was regarded as being representative of the general Japanese population. ⁵⁵ All participants were Japanese and none of the subjects had any history of ocular diseases. 

SNP Selection and Genotyping

To replicate the positive association of the SNPs with high myopia that has been reported in a previous Japanese study, we genotyped rs2075555 from intron 11 of the COL1A1, and rs2269336 from the 5′ upstream region of the COL1A1. The associated functions for these two SNPs have yet to be elucidated. To systematically examine the possible association between the polymorphisms of the COL1A1 gene and the high myopic cases, we used the tag SNP (tSNP) approach. The public dbSNP database build 126 and the HapMap database phase 2 release 22 were used to extract the relevant sequencing information for the COL1A1 gene and the genotyping information for the SNPs. Haplotypes and linkage disequilibrium (LD) blocks were inferred by a solid spine of LD with a minimum D′ of 0.8, according to Haploview version 4.0. ⁵⁶ We selected eight tSNPs to tag the LD blocks harbored within and surrounding the COL1A1 gene (Fig. 1A) . Tagging of the LD blocks was based on the software Tagger (http://www.broad.mit.edu/mpg/tagger/ provided in the public domain by the Broad Institute, Massachusetts of Technology, Cambridge, MA), which used a minimum r ² of 0.8 and a minimum minor allele frequency (MAF) of 20% in the Japanese population of the HapMap dataset. It has been reported in a Japanese study that two SNPs (rs2075555 and rs2269336) are high myopia–susceptible polymorphisms. ³⁶ These SNPs were both included within the eight tSNPs. Genomic DNA was extracted from the leukocytes of the peripheral blood and purified (QuickGene-810; Fujifilm, Tokyo, Japan). All the tSNPs were genotyped with an SNP assay (Taqman; Applied Biosystems, Foster City, CA), according to the manufacturer’s instruction. 

Statistics

The statistical power calculation was performed using the module case–control for discrete traits of the Genetic Power Calculator (http://pngu.mgh.harvard.edu/∼purcell/gpc/ provided in the public domain by the Psychiatric and Neurodevelopmental Genetics Unit, Massachusetts General Hospital, Harvard Medical School, Boston, MA). ⁵⁷ For the calculation, the type 1 error rate was set at 0.05 and the prevalence of high myopia in the general population was set at 1%. The HWE for the genotype distributions was examined by using the χ² test in each group. Differences in the observed genotype and allelic frequencies between the cases with high myopia and the control subjects were also examined by the χ² test. For the current experiment, we combined our results for the single SNP analysis of rs2075555 and rs2269336 with the results of a previous Japanese study, ³⁶ in which the Mantel-Haenszel method based on the fixed-effect model was used to elucidate their predisposing effects on high myopia in a larger Japanese population. We performed the meta-analysis using the R software package Meta (http://cran.r-project.org/web/packages/rmeta/index.html/ provided in the public domain by The Comprehensive R Archive Network, hosted by the Department of Statistics and Mathematics, University of Vienna, Austria). 

Differences in the estimated haplotype frequencies between the cases and the controls were also examined by the χ² test. These SNP and haplotype analyses were performed with Haploview ver. 4.0. The multiple testing correction for P (P _c) was performed by the permutation test (number of iterations, 10,000), also in Haploview, ver. 4.0. The level of statistical significance was set at P < 0.05 and P _c < 0.05. 

Results

The distribution of the genotypes for the eight tSNPs among the cases with high myopia and the control subjects were all in HWE (P > 0.05). The results of the genotyping for rs2075555 and rs2269336 in the cases with high myopia and the control subjects are shown in Table 1 . In this study, there were no significant differences noted for the genotype and allelic frequencies for these two SNPs in the COL1A1 gene between the patient and the control cases. The results of the meta-analysis for rs2075555 and rs2269336 are shown in Table 2 . The Mantel-Haenszel method showed the summary odds ratio (OR) to be 1.19 (95% confidence interval [CI], 1.03–1.38; P = 0.022) for rs2075555 and 1.18 (95% CI, 1.02–1.36; P = 0.026) for rs2269336, respectively. When we performed the subset analysis on the 278 cases with the longer axial lengths (≥28.00 mm in the both eyes), no new significant differences were found for rs2075555 and rs2269336 in our own study (data not shown). The summary OR for the meta-analysis using the subset was 1.27 (95% CI, 1.08–1.48; P = 0.0035) for rs2075555 and 1.25 (95% CI, 1.07–1.46; P = 0.0051) for rs2269336, respectively. 

