Human Molecular Genetics, 2000, Vol. 9, No. 9 1329-1349
© 2000 Oxford University Press
A full genome scan for age-related maculopathy
1Department of Human Genetics, University of Pittsburgh Graduate School of Public Health, Pittsburgh, PA 15261, USA, 2Department of Ophthalmology, University of Pittsburgh School of Medicine, Pittsburgh, PA 15213, USA, 3Smith-Kettlewell Eye Research Institute, San Francisco, CA 94115, USA, 4Department of Ophthalmology, University of California at Davis, Davis, CA 95816, USA, 5409 Heron Place, Davis, CA 95616, USA and 6Erie Retinal Surgery Inc., Erie, PA 16507, USA
Received 6 January 2000; Revised and Accepted 15 March 2000.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Age-related macular degeneration or age-related maculopathy (ARM) is a major public health issue, as it is the leading cause of irreversible vision loss in the elderly in the Western world. Using three diagnostic models, we have genotyped markers in 16 plausible candidate regions and have carried out a genome-wide screen for ARM susceptibility loci. A panel of 225 ARM families comprising up to 212 affected sib pairs was genotyped for 386 markers. Under our most stringent diagnostic model, the regions with the strongest evidence of linkage were on chromosome 9 near D9S301 and on 10 near D10S1230, with peak multipoint heterogeneity LOD scores (HLOD) of 1.87 and 1.42 and peak GeneHunter-Plus non-parametric LOD scores (GHP LOD) of 1.69 and 1.83. After expanding our initial set of families to 364 ARM families with up to 329 affected sib pairs, the linkage signal on chromosome 9 vanished, while the chromosome 10 signal decreased to a GHP LOD of about 1.0, with a SimIBD P-value of 0.008 under the broadest diagnostic model with marker D10S1236. After error filtration, the GHP LOD increased to 1.27 under our most stringent model and 1.42 under our broadest model, peaking near D10S1236. This peak was seen consistently across all three diagnostic models. Our analyses also excluded up to nine different candidate regions and identified a few other regions of potential linkage, suitable for further studies. Of particular interest was the region on chromosome 5 near D5S1480, where a reasonable candidate gene, glutathione peroxidase 3, resides.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Age-related maculopathy (ARM), also known as age-related macular degeneration (AMD), is the leading cause of irreversible vision loss in the elderly population in the USA and the Western world and a major public health issue. Affecting nearly 9% of the population over the age of 65, ARM becomes increasingly prevalent with age such that by age 75 and older nearly 28% of individuals are affected (1–6). As the proportion of the elderly in our population increases, the public health impact of ARM will become even more severe. Currently there is little that can be done to prevent or slow the progression of ARM (7).
ARM is a degenerative disorder involving the retinal pigment epithelium, choriocapillaris (8) and retina which primarily, but not exclusively, affects the macular region. Symptoms of ARM include metamorphopsia, impaired light adaptation and decreased central vision. ARM causes changes in the macula of the retina, which include the presence of drusen (9–12) and/or changes in the retinal pigment epithelium (RPE), geographic atrophy and choroidal neovascularization (13–15). Advanced ARM has been categorized in two forms, an exudative form associated with the development of choroidal neovascular membranes and a ‘dry’ form which includes non-exudative disease and/or geographic atrophy. Either form may be present with hard drusen and pigment migration. Despite morphological differences, there is no direct evidence to suggest that these different clinical forms have distinct etiologies.
Numerous studies have attempted to establish dietary and environmental risk factors that contribute to ARM; however, heredity has emerged as the primary determinant of ARM susceptibility (for a review see ref. 16). First-degree relatives of ARM patients are at between two and four times greater risk of developing ARM than the first-degree relatives of controls (17,18). Twin studies have consistently shown high levels of concordance of the disease among monozygous sibs and strongly support a genetic etiology (19–22). Careful segregation analyses on a large study of 564 families suggest that a single major gene accounts for 89–97% of the genetic variability or 55–57% of the total variability (23). Even so, relatively little is known about identified genetic risk factors for ARM, due to difficulties in diagnosis, late onset and complexity of expression.
Over 20 years ago, Gass (24) proposed that ARM represents a continuum of the earlier onset hereditary macular and retinal degenerations. The clinical similarities among ARM and many of these conditions, such as autosomal dominant drusen, Sorsby fundus dystrophy, peculiar foveomacular dystrophy and adult vitelliform dystrophy, have given rise to the hypothesis that the genes responsible for these monogenic macular dystrophies and other retinal degenerations may play a role in the etiology of ARM. To test this hypothesis, one can attempt to isolate the genes responsible for these conditions and test for mutations in the ARM population. This approach has been attempted for mutations in the peripherin gene, which is known to cause retinal degeneration, including macular degeneration (25–28), and the TIMP3 gene, which has been implicated in Sorsby fundus dystrophy (29,30). However, TIMP3 has been excluded as a major cause of ARM (31,32). Several alleles in the photoreceptor gene ABCR, which is mutated in Stargardt disease, have been shown to be more prevalent in a large ARM population (33) as compared with a control population, suggesting that up to 5% of ARM may be attributable to these two alleles. However, these findings do not clearly establish the extent to which ABCR variants may account for typical ARM. Additional studies have both supported or discounted a role of the ABCR gene in AMD (34–36). Two recent studies indicate that the 
4 allele of the apolipoprotein gene (Apo E) is associated with decreased risk of ARM (37,38). Finally, linkage studies of large families with late onset macular degeneration, clinically indistinguishable from ARM, have identified loci on chromosomes 1 and 6 (39,40).
In addition to testing markers near 16 candidate genes, we conducted a small 20 cM, in-house, genome-wide scan with 120 families followed by a full genome-wide screen for ARM susceptibility loci with the NHLBI Mammalian Genotyping Service (http://www.marshmed.org/genetics/ ). This expanded effort included a panel of 225 ARM families who were genotyped for 386 markers. Four regions of interest were followed up in an expanded set of 364 ARM families (which included the initial 225 families).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
We collected a set of 364 families (2129 individuals) containing relative pairs affected with ARM (Table 1). We obtained a minimum of two affected individuals per family and, whenever possible, additional family members were collected either to be used as affected individuals or for haplotype reconstruction (using unknown status). We specifically avoided specifying individuals as unaffected because of the potential for inadequate documentation and the possibility that seemingly unaffected individuals might still develop ARM in the future. Attempts to create an age-dependent penetrance function for unaffected status of ARM were unsuccessful, largely because the documentation for potentially unaffected individuals was based on records which were not standardized and usually less complete than the documentation for clearly affected individuals. We established three definitions of affected status for genetic analyses. A subset of our population was clearly affected with ARM based upon drusen, pigmentary changes and the presence of endstage disease (Model A). A larger, second group includes these individuals plus those who were considered to be probably affected with ARM (Model B) based on extensive drusen, serous or drusenoid pigment epithelial detachment and/or disruption of the macular pigment epithelium. A third, even more inclusive group (Model C) added individuals for whom there was either insufficient evidence to rule out another type of macular disease due to inadequate documentation or because the presence of endstage degeneration precluded a definitive assessment of the etiology.
