Molecular Vision 2006; 12:523-531 <http://www.molvis.org/molvis/v12/a62/> Received 20 January 2006 | Accepted 15 May 2006 | Published 22 May 2006

Download Reprint

Primary congenital glaucoma and Rieger's anomaly: extended haplotypes reveal founder effects for eight distinct CYP1B1 mutations

Gabriela Chavarria-Soley,¹ Karin Michels-Rautenstrauss,¹ Francesca Pasutto,¹ David Flikier,² Paul Flikier,³ Sebahattin Cirak,⁴ Bassem Bejjani,⁵ Daniel L. Winters,⁵ Richard A. Lewis,⁷ Christian Mardin,⁶ Andre Reis,¹ Bernd Rautenstrauss¹
 
 

¹Institute of Human Genetics and ⁶Ophthalmologic Department, Friedrich-Alexander-University, Erlangen, Germany; ²Instituto de Cirurgia Ocular, San José, Costa Rica; ³Centro Médico la Visión, San José, Costa Rica; ⁴Department of Pediatrics and Pediatric Neurology, University Hospital Essen, Essen, Germany; ⁵Washington State University Spokane, Spokane, WA; ⁷Baylor College of Medicine, Houston, TX

Correspondence to: Bernd Rautenstrauss, Institute of Human Genetics, Friedrich-Alexander-University, Schwabachanlage 10, 91054 Erlangen, Germany; Phone: +49-9131-8522352; FAX: +49-9131-209297; email: bernd.rautenstrauss@humgenet.uni-erlangen.de

Abstract

Purpose: Mutations in the cytochrome P450 1B1 (CYP1B1) gene are a frequent cause of primary congenital glaucoma (PCG) in different ethnic groups. Cytochrome P450 proteins are monooxygenases, which catalyze many reactions involved in the metabolism of drugs as well as steroids and other lipids. The repeated occurence of several mutations in various ethnic groups raises the question if founder effects or mutation-prone sites in CYP1B1 are the cause for this observation.

Methods: A total of 30 individuals (26 PCG patients, three Rieger's anomaly patients, and one variant carrier), presenting 17 variants in CYP1B1 (15 mutations and two variations) were included in our study. We sequenced the entire genomic region of CYP1B1 and analyzed microsatellites flanking the gene in all individuals and constructed haplotypes for all variations using a combination of single nucleotide polymorphisms and microsatellites.

Results: For the CYP1B1 genomic region, we identified five extended haplotypes associated with 17 variations. These haplotypes were complemented with microsatellite information from the region surrounding this gene. A total of eight CYP1B1 mutations were found more than once, each of them presenting one identical haplotype in different individuals. Six mutations were represented in different ethnic groups.

Conclusions: Our results confirm founder effects for most of CYP1B1 mutations. Most of these mutations must have occurred as unique events in the past.

Introduction

Autosomal recessive primary congenital glaucoma (PCG) has been associated with the GLC3A locus on chromosome 2p21 [1]. The PCG-associated cytochrome P450 1B1 (CYP1B1) gene is located here [2]. Cytochrome P450 proteins are monooxygenases, which catalyze many reactions involved in the metabolism of drugs as well as steroids and other lipids. Although the role of CYP1B1 in congenital glaucoma is not well understood, the protein is probably responsible for the metabolism of compounds that are critical for the developing eye [3].

Mutations in CYP1B1 have been shown to cause PCG, which is a major cause of childhood blindness worldwide. CYP1B1 mutations were first identified in Turkish PCG patients [2]. Subsequently, different mutations have been found in a variety of ethnic groups, including those from Saudi Arabia [4,5], Morocco [6], Slovak Gypsies [7], Indonesia [8], India [9-11], Japan [12,13], Europe [8,14,15], North-, Central-, and South America [16-19]. Interestingly, CYP1B1 mutations are also present in patients with malformations in the anterior chamber of the eye and affected with secondary glaucoma types [20-22]. Rieger's anomaly is one malformation of the anterior segment of the eye and occurs frequently in association with secondary glaucoma [23]. More recently CYP1B1 mutations in the heterozygous state (e.g., E229K and Y81N), have been discussed as risk factors for autosomal dominant primary open angle glaucoma (POAG) [14,24].

