Advertisement

BMC Medical Genomics

, 11:70 | Cite as

Local genetic ancestry in CDKN2B-AS1 is associated with primary open-angle glaucoma in an African American cohort extracted from de-identified electronic health records

  • Nicole A. Restrepo
  • Sarah M. Laper
  • Eric Farber-Eger
  • Dana C. Crawford
Open Access
Research

Abstract

Background

Glaucoma is a leading cause of blindness in developed countries. Primary open-angle glaucoma (POAG), the most prevalent clinical subtype of glaucoma in the United States, affects African Americans at a higher rate compared with European Americans. Risk factors identified for POAG include increased age and family history, which coupled with heritability estimates, suggest this complex condition is associated with genetic and environmental factors. To date, several genome-wide studies have identified loci significantly associated with POAG risk, but most of these studies were performed in populations of European-descent.

Methods

To identify population-specific and trans-population genetic associations for POAG, we genotyped 11,521 African Americans using the Illumina Metabochip as part of the Epidemiologic Architecture for Genes Linked to Environment (EAGLE) study accessing BioVU, the Vanderbilt University Medical Center’s biorepository linked to de-identified electronic health records. Among this study population, we identified 138 cases of POAG and 1376 controls and performed Metabochip-wide tests of association. We also estimated local genetic ancestry at CDKN2B-AS1, a POAG-associated locus established in European-descent populations.

Results

Overall, we did not identify significant single SNP-POAG associations after adjusting for multiple testing. We did, however, detect a significant association between POAG risk and local African genetic ancestry at CDKN2B-AS1, where on average cases were of 90% African descent compared with controls at 58% (p = 2 × 10− 6).

Conclusions

These data suggest that CDKN2B-AS1 is an important locus for POAG risk among African Americans, warranting further investigation to identify the variants underlying this association.

Keywords

African Americans Primary open-angle glaucoma Electronic health records 

Abbreviations

EAGLE

Epidemiologic Architecture for Genes Linked to Environment

GWAS

Genome-wide association studies

HDL-C

High density lipoprotein cholesterol

IRB

Vanderbilt Institutional Review Board

LAMP

Local Ancestry in admixed Populations

MAF

Minor allele frequency

OR

Odds ratio

PAGE

Population Architecture using Genomics and Epidemiology

PC

Principal components

POAG

Primary open-angle glaucoma

VUMC

Vanderbilt University Medical Center

Background

Glaucoma is the second leading cause of blindness in the United States, and it is the leading cause of blindness and irreversible vision loss in African Americans [1], with a prevalence approximately double that observed in European-descent populations [1, 2, 3]. The prevalence of glaucoma is similar for European, Japanese, and Indian populations with rates approaching those observed in African descent populations in the oldest age categories [4]. Although African Americans comprise the group of highest risk of developing glaucoma-related vision problems, many cases remain undiagnosed. Previous studies have suggested that nation-wide implementation of screening middle aged African Americans could decrease the rate of undiagnosed glaucoma from 50 to 27% [5]. Earlier screening and diagnosis enables patients to more effectively leverage current treatment options to reduce the risk of bilateral blindness later in life [5].

In addition to African ancestry and age [6], other known risk factors associated with the development of glaucoma include myopia [7] and high intraocular pressure [6, 8, 9]. Family history has also been associated with glaucoma risk [6, 10, 11], albeit inconsistently most likely due to the heterogeneous nature of the disease. The phenotypic heterogeneity of glaucoma has also impacted other studies attempting to establish and quantify the genetic contribution to risk in developing the disease; consequently, the majority of these studies have been conducted on more easily-measured glaucoma endophenotypes such as central corneal thickness (h2 = 0.35–0.72%) [12, 13, 14], intraocular pressure (h2 = 0.35–0.94%) [12, 14], and cup-to-disc ratio (h2 = 0.56–0.66%) [12, 15]. Pulsatility of choroidal blood flow and velocity are additional quantitative traits whose variation from normal parameters has been observed in individuals with glaucoma [16, 17], yet heritability studies have not yet found significant genetic contribution to its variability [12].