|
We typed a total of 422 markers, which included 386 evenly spaced markers designed to span the genome, 18 markers selected to be in the region of reasonable candidate genes (Table 2) and 18 markers to follow up regions on chromosomes 5, 9, 10 and 12. We chose to follow up the region on chromosome 5 based on the prior in-house, genome-wide scan of a subset of our families; however, this scan was done on Pharmacia Alf-Express machines, which employ single dye detection and require widely spaced size standards which reduce the accuracy of the genotyping. Genotyping of the 386 markers in the genome-wide screen was conducted by the NHLBI Mammalian Genotyping Service.
|
The results of our multipoint analyses of the 225 families in the initial genome-wide scan are displayed in Figure 1. Under Model A, the regions with the strongest evidence of linkage were found on chromosome 9 near D9S301 and on 10 near D10S1230, with peak multipoint heterogeneity LOD scores (HLOD) of 1.87 and 1.42 and peak GeneHunter-Plus non-parametric LOD scores (GHP LOD) of 1.69 and 1.83. Under Model B, similar peaks were seen for the same locations with an additional potential locus on chromosome 7 near D7S1808 (HLOD of 1.89, GHP LOD of 0.31). Under Model C, the significance of the loci on chromosome 9 near D9S301 and on 7 were decreased, but there was also some signal at chromosome 9 near D9S1838 and on 10. Under Model C, there was also some evidence of linkage on chromosome 12 near D12S2070, with a peak HLOD of 0.74 and a peak GHP LOD of 1.43. Single marker analyses (Table 3) were fairly consistent with the multipoint results, with the smallest SimIBD P-value of 0.027 obtained under Model A for D10S1237.
|
|
Follow-up genotyping
We genotyped follow-up markers in regions of interest on chromosomes 5, 9, 10 and 12. After adding the additional markers, the linkage signal on chromosome 9 (near D9S301) vanished and the signal on 12 (near D12S395) sank below a GHP LOD of 0.66 or less in all three models (data not shown). In contrast, on chromosome 5 the GHP LOD increased to 1.32 near D5S1462 in Model B (Fig. 2, left column); however, this peak was not seen consistently across all three models. After error filtration, this peak stayed almost exactly the same under Model B, but now a peak occurred under Model C with a GHP LOD of 1.24 near D5S1480 (Fig. 2, right column). On chromosome 10, follow-up genotyping decreased the GHP LOD to about 1 under Models A and C (Fig. 2, left column), but after error filtration this increased to 1.27 under Model A and 1.42 under Model C, peaking near D10S1236. This peak was seen consistently across all three diagnostic models. Single marker results (Table 4) were strongest for loci on chromosome 10, with a SimIBD P-value of 0.008 under Model C for D10S1236.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
The selection and classification of ARM families
In the absence of any clear cut evidence that the phenotypic variations of ARM reflect genetic heterogeneity, we made a conscious decision to collectively assess individuals with either atrophic and/or exudative forms of ARM. It was more critical to clearly establish which individuals were truly affected and the degree of confidence in the diagnosis of ARM. Due to extensive recruitment from vitreo-retinal practices, a majority (>60%) of the affected individuals in our study had historical or photographic evidence of exudative ARM, consisting of a classical and/or occult choroidal neovascular membrane and/or a disciform scar. This is clearly a much higher proportion of patients with exudative disease than would be predicted by population studies of ARM (5–7,41–44). We have previously shown (45) that patients with atrophic ARM are just as likely to have positive family members as are individuals with exudative lesions. Currently, there is no evidence as to whether or not the atrophic and exudative distinctions of ARM may reflect different genetic etiologies.
The distinction of dry and non-exudative ARM is poorly adhered to by many clinicians, including retinal specialists. Many eye care specialists make no distinction between atrophic or dry macular degeneration and patients who have not yet developed a choroidal neovascular membrane or who have been successfully treated for a choroidal neovascular membrane in the past. We have attempted to minimize this confusion by careful review of all available eye care records. Our patient classification distinguishes between those who have had some evidence of a choroidal neovascular membrane at any time during the course of their condition, those who have no evidence of having such an event and those who are reported (or confirmed by fundus photography) to have areas of RPE atrophy unrelated to previous laser treatment.
Candidate gene linkage analyses
There was no evidence of linkage for any of the candidate genes, except for the marker ACCP2 in the vicinity of the rhodopsin gene (HLOD of 0.39, GHP LOD of 0.95 under Model A). In fact, under Model C, nine of the 16 candidate regions were excluded from containing a susceptibility locus with a 
s of 1.5 (Table 2). In particular, we excluded linkage to the ABCR locus under all three models. Based upon our screening of two major ABCR mutations implicated in ARM, we observed that two unrelated individuals carried the G1961E mutation (out of 381 individuals screened). In addition to the proband, the family of the first individual also contained an affected sibling who did not carry the mutation. The family of the second affected individual with the G1961E mutation also contained an affected nephew who did not carry the mutation. We did not detect the D2177N mutation in any of our 447 samples. However, our relative lack of patients with atrophic disease could partly account for this finding. Our exclusion results should be viewed with an appropriate level of skepticism, as exclusion is computed under an assumed genetic model, which may or may not be true. For example, Sheffield and Stone found that their affected pedigree member (APM) linkage analyses of 180 glaucoma sib pairs failed to detect linkage with the GLCA1 locus, despite the fact that direct candidate gene studies indicate that this locus may contribute to as much as 4% of chronic open angle glaucoma (COAG) cases (46; E. Stone, personal communication).
Genome-wide scan and follow-up linkage analyses
Our genome-wide scan provided four regions for follow-up, on chromosome 5, 9, 10 and 12. After follow-up, our only consistent evidence of linkage across all three diagnostic models occured on chromosome 10 near D10S1230. The single marker analyses gave a LOD score of ~1.5 under Models A and C (Table 4) at D10S1230 and a SimIBD P-value of 0.008 at D10S1236 under Model C. The multipoint curves (Fig. 2) all show a narrow well-defined peak, with a GHP LOD of ~1 under Models A and C; these peaks increase a bit after error filtration. While this evidence for linkage appears to be consistent, the level of significance is low and such peaks could have easily been seen by chance in the absence of linkage.
Glutathione peroxidase (GPX) 3
Oxidative damage has been implicated in the pathogenesis of ARM (47). The gene for GPX3, localized to chromosome 5q32 near D51480, was therefore an excellent candidate gene, based on the results of our first, in-house, genome-wide scan. GPX protects cells from oxidative damage by catalyzing the reduction of lipid hydroperoxide and hydrogen peroxide (48). The retina is composed of high concentrations of long chain polyunsaturated fatty acids, making it susceptible to lipid peroxidation through the interaction of light with the retina. An abundance of GPX has been localized to the retina through immunohistochemistry (49). Conjugated diene levels, an indicator of lipid peroxidation, has been shown to increase in the retina of the rat with age, accompanied by a significant decrease in GPX activity (50). Several studies in the human have indicated that plasma GPX levels decrease significantly as we age (51). Given that GPX is present in the retina, that plasma levels of GPX decline as we age and that the retina is an excellent substrate for lipid peroxidation, alterations within the GPX gene could be responsible for ARM. Our examination of the GPX3 gene has resulted in the characterization of two new polymorphisms, GPXPR1 and GPXPR2, in the promoter region. While these polymorphisms have limited informativeness due to high homozygosity, we incorporated the genotypings into the microsatellite marker data from the genome-wide scan done by the NHLBI Mammalian Genotyping Service. After error filtration, the GHP LOD score had a sharp, but modest, peak near the GPX polymorphisms (Fig. 2). Preliminary association testing utilizing one proband per family and age-matched controls indicated evidence of association which warrants further investigation (52). We are currently evaluating our ARM families for family-specific mutations within this gene.