We also performed a systematic tSNP approach to assess the possible association between the COL1A1 and high myopia in Japanese. The distributions of the allelic frequencies for all the eight tSNPs are given in Table 3 . None of the eight tSNPs showed significant differences between the cases with high myopia and the control subjects with regard to the distribution of the genotype and allelic frequency. We also performed a subset analysis on cases with axial lengths ≥ 28.00 mm. However, no new significant differences were found for the subjects in our study (data not shown). 

We identified three haplotype blocks in the COL1A1 gene (Fig. 1B) . The estimated haplotype frequencies in the cases with high myopia and the control subjects are shown in Table 4 . The haplotype frequencies were not significantly different between the patients with high myopia and the control subjects after the multiple testing corrections. Before correction, only one haplotype in block 1 showed a trend for a mildly significant difference in distribution (P = 0.014, P _c = 0.14). However, a haplotype analysis using the subset with axial lengths ≥ 28.00 mm did not show significant results for any of the blocks, even before correction (data not shown). 

In addition, to check the results of our analyses using the same inclusion criteria that were used in a previous Japanese study, ³⁶ we performed another subset analysis on 261 binocular phakic cases with refractions < −9.25 D (mean refraction ± SD, −14.46 ± 3.94 D; mean axial length ± SD, 29.17 ± 1.60 mm). The allelic frequency distributions for all the eight tSNPs in the subset analysis are given in Table 5 . No new significant differences were noted for the genotype and allelic frequencies for rs2075555 and rs2269336. A haplotype in block 1 (the same haplotype described above) showed a trend for a mildly significant different distribution (P = 0.034, P _c = 0.33). However, there were no tSNPs or haplotypes that showed any significant differences after the multiple testing corrections. 

Discussion

The results in this study did not show significant associations with high myopia of the two SNPs of the COL1A1 gene (rs2075555 and rs2269336, which have been reported to be high-myopia–susceptible SNPs in the Japanese population ³⁶ ). A systematic examination using the tSNP approach to access the possible association between the COL1A1 gene and Japanese high myopia also did not find any significant results. The power calculation results that were based on the multiplicative model showed that our own observations rejected the reported ORs of rs2075555 (OR, 1.36) and rs2269336 (OR, 1.31) from the previous Japanese study with 85.9% and 78.7% power, respectively. 

In our study, we defined high myopia by axial length instead of refraction. On the other hand, in the previous Japanese study that was included in our current analyses, they defined high myopia as refraction < −9.25 D. ³⁶ Thus, one possible explanation for the discrepancy that was observed between the previous Japanese study and our own observations might be related to the difference in the way that high myopia was defined. To further examine this possibility, we performed a subset analysis on binocular phakic cases that had refraction < −9.25 D in both eyes, and we found no further significant differences in the present study. High myopia is most commonly defined by refraction. However, corneal curvature and the intraocular lens may also affect the refraction. Among these multiple factors, the axial length is the most important contributor to myopic refraction. ^{1 2 3} Hence, we suggest that the axial length is a more appropriate parameter than refraction when assessing the association between the COL1A1 gene and high myopia. However, our study could not show any significant result whether the axial length or the refraction was chosen as the parameter. We cannot conclude which of the two, axial length or refraction, is the more appropriate parameter to assess the association between the COL1A1 gene and high myopia. 