The strongest evidence for a genetic contribution related to glaucoma comes from studies of primary open-angle glaucoma (POAG), the most prevalent clinical subtype of glaucoma in the United States. Early linkage and family-based genetic association studies identified the MYOC (myocilin), OPTN (optineurin), and WDR36 (WD repeat domain 36) [18, 19, 20] genes as the primary genes for susceptibility to POAG. Mutations in MYOC are known to cause hereditary early-onset POAG in multiple populations [18, 21, 22]. More recently, large-scale genome-wide association studies (GWAS) have identified variants in the CAV1/CAV2, CDKN2B-AS1 and SIX1/SIX6I genes that influence POAG risk in European-descent and Japanese populations [23, 24, 25, 26, 27].

Additional genetic factors that have yet to be discovered are hypothesized to drive POAG risk and to account for the differences in incidence observed across racial/ethnic groups. For example, in a study of African Americans, the frequency of MYOC mutations was comparably lower (~ 1.4%) than in other populations (~ 2–4%) [28] suggesting that other genetic loci are driving risk in this group. It is possible that both population-specific and trans-population genetic variants contribute to POAG risk. To identify population-specific and trans-population genetic factors, we conducted a hypothesis-testing and hypothesis-generating genetic association study in African Americans with and without POAG drawn from a clinical cohort with electronic health records.

Methods

Study population and genotyping

The study population is a subset of the Epidemiologic Architecture for Genes Linked to Environment (EAGLE) study, a study site of the larger Population Architecture using Genomics and Epidemiology (PAGE) I study [29, 30]. In general, the PAGE study is a consortium of diverse epidemiologic and clinical cohorts with broad research goals that include the generalization of genetic associations to multiple populations [29]. To identify non-European Americans for PAGE I, EAGLE accessed the Vanderbilt University Medical Center (VUMC)‘s biorepository linked to de-identified electronic health records known as BioVU [31].

VUMC’s BioVU followed an opt-out model for DNA sample accrual between 2007 and 2015 [31]. That is, DNA was collected from discarded blood samples remaining after routine clinical testing and was linked to de-identified electronic health records. According to the Vanderbilt Institutional Review Board (IRB) and the Federal Office of Human Research Protections provisions, this VUMC protocol is considered nonhuman subjects research (The Code of Federal Regulations, 45 CFR 46.102 (f)) [31, 32].

As previously described [33], EAGLE selected all non-European Americans from BioVU as of 2011 for genotyping on the Metabochip (EAGLE BioVU). A total of 11,521 African Americans samples in EAGLE BioVU were genotyped [33]. From among these patients, billing and procedural codes along with text searches were used to identify POAG cases (n = 138) and controls (n = 1376). In short, controls included patients in BioVU over the age of 60 years whose records did not contain an ICD-9-CM code for any form of glaucoma nor any mention of “glaucoma” in a text search of their ‘Problems List.’ Manual review of all cases and a subset of controls was performed for quality assurance as previously described [34].

The Metabochip is an Illumina (San Diego, CA) custom array designed for fine-mapping of metabolic and cardiovascular traits. Fine-mapping regions cover 257 loci chosen from SNPs that reached genome-wide significance from select consortium meta-analyses [35]. The Metabochip was also designed for replication of GWAS-identified index variants for any phenotype from the GWAS Catalog (http://www.ebi.ac.uk/gwas/) as of 2009. A total of 33 GWAS-index variants representing ocular diseases (including age-related macular degeneration, POAG, normal tension glaucoma, and diabetic retinopathy) as well as related traits (myopia, ocular axial length, HbA1c, cup-to-disc ratio, intraocular pressure, and optic disc size) are directly assayed by the Metabochip (Additional file 1: Table S1).

EAGLE BioVU DNA samples were genotyped using the Metabochip following the manufacturer’s protocol (Illumina, Inc.; San Diego, CA.), and 360 HapMap samples, including YRI samples, were genotyped for PAGE-wide cross-study quality control standards [36]. A description of the genotyping protocols and quality control measures has been previously published [30]. In brief, genetic variants were evaluated for deviations from Hardy Weinberg Equilibrium, which may be a result of poor genotyping. Variants with a genotyping call rate < 95% were removed from further analysis. Principal components (PC) were calculated using EIGENSOFT [37, 38]. At the sample level, DNA samples with poor sample call rate (< 95%), sex discordance, or evidence of cryptic relatedness were removed from analyses.