Genotype errors
While theory clearly establishes that genotyping errors can wipe out evidence for linkage, such errors are very difficult or impossible to detect in small nuclear families in a late-onset disease such as ARM where parents are not available for genotyping. Our disagreement rates between our local genotyping and genotypes done at the NHLBI Mammalian Genotyping Service (Table 5) indicate that undetected genotype errors are much more common than we would hope. While we have attempted to address this frustrating problem through a variety of error detection methodologies, such as comparing allele frequencies, looking for multiple recombinants and employing new statistical methodologies (53), these methods only become effective as the marker density becomes high.
|
Conclusion
Our evidence for linkage of ARM to known loci attributed to macular degenerations (such as TIMP-3, the mallatia leventinese locus and ABCR) fail to demonstrate that these loci contribute to a significant percentage of ARM cases across our diagnostic models and analytical approaches. Our genome-wide scan and follow-up data implicate two potential regions for further follow-up on chromosomes 5 and 10, with the region on chromosome 10 having higher priority. Even so, with a large number of carefully diagnosed individuals and efforts to control for genotyping error, the evidence for linkage remains tentative at best, underscoring the difficulties of using genetic linkage approaches to study common late-onset complex diseases.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Families
ARM families were ascertained by both prospective enrollment of patients who were diagnosed with ARM and who self-reported positive family histories of the condition as well as by a screening program of ARM patients identified within private ophthalmology practices using billing records and a simple postcard survey. This practice-based survey approach for ARM families has been described elsewhere (45). The ophthalmologists who contributed the majority of participants to this study are listed in the Acknowledgements.
Evaluations
Whenever possible, individuals were examined by one of the investigators of the study (M.B.G. or T.O.P.); however, photographic records and ophthalmic examination reports were used for the majority of affected individuals. Photographs were obtained from eye care providers’ offices without standardization. No photographs were rejected; however, photographs that did not include the macula or lacked adequate detail were graded as uncertain for specific pathology. While it would have been best to obtain standardized photographs on all patients, it would have been financially prohibitive and physically impossible. In some instances, fluorescein angiograms were provided and these were incorporated into the gradings. There was no specified protocol given to the eye care providers for photographs and not all of the photographs were as current as the medical records. Eye care records were requested from optometrists, ophthalmologists and retina specialists. The probands were all recruited from general ophthalmology or retinal practices. However, family members received their eye care from a much broader distribution of eye care providers, including optometrists. Photographs were read in a masked fashion by a single reviewer (M.B.G.) and coded for disease severity status, diagnostic certainty of ARM and a general score of ARM severity (by person) using the more severely affected eye. Eye care records were graded in a non-masked fashion using the same final classification categories. If there were discrepancies between the photographic records and clinical reports, then the materials were reviewed together and additional documentation was sought if necessary.
Inclusion/exclusion criteria: grading of ARM pathology
Photographs and records were evaluated for evidence of pathology that could be attributed to ARM as well as abnormalities that might indicate an alternative diagnosis. Age of onset of vision loss was also ascertained from the records, as well as the bilaterality of the clinical findings. The basic features of ARM (54), including evidence of prior laser treatment, were abstracted from the medical records or graded from available photographs.
Individuals were excluded from having a diagnosis of ARM if there was documented evidence of the following conditions: angioid streaks, myopic degeneration (including a history of high axial myopia prior to cataract surgery), central serous chorioretinopathy, ocular histoplasmosis, vitelliform lesions in the absence of drusen (adult foveomacular dystrophy), toxoplasmosis, central areolar choroidal sclerosis (with no history of drusen), Stargardt disease, Best disease, juxtafoveolar telangiectasia, pattern dystrophies, vision loss before the age of 40 or a history of macular disease within the first four decades.
Classification of subjects for linkage analyses
The geneticist’s primary concerns are to establish a level of certainty that the subject is affected and the certainty of the diagnosis (specifically the exclusion of other disorders). There are two fundamental questions regarding each participant in a genetic study of ARM: (i) how confident can one be that an individual is affected with macular degeneration? (Disease severity); (ii) Given an affected individual, how confident can one be of the diagnosis? (Diagnosis). The specific issues regarding the classifications are described in Table 6.
|
We established three clinical grades of ARM for genetic analyses. Under the most stringent model (Model A), individuals were classified as ‘affected’ only if they were clearly affected with ARM based upon extensive/coalescent drusen, pigmentary changes (including pigment epithelial detachment) and/or the presence of endstage disease (geographic atrophy and/or choroidal neovascular membranes). Model B analyses classified as ‘affected’ all affecteds under Model A and those who were considered to be probably affected with ARM based upon moderate to extensive soft drusen, extensive hard drusen with pigmentary changes and suspected pigment epithelial detachments. Model C classified as ‘affected’ all of the affecteds under Models A and B as well as those whose diagnosis of ARM was less clear cut. Under Model C, individuals were considered affected if they were definitely and probably affected with ARM or a related maculopathy (having insufficient evidence to rule out another type of macular disease). Model C also included individuals with endstage disease (choroidal neovascular membranes, disciform scarring and/or geographic atrophy) in the absence of any other documentation of macular pathology. These individuals were considered to be definitely affected, but the determination of whether the etiology was ARM or another maculopathy was considered ambiguous.
Genotyping methods
Genomic DNA was extracted from leukocytes using a simple salting out procedure as described previously (55). PCR was performed to amplify each sample for each marker. The primers were contingent upon which method was utilized to run the gels. Candidate gene markers were genotyped using autoradiography and 32P-labeled primers. The initial genome-wide scan was performed with the 20 cM Research Genetics marker set and ALF-Express automated sequencers. The subsequent and more extensive genome-wide scan was performed by the NHLBI Mammalian Genotyping Service (http://www.marshmed.org/genetics/ ). Markers used for the subsequent focused genotyping were genotyped using an ABI377 automated sequencer (Applied Biosystems Inc., Foster City, CA) and fluorescent-labeled primers, which were purchased from Research Genetics. PCR conditions consisted of 30 cycles of denaturation at 96°C, annealing at 50–55°C and extension at 72°C, each for 1 min in a 96-well Techne thermal cycler (Techne Inc., Princeton, NJ). All of the ABI377 generated genotypes were analyzed using Genotyper 2.0 software (Applied Biosystems Inc.) and were called independently by two experienced individuals. Discrepancies between the two genotype calls were then carefully investigated and resolved.
Candidate loci
Like many other investigators, we have been interested in the relationship of ARM to specific monogenic macular and retinal dystrophies such as Stargardt disease, Sorsby fundus dystrophy, Best disease, cone dystrophies, mallatia leventinese and retinitis pigmentosa. Since the initiation of this project, the genes for the autosomal recessive Stargardt disease and Sorsby fundus dystrophy have been isolated. To explore the potential contribution of these disorders to ARM, we have typed one or more highly polymorphic microsatellite markers that have been shown to closely map to these disease loci. The list of candidate disorders and genetic markers are shown in Table 2. We then merged these markers with all the other markers that had been typed in order to compute a multipoint exclusion curve.