Another difference between the previous study and our own observations is that we used a population-based control. The prevalence of high myopia in the general population has been estimated to be approximately 1% to 5% in elderly adults. ^{2 11 12 13 14 15 16} Even if the control subjects in our study had no history of ocular diseases, the possibility exists that some of the eyes might have had an axial length ≥ 26.50 mm without the presence of vision threatening complications. If this were the case, this would be a possible explanation for the negative results that we found for our case–control association study. To check the results for a different axial-length–based definition, we also performed a subset analysis on cases with longer axial lengths (≥28.00 mm in both of the eyes). However, no new significant differences were found in the present study. Further subset analyses by redefining the cutoff value of axial length (27.00, 27.50, 28.50, and 29.00 mm) did not show any significant results (data not shown). Thus, we can conclude that the results of the present study did not replicate the previously reported Japanese study, which found significant associations for rs2075555 and rs2269336 with high myopia. 

The results of our meta-analysis suggested that there were mildly significant associations between these two SNPs and high myopia in the Japanese population. However, it should be noted that we combined the data of our own study with the data of a previous Japanese study, a study that was the first to report positive results. ³⁶ There was a potential for publication bias in the first positive study, and indeed, the reported ORs in the first positive studies were higher than most of the results that have been reported for subsequent replication studies. ⁵⁸ Therefore, actual ORs of these SNPs are estimated at up to the ORs that are suggested by the results of the meta-analysis in this study. We concluded that the genetic risk in the COL1A1 gene, if any, is weaker than has been originally reported. 

In conclusion, the present study failed to replicate the positive association between the polymorphisms of the COL1A1 gene and high myopia that has been reported in a prior study involving Japanese subjects. To elucidate whether the COL1A1 gene in the MYP5 locus is associated with high myopia in the Japanese population, additional genetic and molecular biological studies are needed. 

 Supported in part by the Ministry of Education, Culture, Sports, Science and Technology of Japan and by the Japanese National Society for the Prevention of Blindness.

 Submitted for publication June 12, 2008; revised July 28 and August 26, 2008; accepted December 3, 2008.

 Disclosure: H. Nakanishi, None; R. Yamada, None; N. Gotoh, None; H. Hayashi, None; A. Otani, None; A. Tsujikawa, None; K. Yamashiro, None; N. Shimada, None; K. Ohno-Matsui, None; M. Mochizuki, None; M. Saito, None; K. Saito, None; T. Iida, None; F. Matsuda, None; N. Yoshimura, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Figure 1.

 

LD structure across the COL1A1 region and selected tag SNPs. LD blocks were inferred by a solid spine of LD with a minimum D′ of 0.8. (A) LD structure in Japanese samples from the HapMap database. SNPs with MAF > 5% are displayed, with the selected 8 tSNPs shown in boxes. (B) LD structure for the samples obtained in the present study (427 unrelated Japanese cases with high myopia [axial length ≥ 26.50 mm in both eyes] and 420 healthy Japanese controls). Three haplotype blocks were identified. The distribution of the haplotypes from each of the three blocks is shown in Table 4 .

Table 1.

 

Frequencies of Genotypes and Alleles of rs2075555 and rs2269336 in the Current Study

SNP ID*	Genotype									Allele
			Case (%)	Control (%)	P , †	OR (95% CI)		Case (%)	Control (%)	P , †	OR (95% CI)
rs2075555	CC	167 (39.2)	158 (37.7)	0.736	1.17 (0.79–1.75)	C	528 (62.0)	505 (60.3)	0.471	1.07 (0.88–1.31)
	CA	194 (45.5)	189 (45.1)		1.14 (0.77–1.68)	A	324 (38.0)	333 (39.7)
	AA	65 (15.3)	72 (17.2)		1.00 (ref.)
rs2269336	CC	141 (33.8)	125 (29.9)	0.444	1.14 (0.78–1.67)	C	469 (56.2)	453 (54.2)	0.400	1.09 (0.90–1.32)
	CG	187 (44.8)	203 (48.6)		0.93 (0.65–1.33)	G	365 (43.8)	383 (45.8)
	GG	89 (21.3)	90 (21.5)		1.00 (ref.)