Statistical methods

Individuals included in this analysis were those identified as POAG cases over the age of 20 years and POAG controls over the age of 60 years. An older age threshold was applied in controls to minimize the probability of including potential future cases. African Americans are at increased risk of glaucoma over the age of 40 years, while other populations have an age-associated risk over 60 years. Age was defined as age at diagnosis in cases and age at last clinical exam in controls. T-tests and chi-square tests, where appropriate, were used to compare demographic clinical characteristics between cases and controls in Stata/SE version 14.2.

All common variants (MAF > 0.05) were tested for an association with POAG using logistic regression separately assuming a log-additive genetic model (Additional file 1: Table S2), a recessive model (Additional file 1: Table S3), and a dominant model (Additional file 1: Table S4) adjusted by age, sex, the first three PCs, and median diastolic blood pressure. Analyses were conducted using PLINKv1.90 [39]. Additionally, we tested for an association between POAG and 258 SNPs that passed quality control in the CDKN2B-AS1 region of chromosome 14. Pair-wise linkage disequilibrium (r2) was calculated in SNAP [40] using YRI 1000 Genomes Project Pilot 1 reference data. Power calculations were performed in Quanto [41] to determine 80% power to detect an association with a case:control ratio of 1:3, assuming a log-additive model and a genome-wide significance threshold (5 × 10− 8).

Local ancestry mapping

Local ancestry for the CDKN2B-AS1 region, located on chromosome 9, was determined for the POAG cases and controls using Local Ancestry in adMixed Populations (LAMP) [42]. Input parameters included the estimated number of generations since admixture (generations = 10), estimated fraction of admixture from each population (African = 0.8, European = 0.2), and predicted recombination rate (3.4 × 10− 7 bases− 1). The number of alleles from the ancestral populations at each SNP that was genotyped in this region was estimated, and the overall fraction of alleles from each ancestral population in this region for each patient was determined. Percent African ancestry was then tested for an association with POAG status using logistic regression using R software version 3.1.3 [43].

Global ancestry mapping

Global ancestry was calculated using fastSTRUCTURE [44] with all of the Metabochip data. Input parameters were set to default, as recommended by the authors, and the analysis was set to determine the proportion of two populations (K = 2). Admixture plots were graphed using the web graphical interface (http://pophelper.com/) of the R module “pophelper” [45].

Results

Population characteristics

A total of 138 African American POAG cases and 1376 controls passed quality control in EAGLE BioVU for the present study. We previously described [34] these cases compared with 4813 controls over the age of 40; in the present study, we compare the same cases with subset of the controls over the age of 60 years (Table 1). Here, cases were younger (p = 0.01) with higher cholesterol levels (183 mg/dL; p = 0.01), more likely to be female with higher average body mass index (30.1 kg/m2) in comparison to controls (28.8 kg/m2 and 169 mg/dL, respectively). Cases also presented with higher triglyceride levels compared to controls (125 mg/dL versus 97 mg/dL; p = 0.0001).
Table 1

Study population characteristics of primary open-angle glaucoma cases and controls among African Americans in EAGLE BioVU

Trait

Cases (n = 138)

Controls (1376)

p-value

Median age at diagnosis or LCV (years)

62.0 (12.0)

67.3 (7.8)

0.01

% female

63.7

56.5

0.10

% hypertensive

55.1

52.5

0.50

Median BMI (kg/m2)

30.1 (6.7)

28.8 (7.35)

0.44

Median diastolic blood pressure (mm/Hg)

74.5 (8.1)

76.0 (8.8)

0.97

Median systolic blood pressure (mm/Hg)

134.5 (14.1)

135 (14.6)

0.88

Median cholesterol (mg/dL)

183 (40.6)

169 (46.7)

0.01

Median HDL-C (mg/dL)

52.5 (25.0)

49 (17.8)

0.88

Median LDL-C (mg/dL)

103 (42.9)

93 (37.4)

0.20

Median triglycerides (mg/dL)

125 (76.3)

97 (68.1)