Follow-up markers
For the follow-up markers, we typed more families in addition to those genotyped by the NHLBI Mammalian Genotyping Service. On chromosome 9 we typed eight follow-up markers in the region of D9S301. Four of these, D9S2025, D9S1874, D9S1876 and D9S1119, were new markers; the other four markers, D9S1118, D9S301, D9S1122 and D9S922, were also typed by the Genotyping Service. For the markers externally typed in the genome-wide scan, we typed mainly families that had not been sent to the Genotyping Service, but we also typed some genotypes in common for quality control. At D9S922 (which we used to help us refine our genotyping protocols before typing the other follow-up markers) we typed 349 genotypes in common and found a disagreement rate of 3.7% (Table 5). Of the 13 genotypes that disagreed, we had one homozygous genotype (8%) while the Genotyping Service had six homozygous genotypes (46%). For two of the other markers, the Genotyping Service had more homozygous calls than we did at genotypes that disagreed.
On chromosome 10 we typed eight follow-up markers in the region of D10S1230. Five of these, D10S198, D10S1236, D10S1246, D10S221 and D10S214, were new markers; the other three markers, D10S677, D10S1237 and D10S1230, were also typed by the Genotyping Service.
On chromosome 12 we typed three follow-up markers in the region of D12S395. Two of these, D12S386 and D12S2082, were new markers not typed by the Genotyping Service; the remaining marker, D12S2070, was also typed by the Genotyping Service.
On chromosome 5 we typed nine follow-up markers in the region of D5S820. Seven of these, D5S1352, D5S1475, D5S2013, D5S2077, D5S2090, D5S412 and D5S812d, were new markers; the remaining two, D5S1505 and D5S816, were also typed by the Genotyping Service. At these two markers, there were substantially more homozygotes called by the Genotyping Service than by us among the genotypes that disagreed (Table 5).
Sample sizes for follow-up markers that were not genotyped by the Genotyping Service unfortunately tend to be smaller, because we had run out of DNA on a fair number of the earlier families. We are currently in the process of re-sampling these families in order to replenish our DNA samples.
GPX3 polymorphisms
Direct sequencing of the promotor region of the GPX3 gene revealed an A/G polymorphism (GPXPR1) 382 nt and an A/C polymorphism (GPXPR2) 703 nt upstream of the transcriptional start site. The base change in GPXPR1 introduces an MboI site and that in GPXPR2 a FokI site. The primers to PCR amplify GPXPR1 were 5'-GCTTCTAAGCCAGGTGGTT-3' (forward) and 5'-CACCTGGCGATGGTCCTCTG-3' (reverse), giving a PCR product of 623 bp using an annealing temperature of 55°C. The primers to PCR amplify GPXPR2 were 5'-GACTCTGGTGGTCAGAGAGAT-3' (forward) and 5'-AGCAAGGCTTGTCTGTTCGT-3' (reverse), giving a PCR product of 302 bp using an annealing temperature of 57°C.
ABCR screening
All of our available ARM samples were screened for the G1961E and D2177N mutations in the ABCR gene that have been implicated in ARM pathogenesis (33); we also genotyped these samples for the non-pathogenic R943Q polymorphism. The G1961E mutation creates a TaqI restriction site and the R943Q polymorphism destroys an MspI restriction site. They were genotyped by amplifying exons 42 and 19, respectively, and performing RFLP analysis. The D2177N mutation does not alter a restriction site, therefore a primer-directed RFLP analysis was performed by utilizing a primer sequence which introduces a G
A substitution destroying an AvaII site in the mutant allele. RFLP analysis was then performed for exon 48. Our allele frequency, based on one affected per family, for the R943Q polymorphism is 36% which is much higher than previously reported (33,56).
Statistical analyses
Error checking.
Our genotyping data were subjected to a variety of quality checks in order to help ensure accurate genotyping. Genotypes at each locus were checked for Mendelian inconsistencies using our program PedCheck (57), which uses genotype elimination to identify the more subtle inconsistencies. We estimated allele frequencies at each locus by simple allele counting in all individuals (disregarding relationships). We compared these estimates via automated Perl scripts to estimates available on the Internet from Genethon (ftp://ftp.genethon.fr/pub/Gmap/Nature-1995/ ), Pericak-Vance’s group at Duke University (http://www2.mc.duke.edu/depts/medicine/medgen/allele_freqs ), the Cooperative Human Linkage Center (http://www.chlc.org ) and the Marshfield Medical Research Center (http://www.marshmed.org/markers/ ). A disparity in allele frequency estimates sometimes indicated miscalling of alleles due to a misalignment of gels or bleed-through from an adjacent lane. We compared the observed number of homozygotes to the expected number of homozygotes based on the estimated allele frequencies; a disparity in these numbers may indicate miscalling of heterozygous genotypes as homozygous. We then used the APM method (58) to assay identity-by-state sharing between all pairs of affected relatives; those markers that displayed much less sharing than expected were double checked for data errors. We also examined the genotypes of several putative monozygotic twin pairs in order to assay the quality of our genotyping data.
For the more densely typed regions of interest, we checked for excessive multiple recombinants by computerized haplotyping using SIMWALK2 (59); however, for most regions of the genome our markers are too sparsely spaced for error detection via haplotyping to be very efficient. We used the relationship estimation program RELPAIR (60) to check the accuracy of our assumed relationships between individuals against the realized level of allele sharing across the whole genome; this led to the detection of two relationship errors where true half-siblings had been entered as full siblings.
Allele frequency estimation.
We estimated our allele frequencies by counting all alleles in all genotypes, ignoring familial relationships. Ignoring familial relationships may lead to slightly biased estimates, but these biases are likely to be quite small in our data as our data contains so many small sparsely typed pedigrees [we verified this for many markers by using the likelihood-based approach implemented in the USERM13 module of MENDEL (61) to estimate the allele frequencies]. Our allele frequency estimates can also be biased by the fact that the vast majority of our genotypes occur in affected individuals. This means that if there were a strong association so a particular allele occurred in the affecteds more often than in the general population, the frequency of that allele would be overestimated. This potential effect is conservative in that it would bias us away from linkage (62,63). As mentioned above, the vast majority of our allele frequency estimates agreed quite well with published frequencies.
Linkage analysis.
We carried out the following analyses.
(i) Single point and multipoint LOD scores under heterogeneity
. Since many of our pedigrees show inheritance patterns that seem to be consistent with dominant inheritance, we chose a priori to compute LOD scores under a single simple dominant model [disease allele frequency 0.0001; penetrance vector (0.01 0.90 0.90)] while allowing for heterogeneity under an admixture model (64). Only two disease phenotypes were used: ‘affected’ and ‘unknown’; no-one was given a ‘normal’ phenotype due to the complexities of the ARM phenotype. The single point LOD scores were computed using MLINK of FASTLINKAGE (65–69) and HOMOG (70) and the multipoint LOD scores were computed using GeneHunter (71) (after correction of the maximization routine for computing heterogeneity LOD scores). Note that the majority of our families are so small that the family-specific LOD score maximizes at either 
= 0 or = 0.50. In a desire to keep the number of models used in the analyses as low as possible, we elected not to employ age-dependent penetrance models in our study. An age-dependent penetrance curve is likely to be of unclear benefit because: (a) such penetrance curves help the most with ‘unaffected’ individuals, but the vast majority of people we genotyped were affected; (b) it is quite difficult to construct an age-dependent penetrance function for ARM, because the ocular changes reflect a continuum with no clear age of onset for most individuals.