 The nucleotides were defined on the forward strand of the reference sequence by dbSNP Build 126.

*  SNP ID in National Center for Biotechnology Information (NCBI; Bethesda, MD) dbSNP Build 126.

†  The nominal probabilities were calculated by the χ² test.

Table 2.

 

Meta-analysis^* of the COL1A1 rs2075555 and rs2269336 in Japanese Subjects with High Myopia

SNP ID, †	Source	Case			Control			OR (95% CI)	P , §
				Subject Number	Risk Allele, ‡ n (%)	Subjects (n)	Risk Allele, ‡ n (%)
rs2075555	Current study	426	528 (62.0)	419	505 (60.3)	1.07 (0.88–1.31)	0.47
	Inamori et al. 36	328	422 (64.3)	326	372 (57.1)	1.36 (1.09–1.70)	0.0071
	Total	754		745		1.19 (1.03–1.38)	0.022
rs2269336	Current study	417	469 (56.2)	418	453 (54.2)	1.09 (0.90–1.32)	0.40
	Inamori et al. 36	329	397 (60.3)	330	354 (53.6)	1.31 (1.06–1.64)	0.014
	Total	746		748		1.18 (1.02–1.36)	0.026

*  This meta-analysis was performed using the Mantel-Haenszel method based on the fixed-effect model.

†  SNP ID in NCBI dbSNP Build 126.

‡  Risk alleles of rs2075555 (A/C) and rs2269336 (C/G) are allele C and allele C, respectively. The nucleotides were defined on the forward strand of the reference sequence by the dbSNP Build 126. We confirmed through personal communication that the nucleotides for rs2269336 (C/G) in a previous study by Inamori et al. ³⁶ were defined on the reverse strand of the reference sequence by dbSNP.

§  The nominal probabilities were calculated by the χ² test.

Table 3.

 

Association of Eight Tagged SNPs of the COL1A1 with High Myopia in the Current Study

SNP ID*	Position, †	Ref, ‡	Var, ‡	Case–Control			OR (95% CI)	P , §
									Allele Counts	Allele Frequencies		Nominal	Corrected
rs2696296	45601230	G	A	427:421, 395:445	0.504, 0.470	1.14 (0.94–1.38)	0.171	0.886
rs1061237	45617774	T	C	469:377, 477:339	0.554, 0.585	0.88 (0.73–1.07)	0.214	0.933
rs2696247	45624902	A	G	589:257, 567:271	0.696, 0.677	1.10 (0.89–1.35)	0.386	0.997
rs2586494	45628154	A	C	407:441, 425:409	0.480, 0.510	0.89 (0.73–1.08)	0.224	0.940
rs2141279	45628463	T	C	270:574, 268:568	0.320, 0.321	1.00 (0.81–1.22)	0.977	1.000
rs2075555	45629290	A	C	324:528, 333:505	0.380, 0.397	0.93 (0.77–1.13)	0.471	1.000
rs2269336	45635355	C	G	469:365, 453:383	0.562, 0.542	1.09 (0.90–1.32)	0.400	0.998
rs2246839	45643395	C	T	321:527, 321:511	0.379, 0.386	0.97 (0.80–1.18)	0.759	1.000

 Axial length ≥ 26.5 mm in both eyes.

*  SNP ID in NCBI dbSNP Build 126.

†  Position of the polymorphism in the reference sequence NT_010783.14.

‡  Ref and Var were the reference and variant nucleotides, respectively, that were defined on the forward strand of the reference sequence by dbSNP.

§  The nominal probabilities were calculated by the χ² test, and the multiple testing corrections were performed by the permutation test (number of iterations = 10,000).

Table 4.