The CYP1B1 gene consists of three exons, two of which are coding. Approximately 60 PCG-causing mutations have been identified, including missense and frameshift mutations, as well as small insertions and deletions. Most missense mutations occur in highly conserved functional regions. Several mutations have been reported in the literature repeatedly [2,4-19,25].

To determine whether the repeated occurrence of mutations is explained by a "hot spot" for a mutation or by a founder effect, we constructed extended haplotypes for 30 CYP1B1 variation carriers with a combination of SNPs in the complete CYP1B1 genomic region and microsatellites flanking the gene [17].

Methods

Individuals

Our research program followed the tenets of the Declaration of Helsinki and was approved by the ethics committee of the Medical Faculty of the Friedrich-Alexander-University and the affiliated academic institutions of the authors. Informed consent was obtained from the participants after we explained the nature and possible consequences of the study.

In total, 30 (29 patients and one variation carrier) individuals with CYP1B1 variants were included in the study. Three of them belong to two PCG families originating in Costa Rica (PCG-CR1) and one in Russia (PCG-R1). The two families had a pedigree structure consistent with autosomal recessive inheritance (Figure 1). The index patient of each family was included in our study along with one more individual (II:1) from family PCG-R1 who carries the E229K variant. The PCG patients in family PCG-R1 are compound heterozygous for the c.1064-1076del and c.155insC mutations. The unaffected mother (I:2) of patients II:2 and II:3 carries the c.1064-1076del deletion and the E229K variant in compound heterozygous state.

Two other patients have an affected sibling (PCG-14 and PCG-15), 22 are simplex PCG patients (PCG-1 to PCG-13 and PCG16 to PCG-24), and three are simplex Rieger's anomaly patients (RA-1 to RA-3). All but two individuals carry two CYP1B1 mutations, either in the homozygous or the compound heterozygous state. For patient PCG-6 and patient RA-1, only one heterozygous mutation is known.

Available family members from the simplex cases were included to establish phase. The list of patients with country of origin and the mutations are presented in Table 1.

All patients underwent a complete eye examination, including anterior segment evaluation with slit lamp, fundoscopy, Goldmann applanation tonography, and gonioscopy. Congenital glaucoma was defined as the presence of intraocular pressure higher than 21 mm Hg in both eyes before the age of three years, presence of optic disc cupping, enlarged axial diameter of the globe, and an increased corneal diameter with or without Haab's lines. Rieger's anomaly was defined by the presence of iris hypoplasia, posterior embryotoxon, and anterior synechiae to the prominent Schwalbe's line or posterior cornea.

Sequence analysis

Genomic DNA was prepared from peripheral blood. Twenty primer pairs were used to amplify the genomic region of CYP1B1 and four to amplify the region flanking M4 (primer sequences are available on request). For amplification we used a touchdown PCR program with an annealing temperature decreasing from 65 °C to 55 °C over 9 cycles, followed by 24 cycles with an annealing temperature of 55 °C in a 25 μl mixture (PCR conditions available on request). Sequencing reactions were performed on both strands using the BigDye Terminator Cycle Sequencing Kit v3.1 (Applied Biosystems, Foster City, CA) according to the manufacturer's instructions. The products were analyzed on an ABI Genetic Analyzer 3730 (Applied Biosystems).

Microsatellite analysis

Seven microsatellite markers flanking CYP1B1 [17] were amplified in singleplex reactions in a final reaction volume of 15 μl containing 10 mM Tris, 1.5 mM MgCl₂, 100 μM each dNTP, 0.35 U DNA polymerase (Invitrogen), 7.0 pmol of each primer, and 20 ng of genomic DNA. One of the primers was end-labeled with a fluorescent dye (FAM, TET, or HEX). For amplification we used a touchdown PCR program with an annealing temperature decreasing from 61 °C to 55 °C over 6 cycles, followed by 31 cycles with an annealing temperature of 55 °C. Products were analyzed on an ABI Genetic Analyzer 3100 (Applied Biosystems). Microsatellite 3 could not be properly genotyped in our samples and was therefore excluded from our analysis.