(ii) Single marker ‘model-free’ methods.
For initial fast analyses, we applied the APM method of linkage analysis (58), which is particularly appropriate for our data due to the late onset of the disease. However, we present here only simulation-based SimIBD results (72,73), as we have shown that our SimIBD approach is more powerful and less sensitive to mis-specification of allele frequencies because this statistic generates an empirical P-value conditional on any genotyped non-affecteds (74).
(iii) Multiple marker ‘model-free’ analyses.
We also computed multipoint ‘model-free’ LOD scores using the All statistic with GeneHunter-Plus (71,75), which will handle small to medium sized pedigrees without altering their structure, but trims our four largest pedigrees due to computational constraints. We used the kac program to compute GHP LOD under the linear model. We also used the sib_phase module from the ASPEX package (76) to compute multipoint exclusion curves at an assumed s value of 1.5, where s is the ratio of the sibling risk to the population prevalence. A s of 1.5 would represent a major gene effect in ARM. Note that ASPEX has a disadvantage in that it analyzes only nuclear families, requiring one to break our larger pedigrees into their component nuclear pedigrees (and thus losing all information from our more distant relative pairs; Table 1).
(iv) Error filtration.
We used Mendel 4 (by Kenneth Lange) to compute the a posteriori probability of a genotype error for each individual and family conditional on the multipoint marker genotype data. We then untyped each family at each locus for which their a posteriori probability of a genotype error was >50%. Such an approach has been shown to ‘recover’ the evidence for linkage that is obscured by genotyping error and is unlikely to produce false evidence for linkage when none is present (53).
(v) Robustness to allele frequency estimation.
As mentioned above, counting all alleles while ignoring familial relationships may lead to slightly biased allele frequency estimates. In order to check the robustness of our results, we derived unbiased estimates by counting alleles in founders when a pedigree had at least one genotyped founder plus a randomly chosen genotyped person from each pedigree without any genotyped founders. This results in a sample size of ~422–430 alleles. We then adjusted those alleles with 0 frequency estimates by increasing them to 0.002367 = 1/422, under the assumption that the missing alleles occur at least once in the data set but were not seen due to the random sampling (this makes these rare alleles a bit more frequent than they probably really are). We then made the allele frequencies sum to 1 by reducing the frequency of the most common allele. When we used these new allele frequency estimates on the genome-wide scan data, the HLOD and GHP LOD were almost identical, with most deviations <0.05. The largest positive change we saw under Model A was 0.14 for HLOD (new, 1.83; old, 1.69) on chromosome 9 (all other positive changes on a chromosome were <0.09 and negative changes were less than –0.04); for GHP LOD, the largest positive change was 0.11 on chromosome 21 (new, 0.54; old, 0.43) (all other positive changes on a chromsome were <0.09 and negative changes were less than –0.06). Maximum changes under the other models were of the same or smaller magnitude. Such minor differences do not alter the conclusions of this study at all and indicate that our analyses are conservative, as anticipated. Also, one should note that the exclusion curves computed with ASPEX are not influenced by this issue, as ASPEX internally estimates marker allele frequencies itself using an unbiased maximum likelihood-based approach.
(vi) Power to detect linkage.
While power estimates for complex diseases may not be that helpful, we briefly present here some simple power estimates which indicate that these data have substantial power to detect a major gene. We took the families used in the follow-up genotyping on chromosome 10 and, holding the Model A disease statuses constant, we used SLINK (77,78) to simulate a four-allele marker with equally frequent alleles linked at = 0 under the assumption that the disease was inherited in an autosomal dominant manner with a (common) disease allele frequency of 0.01, a penetrance of 70% and a high phenocopy rate of 0.01. One hundred simulated data sets were then analyzed exactly as we did our analyses of the real data and we found that, with 100% of the families linked, the mean LOD was 7.6, the mean HLOD was 7.7 and the mean GHP LOD was 9.4. Under heterogeneity, with only 50% of the families linked, the mean LOD was –0.7, the mean HLOD was 1.9 and the mean GHP LOD was 2.1; with 70% of the families linked, these values were 2.4, 3.7 and 4.1, respectively.
(vii) False positive rates.
We simulated six linked markers with realistic allele frequencies in a region spanning 78 cM, using 282 pedigrees (the number of pedigrees is slightly less than the 292 pedigrees used on chromosome 10 since this simulation was done at an earlier stage of this study). To do this, we used SIMULATE (79) to simulate marker genotypes for exactly the same people genotyped and the same Model C disease phenotypes as in the original real data, under the assumption that the set of markers was segregating independently of ARM disease status. We then analyzed each simulated replicate using GeneHunter-Plus in exactly the same manner as we had analyzed the real data and recorded the results. We obtained essentially identical tail distributions for the maximum HLOD and the maximum GHP LOD in the region, based on 9431 replicates: we have a 1% chance of seeing a statistic greater than ~1.6 and a 0.1% chance of a statistic exceeding ~2.6 in a region of 78 cM solely by chance. Note that these empirical distributions are looking at the maximum over the region, which is different than looking at the marginal distribution at a particular map position.
|
ACKNOWLEDGEMENTS |
|---|
This study would not have been possible without the generous support in terms of time and effort of a large number of families and clinicians who assisted in the recruitment and ascertainment of ARM families. It is impossible to cite all of the clinicians who provided records and photographs, but we wish to thank those who allowed us to conduct screening programs through their clinical practices. The physicians’ names are specifically included when only certain members of a practice group participated in the screening program. This includes: Seneca Eye Surgeons (Warren, PA); Erie Retinal Surgery Inc. (Erie, PA); Health Center Ophthalmology, University of Pittsburgh (Eller and Friberg) (Pittsburgh, PA); the private offices of Drs D. Berk, S. Portnoy, R. Barad, G. Gerneth, D. Nadler and T. Gordon (Pittsburgh, PA); Retinal Associates of Cleveland (Cleveland, OH); Division of Vitreoretinal Surgery, Department of Ophthalmology, The Cleveland Clinic Foundation (Cleveland, OH); Associated Retinal Consultants PC (Detroit, MI); Texas Retina Associates (Dallas, TX); the Division of Vitreoretinal Surgery, Department of Ophthalmology, Milton S. Hershey-Geisinger Medical Center, Pennsylvania State University (Hershey, PA); Division of Vitreoretinal Surgery, Department of Ophthalmology, The Emory Clinic Inc. (Atlanta, GA). We also wish to acknowledge the efforts of the laboratory staff and clinical co-ordinators that were devoted to this project: Ms Anne Deka, Ms Ling Mei Yu, Ms Catherine Klump, Ms Maria Shaffer-Gordon, Dr Anna Burchis, Dr Marina Blagodatny, Dr Heena Sheth, Dr Birgul Gur and Ms Laura Barnes. We gratefully acknowledge the contribution of Dr Margaret Pericak-Vance who made available the allele frequencies from the Duke data sets (grant NS26630). This study was initially funded by the Smith-Kettlewell Eye Research Foundation, San Francisco, CA (M.B.G. and T.O.P.) and the Pennsylvanian Lions Sight Conservation and Eye Research Foundation (M.B.G.). Additional support was subsequently provided by: The Eye and Ear Foundation of Pittsburgh (Pittsburgh, PA); NEI R01-EY09859 (M.B.G.); Research to Prevent Blindness (New York, NY); The Steinbach Foundation (New York, NY) (M.B.G.). This study also received substantial support in the form of marker genotypes from Dr James Weber and colleagues at the NHLBI Mammalian Genotyping Service (http://www.marshmed.org/genetics/ ).