 

Association of Haplotypes across the COL1A1 Region with High Myopia in the Current Study

Haplotype*	Frequency	Case–Control			OR (95% CI)	P , †
				Ratio Counts		Frequencies		Nominal	Corrected
Block 1
ATA	0.489	403.5:450.5, 424.3:415.7	0.472, 0.505	0.88 (0.73–1.06)	0.179	0.889
GCG	0.293	243.0:611.0, 253.0:587.0	0.285, 0.301	0.92 (0.75–1.14)	0.452	0.998
GCA	0.132	130.0:724.0, 93.7:746.3	0.152, 0.112	1.43 (1.08–1.90)	0.014	0.142
GTA	0.062	57.3:796.7, 48.3:791.7	0.067, 0.057	1.18 (0.79–1.75)	0.411	0.997
ATG	0.017	11.8:842.2, 17.3:822.7	0.014, 0.021	0.67 (0.32–1.41)	0.288	0.980
Block 2
AC	0.488	405.6:448.4, 420.6:419.4	0.475, 0.501	0.90 (0.75–1.09)	0.289	0.980
CT	0.315	269.6:584.4, 263.3:576.7	0.316, 0.313	1.01 (0.82–1.24)	0.922	1.000
CC	0.192	175.9:678.1, 150.2:689.8	0.206, 0.179	1.19 (0.94–1.52)	0.156	0.842
Block 3
CC	0.545	474.3:379.7, 448.3:391.7	0.555, 0.534	1.09 (0.90–1.32)	0.369	0.993
AG	0.381	319.3:534.7, 326.4:513.6	0.374, 0.389	0.94 (0.77–1.14)	0.533	0.999
CG	0.067	54.8:799.2, 57.9:782.1	0.064, 0.069	0.93 (0.63–1.36)	0.699	1.000

 Axial length ≥ 26.5 mm in both eyes. The nucleotides were defined on the forward strand of the reference sequence by dbSNP Build 126.

*  Haplotypes and linkage disequilibrium (LD) blocks were inferred by a solid spine of LD with a minimum D′ of 0.8. The LD structure for the Japanese samples in the current study is shown in Figure 1(B) .

†  The nominal probabilities were calculated by the χ² test, with the multiple testing corrections performed by the permutation test (no. of iterations = 10,000).

Table 5.

 

Association of Eight Tagged SNPs of COL1A1 with a Subset of High Myopia in the Current Study

SNP ID*	Position, †	Ref, ‡	Var, ‡	Case–Control			OR (95% CI)	P , §
									Allele Frequencies	Allele Counts		Nominal	Corrected
rs2696296	45601230	G	A	266:250, 395:445	0.516, 0.470	1.20 (0.96–1.49)	0.105	0.743
rs1061237	45617774	T	C	277:239, 477:339	0.537, 0.585	0.82 (0.66–1.03)	0.087	0.640
rs2696247	45624902	A	G	348:168, 567:271	0.674, 0.677	0.99 (0.78–1.25)	0.933	1.000
rs2586494	45628154	A	C	259:259, 425:409	0.500, 0.510	0.96 (0.77–1.20)	0.732	1.000
rs2141279	45628463	T	C	158:358, 268:568	0.306, 0.321	0.94 (0.74–1.19)	0.581	1.000
rs2075555	45629290	A	C	197:323, 333:505	0.379, 0.397	0.92 (0.74–1.16)	0.496	1.000
rs2269336	45635355	C	G	296:218, 453:383	0.576, 0.542	1.15 (0.92–1.43)	0.222	0.956
rs2246839	45643395	C	T	197:323, 321:511	0.379, 0.386	0.97 (0.77–1.22)	0.798	1.000

 Eyes were binocular phakic and had refraction < −9.25 D.

*  SNP ID in NCBI dbSNP Build 126.

†  Position of the polymorphism in the reference sequence NT_010783.14.

‡  Ref and Var are, respectively, reference and variant nucleotides defined on the forward strand of the reference sequence by dbSNP.

§  The nominal probabilities were calculated by the χ² test, and the multiple testing corrections were performed by the permutation test (number of iterations = 10,000).

The authors thank Yasuo Kurimoto for assistance in recruiting the patients.