Linkage disequilibrium structure

The Haploview program [26] was used for the construction of the linkage disequilibrium plot with HapMap data [27] and calculating the haplotype frequencies.

Nomenclature

GenBank accession NM_000104 was used as the cDNA reference sequence. The nomenclature recommendations of den Dunnen and Antonarakis [28] were followed. Nucleotide +1 is the A from the ATG-translation initiation codon. For amino acid numbering the translation initiation methionine is considered +1.

Results

The CYP1B1 variations Q42X in patient PCG-4, W434X in patient PCG-7, and c.1033-1035del in patient RA-3 are, to the best of our knowledge, reported here for the first time.

Initially we searched publicly available data for haplotype information on CYP1B1 mutations. Six well-studied intragenic SNPs in CYP1B1 are frequently used to form haplotypes. Five single nucleotide polymorphisms (SNPs) are located in the coding region and one is 12 bp upstream of the first coding exon [19]. The resulting intragenic haplotypes as published for different mutations, and the countries of origin are summarized in Table 1. This analysis clarifies the association of certain haplotypes with known mutations. We also applied modern uniform nomenclature rules to these mutations [28]. Twelve mutations in our patients are associated with the common 5'-CCGGTA-3' SNP haplotype previously discussed by Stoilov [19] and Sena [17]. This haplotype is associated with 54.5% of the CYP1B1 mutations for which haplotype information is available (Table 2). The Y81N, E229K, and E387K variants reside on the 5'-TGTCCA-3' haplotype, and the c.1033-1035del mutation presents the 5'-CCGCCG-3' SNP haplotype. According to public available data, these two less frequent haplotypes are associated with about 9.7% and 7% of CYP1B1 mutations, respectively (Table 2).

We sequenced the complete genomic region in our patients to find SNPs beyond the six previously reported [19], allowing us to construct extended haplotypes for the mutations. We discovered 11 further intronic SNPs in the CYP1B1 genomic region (Table 3) and constructed haplotypes for each mutation with a combination of SNPs and six microsatellites [17] that flank the gene. CYP1B1 is located within one linkage disequilibrium (LD) block (LD block A) comprising approximately 63 kb on chromosome 2 (Figure 2). Microsatellites M1 and M2 are located in LD block A as well. M4 is located in LD block B, while M5 lies in a region of low LD between two LD blocks, and M6 and M7 are together in LD block C. The extensive haplotypes constructed here show that the previously reported 5'-CCGGTA-3' haplotype [19] represents in reality two different ancestral haplotypes (CYP1B1-A and CYP1B1-B), which happen to share some common SNPs. For the 17 mutations included in our study, we found five possible SNP haplotypes in LD block A (not including the microsatellites; Table 3). Eight mutations are associated with the CYP1B1-A haplotype and four with CYP1B1-B. From the three other haplotypes, we found CYP1B1-C is associated with three variants, while CYP1B1-D and CYP1B1-E are associated with just one mutation each (Table 3). We also obtained population frequencies for our haplotypes from HapMap data, which in the genomic region of CYP1B1 include genotypes for 10 of the 17 SNPs of our study. According to HapMap data, seven distinct haplotypes can be detected in Caucasians for the CYP1B1 region (Table 4). The mutations analyzed here are associated with the five most frequent ones.

The LD structure for CYP1B1 explains the observation that mutations are also tightly associated with microsatellites, especially M1 and M2, which are within the same high LD block. Therefore, every mutation is associated with certain M1 and M2 microsatellite alleles and variation observed at these markers is due to mutation rather than recombination. For example, the G61E mutation is associated with alleles 159 and 253 for microsatellites M2 and M1, respectively (Table 3). Microsatellite marker M4, however, varies within mutations. It is located in LD block B, which shows recombination with the LD block A in historical time scales. In order to determine whether variation at M4 was indeed due to recombination or mutation, we analyzed six further SNPs flanking the microsatellite (Table 3).