|
FOOTNOTES |
|---|
+ To whom correspondence should be addressed. Tel: +1 412 647 2111; Fax: +1 412 647 5880; Email: gorin@vision.eei.upmc.edu
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
1 Hakkinen, L. (1984) Vision in the elderly and its use in the social environment. Scand. J. Soc. Med., 35 (Suppl), 5–60.
2 Kahn, H.A., Leibowitz, H.M., Ganley, J.P., Kini, M.M., Colton, T., Nickerson, R.S. and Dawber, T.R. (1977) The Framingham Eye Study. I. Outline and major prevalence findings. Am. J. Epidemiol., 106, 17–32.
3 Klein, B. and Klein, R. (1982) Cataracts and macular degeneration in older Americans. Arch. Ophthalmol., 100, 571–573.[Abstract]
4 Martinez, G. and Campbell, A. (1982) Prevalence of ocular disease in a population study of subjects 65 years old and older. Am. J. Ophthalmol., 94, 181–189.[ISI][Medline]
5 Sperduto, R. and Seigel, D. (1980) Senile lens and senile macular changes in a population-based sample. Am. J. Ophthalmol., 90, 86–91.[ISI][Medline]
6 Vinding, T. (1989) Age-related macular degeneration. Macular changes, prevalence and sex ratio. An epidemiological study of 1000 aged individuals. Acta Ophthalmol., 67, 609–616.
7 Klein, R., Klein, B.E. and Moss, S.E. (1998) Relation of smoking to the incidence of age-related maculopathy. The Beaver Dam Eye Study. Am. J. Epidemiol., 147, 103–110.
8 Kornzweig, A. (1977) Changes in the choriocapillaris associated with senile macular degeneration. Ann. Ophthalmol., 9, 753–756.[ISI][Medline]
9 Bressler, N., Bressler, S. and Fine, S. (1988) Age-related macular degeneration. Surv. Ophthalmol., 32, 375–413.[ISI][Medline]
10 Sarks, S. (1976) Aging and degeneration in the macular region: a clinicopathological study. Br. J. Ophthalmol., 60, 324–341.
11 Sarks, S., Van Driel, D., Maxwell, L. and Killingsworth, M. (1980) Softening of drusen and subretinal neovascularization. Trans. Ophthalmol. Soc. UK, 100, 414–422.[ISI][Medline]
12 Sarks, S.H. (1980) Drusen and their relationship to senile macular degeneration. Aust. J. Ophthalmol., 8, 117–130.[ISI][Medline]
13 Sarks, S. (1982) Drusen in patients predisposing to geographic atrophy of the retinal pigment epithelium. Aust. J. Ophthalmol., 10, 91–97.[ISI][Medline]
14 Sarks, J., Sarks, S. and Killingsworth, M. (1988) Evolution of geographic atrophy of the retinal pigment epithelium. Eye, 2, 552–577.
15 Chisholm, I. (1983) The recurrence of neovascularization and late visual failure in senile disciform lesions. Trans. Ophthalmol. Soc. UK, 103, 354–359.
16 Gorin, M., Sarneso, C., Paul, T., Ngo, J. and Weeks, D. (1993) The genetics of age-related maculopathy. In Hollyfield, J., Anderson, R. and LaVail, M. (eds), Retinal Degeneration. Plenum Press, New York, NY, pp. 35–47.
17 Seddon, J.M., Ajani, U.A. and Mitchell, B.D. (1997) Familial aggregation of age-related maculopathy. Am. J. Ophthalmol., 123, 199–206.[ISI][Medline]
18 18. Klaver, C.C.W. (2000) Genetic epidemiologic studies on age-related maculopathy: a population-based approach. PhD thesis, Erasmus University, Rotterdam, The Netherlands.
19 Gottfredsdottir, M.S., Sverrisson, T., Musch, D.C. and Stefansson, E. (1999) Age related macular degeneration in monozygotic twins and their spouses in Iceland. Acta Ophthalmol. Scand., 77, 422–425.[ISI][Medline]
20 Meyers, S.M., Greene, T. and Gutman, F.A. (1995) A twin study of age-related macular degeneration. Am. J. Ophthalmol., 120, 757–766.[ISI][Medline]
21 Sverrisson, T., Gottfredsdottir, M.S. and Stefansson, E. (1995) Age-related macular degeneration in monozygotic twins and their spouses. Invest. Ophthalmol. Vis. Sci., 36, S10.
22 Klein, M.L., Mauldin, W.M. and Stoumbos, V.D. (1994) Heredity and age-related macular degeneration. Observations in monozygotic twins. Arch. Ophthalmol., 112, 932–937.[Abstract]
23 Heiba, I.M., Elston, R.C., Klein, B.E. and Klein, R. (1994) Sibling correlations and segregation analysis of age-related maculopathy: the Beaver Dam Eye Study. Genet. Epidemiol., 11, 51–67.[ISI][Medline]
24 Gass, J.D.M. (1973) Drusen and disciform macular detachment and degeneration. Arch. Ophthalmol., 90, 206–217.[ISI][Medline]
25 Gorin, M.B., Jackson, K.E., Ferrell, R.E., Sheffield, V.C., Jacobson, S.G., Gass, J.D., Mitchell, E. and Stone, E.M. (1995) A peripherin/retinal degeneration slow mutation (Pro-210-Arg) associated with macular and peripheral retinal degeneration. Ophthalmology, 102, 246–255.[ISI][Medline]
26 Kemp, C.M., Jacobson, S.G., Cideciyan, A.V., Kimura, A.E., Sheffield, V.C. and Stone, E.M. (1994) RDS gene mutations causing retinitis pigmentosa or macular degeneration lead to the same abnormality in photoreceptor function. Invest. Ophthalmol. Vis. Sci., 35, 3154–3162.