0.0001

Case extraction was described in Restrepo et al. [34]. Control extraction from EAGLE BioVU was also described in Restrepo et al. [34] but restricted to controls > 40 years of age (as opposed to > 60 years of age here). Values were defined or calculated for the following: Age at POAG diagnosis was determined by the date of when the POAG billing code (ICD-9-CM 365.11) was first mentioned in the records. Age at last clinic visit (LCV) was taken as the date of the last current procedure terminology (CPT) code mentioned in the records for controls. An individual was classified as hypertensive if he/she met one of three criteria: systolic blood pressure > 140 mm/Hg, diastolic blood pressure > 90 mm/Hg, or on hypertension medications all within a 2-year window of when he or she was diagnosed with POAG for cases and a 2-year window of his or her LCV date for controls. Median (and standard deviation) blood pressure (systolic and diastolic), lipids (total cholesterol, high-density lipoprotein cholesterol or HDL-C, low-density lipoprotein cholesterol or LDL-C, and triglycerides), and body mass index (height and weight) were calculated from all labs or measurements available within 2 years of POAG diagnosis or LCV. T-tests and chi-square tests, where appropriate, were used to compare demographic clinical characteristics between cases and controls. Abbreviations: standard deviation (SD)

Local ancestry in the CDKN2B-AS1 region

We previously reported on preliminary tests of association in the CDKN2B-AS1 region [34], which was fine-mapped by the Illumina Metabochip. None of the tests of association were significant after correcting for the 258 common variants tested [34]. As we have already noted, there are multiple explanations for the lack of significant results including limited power, variability in linkage disequilibrium patterns across populations, and genetic heterogeneity. Another possible explanation for the observed null results is that the previous analysis did not account for local genetic ancestry. We therefore sought to determine whether the total composition of African and European ancestry at this region could account for POAG risk.

Local ancestry for the CDKN2B-AS1 region was determined for POAG cases and controls using LAMP [42]. The number of alleles from the ancestral populations at each SNP was estimated, and the overall fraction of alleles from each ancestral population in this region for each patient was calculated. Logistic regression was then performed between POAG case status and percent African ancestry to assess whether ancestry might alter POAG risk. The mean African ancestry for POAG cases and controls at CDKN2B-AS1 was 0.90 and 0.58, respectively (Fig. 1), and the percent of African ancestry in the CDKN2B-AS1 region was significantly associated with POAG at p = 2 × 10− 6. In contrast, the average Metabo-wide global African ancestry for cases and controls was 81.5 and 79.4%, respectively, in agreement with previous estimates [46, 47].
Fig. 1

Distribution of African ancestry at CDKN2B-AS1 in African American primary open-angle glaucoma cases and controls from EAGLE BioVU. Fraction of African ancestry, estimated by LAMP using genotype data available for CDKN2B-AS1 from the Illumina Metabochip, is plotted on the x-axis with frequency on the y-axis for primary open-angle glaucoma (POAG) a cases (n = 138) and b controls (n = 1376). Plots were graphed in R

Metabochip-wide association of POAG in African Americans

We tested all SNPs genotyped on the Illumina Metabochip for an association with POAG adjusted for age, sex, the first three principal components, and median diastolic blood pressure (Fig. 2; Additional file 1: Tables S2-S4). No SNP was significantly associated with POAG after adjusting for a strict Bonferroni correction (p < 4.04 × 10− 7). The two most significant associations [chr1:228347779 (rs4846835) and chr1:228354829 (rs34783939)] under the log-additive genetic model are located within the protein coding gene of GALNT2, a member of the glycosyltransferase 2 protein family. GALNT2 was targeted for fine-mapping by the Metabochip based on earlier reported associations with high density lipoprotein cholesterol (HDL-C) and triglyceride levels [48, 49]. It is interesting to note that GALNT2 rs4846835 was associated with dementia and core Alzheimer’s disease neuropathologic changes, albeit not at the genome-wide level [50]. Both of these variants also appear as marginally significant under a dominant genetic model (Table 3 [OR = 2.43 & 2.24 respectively]. Homozygous carriers for either of the two SNPs are rare in cases and controls at only a frequency of 1 to 2%. Variant rs4846835 heterozygotes account for 9.4% of cases and 17.5% of controls, while rs34783939 heterozygotes make up 12.3% of cases and 23.6% of controls. The variants are not in strong linkage disequilibrium with one another (r2 = 0.304 in YRI, phase 1 1000 Genomes Project). It is important to note that rs34783939 is most likely multi-allelic based on later versions of the 1000 Genomes Project and other large-scale sequencing efforts.
Fig. 2