Eight of the mutations that we studied were found more than once (Table 1). Our extended haplotype reconstruction shows each mutation that is repeated is always associated with only one CYP1B1 haplotype, supporting the hypothesis of a founder effect for each of these mutations. The shared haplotypes extend in many cases to the M4 region, or even further to the more distal microsatellites M6 and M7, located in different LD blocks.

The c.1064-1076del mutation occurs six times in patients with different ethnic origin (Germany, Russia, United States, and Switzerland) and in all instances the individuals share a haplotype up to the M6-M7 block (approximately 160 kb), except for microsatellite M4. At M4 there is some variation (e.g., patients PCG-R1-II:2 and PCG-4 present allele 198 instead of 196). This is probably due to a mutation at the microsatellite locus, not recombination, because the remaining SNP-based haplotype is shared by the chromosomes with this small deletion. Hence our data support a founder effect for the c.1064-1076del variation. This variation is old enough for a mutation to have occurred at a microsatellite locus but not old enough that recombination events within the shared haplotype are observed.

The G61E mutation was present in nine Saudi Arabian patients and one from the United States. In general, they share one haplotype including the M4 region, indicating a common origin. However, a recombination has occurred in two Saudi Arabian patients (PCG-20 and PCG-21). The site of the recombination events between LD blocks A and B fits with a region of LD breakdown and thus historical recombination as seen in HapMap data (Figure 2).

That there are two independent recombinations and there is a large variation in the more distant microsatellite marker M5 suggests that additional recombination/mutation events have occurred. It is evident this is a historically old mutation, which originated outside America and became frequent in Saudi Arabia due to a founder effect.

The R368H, c.1377-1403dup, and R469W mutations were each found only twice. The respective patients share a complete SNP-based ancestral CYP1B1-A haplotype but show variation at the microsatellites, again indicating a common respective founder and additional mutation events in the microsatellites.

The c.1200-1209dup duplication clearly represents an ancestral founder mutation, as patients from the United States, Germany, and Costa Rica share a single haplotype CYP1B1-B in the genomic region including the adjacent regions at microsatellites M4-M7. Only patient RA-2 presents a recombination event between the CYP1B1 and M4 LD blocks. In addition, patient PCG-12 has one mutation event in microsatellite M4. The presence of variations in the microsatellites and recombination detected by means of SNP analysis suggest that this is also a historically old mutation. Further evidence of the founder effect is implied by the observation that four of five patients carrying this small duplication carry the SNP-haplotype 5'-CTATGG-3' for the M4 region (excluding the microsatellite), which according to data obtained from HapMap, has only a frequency of 8% in Caucasians (Table 4).

The three patients with the W57X mutation share the complete CYP1B1-B haplotype as well as the microsatellites on both sides of the gene. This suggests a German origin of the mutation and a later migration to the United States.

Six patients from the United States carrying the E387K mutation share the CYP1B1-C haplotype and have the same haplotype for the M4 block as well. In the M5-M7 block there is some variation for these individuals, again indicating a common founder. The variation pattern observed with most patients sharing either the 180 or 182 base pair allele indicates this is a recent mutation, where probably only one mutation event at microsatellite M6 has occurred in a common ancestor (180 to 182 or vice versa).

The remaining mutations (c.155insC, R355X, c.279delC, Q42X, E229K, Y81N, W434X, c.433-442del, and c.1033-1035del; Table 1) were found only once and thus could not be analyzed for a founder effect.