27 Wells, J., Wroblewski, J., Keen, J., Inglehearn, C., Jubb, C., Eckstein, A., Jay, M., Arden, G., Bhattacharya, S., Fitzke, F. et al. (1993) Mutations in the human retinal degeneration slow (RDS) gene can cause either retinitis pigmentosa or macular dystrophy. Nature Genet., 3, 213–218.[ISI][Medline]
28 Wroblewski, J.J., Wells, J.A.R., Eckstein, A., Fitzke, F., Jubb, C., Keen, T.J., Inglehearn, C., Bhattacharya, S., Arden, G.B., Jay, M. et al. (1994) Macular dystrophy associated with mutations at codon 172 in the human retinal degeneration slow gene. Ophthalmology, 101, 12–22.[ISI][Medline]
29 Weber, B.H., Vogt, G., Pruett, R.C., Stohr, H. and Felbor, U. (1994) Mutations in the tissue inhibitor of metalloproteinases-3 (TIMP3) in patients with Sorsby’s fundus dystrophy. Nature Genet., 8, 352–356.[ISI][Medline]
30 Weber, B.H.F., Felbor, U., Schneider, U. and Doepner, D. (1995) Mutational analysis of TIMP3 in Sorsby’s fundus dystrophy (SFD) and age-related macular dystrophy (AMD). Invest. Ophthalmol. Vis. Sci., 36, S1064.
31 De La Paz, M.A., Pericak-Vance, M.A., Lennon, F., Haines, J.L. and Seddon, J.M. (1997) Exclusion of TIMP3 as a candidate locus in age-related macular degeneration. Invest. Ophthalmol. Vis. Sci., 38, 1060–1065.
32 Felbor, U., Doepner, D., Schneider, U., Zrenner, E. and Weber, B.H. (1997) Evaluation of the gene encoding the tissue inhibitor of metalloproteinases-3 in various maculopathies. Invest. Ophthalmol. Vis. Sci., 38, 1054–1059.
33 Allikmets, R., Shroyer, N., Singh, N., Seddon, J., Lewis, R., Bernstein, P., Peiffer, A., Zabriskie, N., Li, Y., Hutchinson, A., Dean, M., Lupski, J. and Leppert, M. (1997) Mutation of the Stargardt disease gene (ABCR) in age-related macular degeneration. Science, 277, 1805–1807.
34 De La Paz, M.A., Guy, V.K., Abou-Donia, S., Heinis, R., Bracken, B., Vance, J.M., Gilbert, J.R., Gass, J.D., Haines, J.L. and Pericak-Vance, M.A. (1999) Analysis of the Stargardt disease gene (ABCR) in age-related macular degeneration. Ophthalmology. 106, 1531–1536.[ISI][Medline]
35 Souied, E.H., Ducroq, D., Gerber, S., Ghazi, I., Rozet, J.M., Perrault, I., Munnich, A., Dufier, J.L., Coscas, G., Soubrane, G. and Kaplan, J. (1999) Age-related macular degeneration in grandparents of patients with Stargardt disease: genetic study. Am. J. Ophthalmol., 128, 173–178.[ISI][Medline]
36 Stone, E., Webster, A., Vandenburgh, K., Streb, L., Hockey, R., Lotery, A. and Sheffield, V. (1988) Allelic variation in ABCR associated with Stargardt disease but not age-related macular degeneration. Nature Genet., 20, 328–329.[ISI][Medline]
37 Klaver, C.C., Kliffen, M., van Duijn, C., Hofman, A., Cruts, M., Grobbee, D.E., van Broeckhoven, C. and de Jong, P.T. (1998) Genetic association of apolipoprotein E with age-related macular degeneration. Am. J. Hum. Genet., 63, 200–206.[ISI][Medline]
38 Souied, E.H., Benlian, P., Amouyel, P., Feingold, J., Lagarde, J.P., Munnich, A., Kaplan, J., Coscas, G. and Soubrane, G. (1998) The epsilon4 allele of the apolipoprotein E gene as a potential protective factor for exudative age-related macular degeneration. Am. J. Ophthalmol., 125, 353–359.[ISI][Medline]
39 Klein, M.L., Schultz, D.W., Edwards, A., Matise, T.C., Rust, K., Berselli, C.B., Trzupek, K., Weleber, R.G., Ott, J., Wirtz, M.K. and Acott, T.S. (1998) Age-related macular degeneration. Clinical features in a large family and linkage to chromosome 1q. Arch. Ophthalmol., 116, 1082–1088.
40 Kniazeva, M., Traboulsi, E.I., Stefko, S.T., Gorin, M.B., Blaschak, C.J., Cutting, G., Han, M. and Zhang, K. (2000) A new locus for drusen with macular degeneration maps to chromosome 6. Am. J. Ophthalmol, in press.
41 Hyman, L.G., Lilienfeld, A.M., Ferris, F.L.I. and Fine, S.L. (1983) Senile macular degeneration: a case-control study. Am. J. Epidemiol., 118, 213–227.
42 Leibowitz, H.M., Krueger, D.E., Maunder, L.R., Milton, R.C., Kini, M.M., Kahn, H.A., Nickerson, R.J., Pool, J., Colton, T.L., Ganley, J.P. et al. (1980) The Framingham Eye Study monograph: an ophthalmological and epidemiological study of cataract, glaucoma, diabetic retinopathy, macular degeneration and visual acuity in a general population of 2631 adults, 1973–1975. Surv. Ophthalmol., 24, 335–610.
43 Wu, L. (1987) Study of aging macular degeneration in China. Jpn. J. Ophthalmol., 31, 349–367.[Medline]
44 Vinding, T. (1990) Visual impairment of age-related macular degeneration. An epidemiological study of 1000 aged individuals. Acta Ophthalmol., 68, 162–167.[Medline]
45 Keverline, M.R., Mah, T.S., Keverline, P.O. and Gorin, M.B. (1998) A practice-based survey of familial age-related maculopathy. Ophthalmic Genet., 19, 19–26.[ISI][Medline]
46 Stone, E.M., Fingert, J.H., Alward, W.L.M., Nguyen, T.D., Polansky, J.R., Sunden, S.L.F., Nishimura, D., Clark, A.F., Nystuen, A., Nichols, B.E. et al. (1997) Identification of a gene that causes primary open angle glaucoma. Science, 275, 668–670.
47 De La Paz, M. and Anderson, R.E. (1992) Region and age-dependent variation in susceptibility of the human retina to lipid peroxidation. Invest. Ophthalmol. Vis. Sci., 33, 3497–3499.
48 Yoshimura, S., Suemizu, H., Taniguchi, Y., Arimori, K., Kawabe, N. and Moriuchi, T. (1994) The human plasma glutathione peroxidase-encoding gene: organization, sequence and localization to chromosome 5q32. Gene, 145, 293–297.[ISI][Medline]
49 Atalla, L.R., Sevanian, A. and Rao, N.A. (1990) Immunohistochemical localization of peroxidative enzymes in ocular tissue. CLAO J., 16, S30–S33.
50 Castorina, C., Campisi, A., Di Giacomo, C., Sorrenti, V., Russo, A. and Vanella, A. (1992) Lipid peroxidation and antioxidant enzymatic systems in rat retina as a function of age. Neurochem. Res., 17, 599–604.[ISI][Medline]
51 Samiec, P.S., Drews-Botsch, C., Flagg, E.W., Kurtz, J.C., Sternberg Jr, P., Reed, R.L. and Jones, D.P. (1998) Glutathione in human plasma: decline in association with aging, age-related macular degeneration and diabetes. Free Radic. Biol. Med., 24, 699–704.[ISI][Medline]
52 Conley, Y.P., Gorin, M.B., Mah, T.S., Weeks, D.E. and Ferrell, R.E. (1999) Glutathione peroxidase 3: identification of promoter region polymorphisms and association with age-related maculopathy. Am. J. Hum. Genet., 65, A104.