Manhattan plot of EAGLE BioVU primary open-angle glaucoma Metabochip-wide tests of association in African Americans. Logistic regression assuming an additive genetic model was performed for 138 cases and 1376 controls adjusted by age, sex, principal components, and median diastolic blood pressure. P-values [(−log10) on the y-axis] for each test of association are plotted by chromosome (x-axis). The blue line depicts a suggestive significance threshold of p < 5.0 × 10− 4

Additional variants of interest for future studies that were marginally significant in both additive and dominant genetic modes are rs13423742 (ORadditive = 3.04; p = 1.14 × 10− 4), rs9479726 (ORadditive = 0.41; p = 1.54 × 10− 4), and rs1671152 (ORadditive = 1.91; p = 1.6 × 10− 4). Variant rs1671152 is a known missense variant in the glycoprotein VI (GP6) gene. GP6, a collagen receptor, is involved in platelet aggregation [54]. GP6 RNAs are expressed in the retina and brain as shown in the FANTOM5 and GTEx datasets [55, 56]. In 1000 Genomes Project phase 3 CEU the rs1671152 (A) allele has a frequency of 0.182 while the YRI population has a frequency of 0.324, consistent with that observed in EAGLE BioVU African Americans (coded allele frequency = 0.32). The frequency of rs1671152 is twice as high in African Americans compared with European Americans suggesting it could potentially be a population-specific factor.

The intergenic SNP rs9479726 was less frequent in cases compared with controls (ORadditive = 0.41; 1.54 × 10–4 and ORdominant = 0.42; 2.5 × 10− 4). None of the cases were found to be homozygous for the coded allele, while 5.7% of controls were homozygous. The ‘A’ allele had a frequency of 24% in the overall population, with 10.8% of cases and 37.6% of controls being heterozygous carriers.

One SNP (i.e., rs7454156) was consistently associated with POAG at p < 5.0 × 10− 4 in both the additive (Table 2) and recessive genetic models (Table 3). This intronic variant (G) in the bone morphogenetic protein 6 (BMP6) was found to be homozygous in 5% of controls and 2.5% of cases. In a mouse model of hemochromatosis, mutations in a BMP6 co-receptor (i.e., HJV) were found to result in an accumulation of iron in mouse retinal tissues and upregulation of BMP6 along with upregulation of VEGF that resulted in subsequent abnormal vascularization of the retina [57].
Table 2

Ten most significant results for primary open-angle glaucoma Metabochip-wide genetic associations in African Americans

CHR

SNP

Gene

CA

CAF

OR

95% CI

p-value

1

rs4846835

GALNT2

A

0.11

2.37

1.56–3.60

5.00 × 10− 5

1

rs34783939

GALNT2

C

0.15

2.09

1.44–3.02

8.73 × 10− 5

21

rs9982695

C21orf33

A

0.24

2.09

1.44–3.02

8.74 × 10− 5

4

rs3775202

VEGFC

G

0.43

1.92

1.38–2.66

9.70 × 10−5

2

rs13423742

FN1

C

0.06

3.04

1.73–5.36

1.14 × 10− 4

6

rs7454156

BMP6

G

0.18

2.08

1.42–3.02

1.37 × 10− 4

6

rs9479726

RGS17-OPRM1

A

0.24

0.41

0.25–0.64

1.54 × 10−4

19

rs1671152

GP6

A

0.32

1.91

1.36–2.68

1.60 × 10−4

10

rs286489

LOC101929727

A

0.28

1.90

1.35–2.66

1.80 × 10−4

5

rs4336354

HTR4

G

0.09

2.51

1.54–4.07

1.86 × 10−4

Logistic regression assuming an additive genetic model was performed for 138 cases and 1376 controls adjusted by age, sex, principal components, and median diastolic blood pressure. For the ten most significant associations, chromosome (CHR), SNP ID (rs number), gene, coded allele (CA), coded allele frequency (CAF), odds ratio (OR), 95% confidence interval (CI), and p-value are given