Discussion

Most CYP1B1 mutations in the Western world are associated with the 5'-CCGGTA-3' haplotype. As shown in Table 2, the mutations found in Japan tend to be present on other haplotypes and the respective mutations differ from those in other populations [13]. Our results show that the 5'-CCGGTA-3' haplotype is actually misleading, as it is in reality part of two different LD-based haplotypes of the seven found in the CYP1B1 genomic region in Caucasians. In our study we found mutations presenting the five most common haplotypes. Some speculation exists over why so many mutations in CYP1B1 present the inner 5'-CCGGTA-3' haplotype; some have proposed that the haplotype could somehow be "mutation prone" [17]. After analyzing the LD structure of the CYP1B1 gene and the haplotype frequency in this region, we propose that the 5'-CCGGTA-3' associated mutations are frequent simply because the 5'-CCGGTA-3' haplotype includes not one, but two, common haplotypes associated with widespread founder mutations.

In family PCG-R1, the healthy mother (I:2) of two affected children (II:2, II:3) is a heterozygous carrier of the c.1064-1076del mutation. Her second CYP1B1 allele carries the E229K variant. This variation was previously thought to be pathogenic [15] and, furthermore, a dominant effect was proposed [14]. The healthy mother (I:2) as compound heterozygous mutation carrier indicates that this variation has the nature of a benign variation. On the other hand, additional modifying factors cannot be excluded at this stage of our investigation.

For patient PCG-6, only one CYP1B1 mutation, Y81N, was found. Similarly, patient RA-2 presented with a single R368H mutation. Other variations in noncoding regions have not yet been ruled out, however, the missense mutation present in patient PCG-6 had been reported previously in heterozygous form in patients with early onset primary open angle glaucoma [24]. Mutations in the CYP1B1 gene have been reported mainly in PCG patients but also in individuals with secondary glaucoma [20-22], Peters' and Rieger's anomaly, and even in heterozygous state in persons with adult glaucoma [24]. There seems to be a broad spectrum of phenotypes associated with homo- and heterozygous CYP1B1 mutations.

In the case of a founder effect the genetic markers on the mutant chromosome would be identical, or almost identical, in every ethnic group. If the recurrence of the mutations is explained by a mutation "hot spot," then we would observe different genetic markers in discrete ethnic groups. Hence our data provide evidence for a founder effect in all CYP1B1 mutations that could be investigated.

Having both microsatellite and SNP data is useful, because these markers mutate at different rates [29,30]. SNPs with a high frequency of the minor allele, such as those used in this study and in the HapMap project, are old and thus often found in all living human populations, although at different frequencies. Microsatellites, though, are less stable and have a mutation rate of about 1 every 1,000 generations [29]. The variation observed usually consists of an increment or decrease of one repeat unit (i.e., two base pairs). Comparison of these two types of polymorphisms thus allows inferences to be made about the age of the mutations. Based on the amount of variation in our microsatellite and SNP information, we can establish three different broad age groups for the mutations: old mutations (c.1200-1209dup and G61E), intermediate (c.1064-1076del, c.1377-1403dup, R368H, and R469W), and young (E387K and W57X). In all age groups mutations in different individuals share a haplotype in LD block A. In old mutations, we observe mutations in the microsatellites in LD blocks B and C and recombination between block A and the other blocks. In intermediate mutations, we find some variation due to mutation in the microsatellites in LD blocks B and C, but no recombination is observed. In young mutations, the haplotype is basically conserved across the whole analyzed region.

Eight mutations with multiple occurrences share extended haplotypes up to 160 kb in size. Six of these mutations are in patients in our study who are from different ethnic origins, meaning that the ancestral mutation has spread to several countries. The Saudi Arabian patients in our study were all homozygous for the CYP1B1 mutations due to the high level of consanguinity customary in this country. The other homozygous CYP1B1 mutation carriers originate from Costa Rica (one from a consanguineous family), from Turkey (two), and from the United States (two). Eleven patients with German, American, Swiss, or Russian origin are compound heterozygotes. In countries where consanguinity is uncommon, the majority of patients are expected to be compound heterozygotes. Most mutations found in the compound heterozygous state are ancestral, widespread mutations, which explains their frequent observation worldwide in different ethnic groups.

Acknowledgements

We thank the patients, and their families for their cooperation in these studies. This work was funded by the DFG, SFB539. GC was a Gottlieb-Daimler-Fellow and is currently a DAAD fellow. RAL is a Senior Scientific Investigator of Research to Prevent Blindness, New York, NY.