53 Douglas, J.A., Boehnke, M. and Lange, K. (2000) A multipoint method for detecting genotyping errors and mutations in sibling-pair linkage data. Am. J. Hum. Genet., 66, 1287–1297.[ISI][Medline]
54 Bird, A.C., Bressler, N.M., Bressler, S.B., Chisholm, I.H., Coscas, G., Davis, M.D., de Jong, P.T., Klaver, C.C., Klein, B.E., Klein, R. et al. (1995) An international classification and grading system for age-related maculopathy and age-related macular degeneration. The International ARM Epidemiological Study Group. Surv. Ophthalmol., 39, 367–374.[ISI][Medline]
55 Miller, S., Dykes, D. and Polesky, H. (1988) A simple salting out procedure for extracting DNA from human nucleated cells. Nucleic Acids Res., 16, 1215.
56 Allikmets, R. and the International ABCR Screening Consortium (1999) Statistically significant association of sequence variants in the ABCR gene with age-related macular degeneration. Am. J. Hum. Genet., 65, A103.
57 O’Connell, J.R. and Weeks, D.E. (1998) PedCheck: a program for identifying genotype incompatibilities in linkage analysis. Am. J. Hum. Genet., 63, 259–266.[ISI][Medline]
58 Weeks, D. and Lange, K. (1988) The affected-pedigree-member method of linkage analysis. Am. J. Hum. Genet., 42, 315–326.[ISI][Medline]
59 Sobel, E. and Lange, K. (1996) Descent graphs in pedigree analysis: applications to haplotyping, location scores and marker-sharing statistics. Am. J. Hum. Genet., 58, 1323–1337.[ISI][Medline]
60 Boehnke, M. and Cox, N.J. (1997) Accurate inference of relationships in sib-pair linkage studies. Am. J. Hum. Genet., 61, 423–429.[ISI][Medline]
61 Boehnke, M. (1991) Allele frequency estimation from data on relatives. Am. J. Hum. Genet., 48, 22–25.[ISI][Medline]
62 Ott, J. (1992) Strategies for characterizing highly polymorphic markers in human gene mapping. Am. J. Hum. Genet., 51, 283–290.[ISI][Medline]
63 Göring, H.H.H. and Terwilliger, J.D. (2000) Linkage analysis in the presence of errors III: Marker loci and their map as nuisance parameters. Am. J. Hum. Genet., 66, 1298–1309.[ISI][Medline]
64 Hodge, S.E., Anderson, C.E., Neiswanger, K., Sparkes, R.S. and Rimoin, D.L. (1983) The search for heterogeneity in insulin-dependent diabetes mellitus (IDDM): linkage studies, two-locus models and genetic heterogeneity. Am. J. Hum. Genet., 5, 1139–1155.
65 Lathrop, G.M. and Lalouel, J.M. (1988) Efficient computations in multilocus linkage analysis. Am. J. Hum. Genet., 42, 498–505.[ISI][Medline]
66 Lathrop, G.M., Lalouel, J.M., Julier, C. and Ott, J. (1984) Strategies for multilocus linkage analysis in humans. Proc. Natl Acad. Sci. USA, 81, 3443–3446.
67 Cottingham, R.W., Idury, R.M. and Schäffer, A.A. (1993) Faster sequential genetic linkage computations. Am. J. Hum. Genet., 53, 252–263.[ISI][Medline]
68 Schäffer, A.A., Gupta, S.K., Shriram, K. and Cottingham, R.W. (1994) Avoiding recomputation in linkage analysis. Hum. Hered., 44, 225–237.[ISI][Medline]
69 Schäffer, A.A. (1996) Faster linkage analysis computations for pedigrees with loops or unused alleles. Hum. Hered., 46, 226–235.[ISI][Medline]
70 Ott, J. (1986) Linkage probability and its approximate confidence interval under possible heterogeneity. Genet. Epidemiol., Suppl. 1, 251–257.
71 Kruglyak, L., Daly, M.J., Reeve-Daly, M.P. and Lander, E.S. (1996) Parametric and nonparametric linkage analysis: a unified multipoint approach. Am. J. Hum. Genet., 58, 1347–1363.[ISI][Medline]
72 Davis, S., Schroeder, M., Goldin, L.R. and Weeks, D.E. (1996) Nonparametric simulation-based statistics for detecting linkage in general pedigrees. Am. J. Hum. Genet., 58, 867–880.[ISI][Medline]
73 Davis, S. and Weeks, D.E. (1997) Improvements to SimIBD, a nonparametric method for detecting linkage in general pedigrees. Am. J. Hum. Genet., 61 (suppl.), A272.
74 Davis, S., Goldin, L.R. and Weeks, D.E. (1997) SimIBD: a powerful robust nonparametric method for detecting linkage in general pedigrees. In Pawlowitzki, I.H., Edwards, J.H. and Thompson, E.A. (eds), Genetic Mapping of Disease Genes. Academic Press, London, UK, pp. 189–204.
75 Kong, A. and Cox, N.J. (1997) Allele-sharing models: LOD scores and accurate linkage tests. Am. J. Hum. Genet., 61, 1179–1188.[ISI][Medline]
76 Hinds, D. and Risch, N. (1996) The ASPEX package: affected sib-pair mapping. ftp://lahmed.stanford.edu/pub/aspex
77 Ott, J. (1989) Computer-simulation methods in human linkage analysis. Proc. Natl Acad. Sci. USA, 86, 4175–4178.
78 Weeks, D.E., Ott, J. and Lathrop, G.M. (1990) SLINK: a general simulation program for linkage analysis. Am. J. Hum. Genet., 47, A204.
79 Terwilliger, J.D., Speer, M. and Ott, J. (1993) Chromosome-based method for rapid computer simulation in human genetic linkage analysis. Genet. Epidemiol., 10, 217–224.[ISI][Medline]
80 Kaplan, J., Gerber, S., Larget-Piet, D., Rozet, J.M., Dollfus, H., Dufier, J.L., Odent, S., Postel-Vinay, A., Janin, N., Briard, M.L. et al. (1993) A gene for Stargardt’s disease (fundus flavimaculatus) maps to the short arm of chromosome 1. Nature Genet., 5, 308–311.[ISI][Medline]
81 Heon, E., Piguet, B., Munier, F., Sneed, S.R., Morgan, C.M., Forni, S., Grasiano, P., Schorderet, D., Taylor, C.M., Streb, L.M., Wiles, C.D., Nishimura, D.Y., Sheffield, V.C. and Stone, E.M. (1996) Linkage of autosomal dominant radial drusen (Malattia Leventinese) to chromosome 2p16-21. Arch. Ophthalmol., 114, 193–198.[Abstract]
82 Dryja, T., McGee, T., Reichel, E., Hahn, L., Cowley, G., Yandell, D., Sandberg, M. and Berson, E. (1990) A point mutation of the rhodopsin gene in one form of retinitis pigmentosa. Nature, 343, 364–366.[Medline]
83 Kajiwara, K., Hahn, L.B., Mukai, S., Travis, G.H., Berson, E.L. and Dryja, T.P. (1991) Mutations in the human retinal degeneration slow gene in autosomal dominant retinitis pigmentosa.