References

1. Sarfarazi M, Akarsu AN, Hossain A, Turacli ME, Aktan SG, Barsoum-Homsy M, Chevrette L, Sayli BS. Assignment of a locus (GLC3A) for primary congenital glaucoma (Buphthalmos) to 2p21 and evidence for genetic heterogeneity. Genomics 1995; 30:171-7.

2. Stoilov I, Akarsu AN, Sarfarazi M. Identification of three different truncating mutations in cytochrome P4501B1 (CYP1B1) as the principal cause of primary congenital glaucoma (Buphthalmos) in families linked to the GLC3A locus on chromosome 2p21. Hum Mol Genet 1997; 6:641-7.

3. Stoilov I, Jansson I, Sarfarazi M, Schenkman JB. Roles of cytochrome p450 in development. Drug Metabol Drug Interact 2001; 18:33-55.

4. Bejjani BA, Lewis RA, Tomey KF, Anderson KL, Dueker DK, Jabak M, Astle WF, Otterud B, Leppert M, Lupski JR. Mutations in CYP1B1, the gene for cytochrome P4501B1, are the predominant cause of primary congenital glaucoma in Saudi Arabia. Am J Hum Genet 1998; 62:325-33.

5. Bejjani BA, Stockton DW, Lewis RA, Tomey KF, Dueker DK, Jabak M, Astle WF, Lupski JR. Multiple CYP1B1 mutations and incomplete penetrance in an inbred population segregating primary congenital glaucoma suggest frequent de novo events and a dominant modifier locus. Hum Mol Genet 2000; 9:367-74. Erratum in: Hum Mol Genet 2000; 9:1141.

6. Belmouden A, Melki R, Hamdani M, Zaghloul K, Amraoui A, Nadifi S, Akhayat O, Garchon HJ. A novel frameshift founder mutation in the cytochrome P450 1B1 (CYP1B1) gene is associated with primary congenital glaucoma in Morocco. Clin Genet 2002; 62:334-9.

7. Plasilova M, Stoilov I, Sarfarazi M, Kadasi L, Ferakova E, Ferak V. Identification of a single ancestral CYP1B1 mutation in Slovak Gypsies (Roms) affected with primary congenital glaucoma. J Med Genet 1999; 36:290-4.

8. Sitorus R, Ardjo SM, Lorenz B, Preising M. CYP1B1 gene analysis in primary congenital glaucoma in Indonesian and European patients. J Med Genet 2003; 40:e9.

9. Chakrabarti S, Kaur K, Kaur I, Mandal AK, Parikh RS, Thomas R, Majumder PP. Globally, CYP1B1 mutations in primary congenital glaucoma are strongly structured by geographic and haplotype backgrounds. Invest Ophthalmol Vis Sci 2006; 47:43-7.

10. Reddy AB, Kaur K, Mandal AK, Panicker SG, Thomas R, Hasnain SE, Balasubramanian D, Chakrabarti S. Mutation spectrum of the CYP1B1 gene in Indian primary congenital glaucoma patients. Mol Vis 2004; 10:696-702 <http://www.molvis.org/molvis/v10/a84/>.

11. Reddy AB, Panicker SG, Mandal AK, Hasnain SE, Balasubramanian D. Identification of R368H as a predominant CYP1B1 allele causing primary congenital glaucoma in Indian patients. Invest Ophthalmol Vis Sci 2003; 44:4200-3.

12. Kakiuchi-Matsumoto T, Isashiki Y, Ohba N, Kimura K, Sonoda S, Unoki K. Cytochrome P450 1B1 gene mutations in Japanese patients with primary congenital glaucoma(1). Am J Ophthalmol 2001; 131:345-50.

13. Mashima Y, Suzuki Y, Sergeev Y, Ohtake Y, Tanino T, Kimura I, Miyata H, Aihara M, Tanihara H, Inatani M, Azuma N, Iwata T, Araie M. Novel cytochrome P4501B1 (CYP1B1) gene mutations in Japanese patients with primary congenital glaucoma. Invest Ophthalmol Vis Sci 2001; 42:2211-6. Erratum in: Invest Ophthalmol Vis Sci 2001; 42:2775.

14. Colomb E, Kaplan J, Garchon HJ. Novel cytochrome P450 1B1 (CYP1B1) mutations in patients with primary congenital glaucoma in France. Hum Mutat 2003; 22:496.

15. Michels-Rautenstrauss KG, Mardin CY, Zenker M, Jordan N, Gusek-Schneider GC, Rautenstrauss BW. Primary congenital glaucoma: three case reports on novel mutations and combinations of mutations in the GLC3A (CYP1B1) gene. J Glaucoma 2001; 10:354-7.

16. Curry SM, Daou AG, Hermanns P, Molinari A, Lewis RA, Bejjani BA. Cytochrome P4501B1 mutations cause only part of primary congenital glaucoma in Ecuador. Ophthalmic Genet 2004; 25:3-9.

17. Sena DF, Finzi S, Rodgers K, Del Bono E, Haines JL, Wiggs JL. Founder mutations of CYP1B1 gene in patients with congenital glaucoma from the United States and Brazil. J Med Genet 2004; 41:e6.

18. Soley GC, Bosse KA, Flikier D, Flikier P, Azofeifa J, Mardin CY, Reis A, Michels-Rautenstrauss KG, Rautenstrauss BW. Primary congenital glaucoma: a novel single-nucleotide deletion and varying phenotypic expression for the 1,546-1,555dup mutation in the GLC3A (CYP1B1) gene in 2 families of different ethnic origin. J Glaucoma 2003; 12:27-30.

19. Stoilov IR, Costa VP, Vasconcellos JP, Melo MB, Betinjane AJ, Carani JC, Oltrogge EV, Sarfarazi M. Molecular genetics of primary congenital glaucoma in Brazil. Invest Ophthalmol Vis Sci 2002; 43:1820-7.

20. Churchill AJ, Yeung A. A compound heterozygous change found in Peters' anomaly. Mol Vis 2005; 11:66-70 <http://www.molvis.org/molvis/v11/a7/>.

21. Edward D, Al Rajhi A, Lewis RA, Curry S, Wang Z, Bejjani B. Molecular basis of Peters anomaly in Saudi Arabia. Ophthalmic Genet 2004; 25:257-70.

22. Vincent A, Billingsley G, Priston M, Williams-Lyn D, Sutherland J, Glaser T, Oliver E, Walter MA, Heathcote G, Levin A, Heon E. Phenotypic heterogeneity of CYP1B1: mutations in a patient with Peters' anomaly. J Med Genet 2001; 38:324-6.

23. Gould DB, John SW. Anterior segment dysgenesis and the developmental glaucomas are complex traits. Hum Mol Genet 2002; 11:1185-93.

24. Melki R, Colomb E, Lefort N, Brezin AP, Garchon HJ. CYP1B1 mutations in French patients with early-onset primary open-angle glaucoma. J Med Genet 2004; 41:647-51.

25. Stenson PD, Ball EV, Mort M, Phillips AD, Shiel JA, Thomas NS, Abeysinghe S, Krawczak M, Cooper DN. Human Gene Mutation Database (HGMD): 2003 update. Hum Mutat 2003; 21:577-81.

26. Barrett JC, Fry B, Maller J, Daly MJ. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 2005; 21:263-5.

27. The International HapMap Consortium. The International HapMap Project. Nature 2003; 426:789-96.

28. den Dunnen JT, Antonarakis SE. Nomenclature for the description of human sequence variations. Hum Genet 2001; 109:121-4.

29. Ellegren H. Heterogeneous mutation processes in human microsatellite DNA sequences. Nat Genet 2000; 24:400-2.

30. Lercher MJ, Hurst LD. Human SNP variability and mutation rate are higher in regions of high recombination. Trends Genet 2002; 18:337-40.