Table 3

Primary open-angle glaucoma Metabochip-wide genetic associations in African Americans for dominant and recessive models that overlap with the most significant results for the additive genetic model

CHR

SNP

Gene

CA

OR

95% CI

p-value

Genetic model

1

rs4846835

GALNT2

A

2.43

1.56–3.74

6.59 × 10−5

dominant

1

rs34783939

GALNT2

C

2.24

1.47–3.39

1.6 × 10−4

dominant

2

rs13423742

FN1

C

2.82

1.63–4.86

2.02 × 10−4

dominant

6

rs9479726

RGS17-OPRM1

A

0.42

0.26–0.67

2.5 × 10−4

dominant

19

rs1671152

GP6

A

2.24

1.42–3.53

5.2 × 10−4

dominant

6

rs7454156

BMP6

G

5.14

2.54–10.37

4.87 × 10−6

recessive

Logistic regression assuming a dominant genetic model adjusted by age, sex, principal components, and median diastolic blood pressure. Chromosome (CHR), SNP ID (rs number), gene, coded allele (CA), odds ratio (OR), 95% confidence interval (CI), p-value, and assumed genetic model are given

While certain variants were found to be associated with POAG in both dominant/additive and recessive/additive models, we found that no SNPs were consistently associated with POAG in both dominant and recessive models.

Discussion

Epidemiologic and clinical studies have demonstrated that POAG risk is higher in African-descent populations compared with other populations such as European Americans. To identify genetic variants associated with POAG risk that are specific to African-descent populations or shared across world populations, we identified African American POAG cases and controls in a clinic setting using electronic health records to conduct genetic associations studies in the fine-mapped region of CDKN2B-AS1 and Metabochip-wide [34]. Overall, we found evidence that the percentage of African ancestry at CDKN2B-AS1 was strongly correlated with POAG case status (p = 2 × 10− 6). POAG cases on average contained 90% African ancestral alleles at the CDKN2B-AS1 region versus controls which were only 58% African, suggesting that African-specific variation may indeed being driving risk at this locus. Additionally, the lack of strong statistical associations with individual SNPs but an association with gene-based African ancestry suggests that gene x gene or gene x environment interactions may be involved but which will require larger sample sizes to accurately assess the possibility.

Common variants in CDKN2B-ASI are consistently associated with POAG in European-descent populations [25, 27, 51]. We [34] and others [27, 52, 53] have demonstrated that these same variants are inconsistently associated with POAG in African-descent populations. The lack of association in African-descent populations is likely due to limited power, a consequence of smaller sample sizes and considerably lower allele frequencies compared with studies of European-descent populations. For example, CDKN2B-AS1 rs2157719 has a minor allele frequency of 3 and 46% in African Americans (ASW) and Europeans (CEU), respectively, in phase 3 of the 1000 Genomes Project. Originally discovered in a European POAG cohort [25], rs2157719 was found to be associated with optic nerve degeneration in glaucoma patients with an odds ratio of 1.45. The present study is powered (80%) to detect associations for common variants (≥15% MAF) with large effect sizes (at least 2.9 odds ratio) at genome-wide significance (5 × 10− 8). The small sample size of this study is underpowered to detect associations for less frequent variants and/or variants with smaller effect sizes. Although we could not generalize these associations in this study sample, we note that this locus is still important in POAG risk as evidenced by the association between African ancestry at this locus and POAG. It is also interesting to note that the CDKN2B-AS1 rs2157719 allele associated with lower odds of POAG (C/G) is the ancestral allele yet the minor allele in all 1000 Genomes Project populations. The high frequency of the derived allele at rs2157719 may be due to chance, positive selection (and possible antagonistic pleiotropy), or an error in ancestral allele assignment, among other possibilities.

Conclusions

Here, we show a significant association between POAG risk and local African genetic ancestry at CDKN2B-AS1 (p = 2 × 10− 6). While not identifying significant single SNP-POAG associations after adjusting for multiple testing, the results still suggest that CDKN2B-AS1 is an important locus of POAG risk among African Americans, warranting further investigation to identify genetic variants or epigenetic regulators that may be acting in conjunction with this locus. When gauging the strengths and limitations of this study, perhaps its greatest strength is the expansion of knowledge in African Americans, a population far too often underrepresented in biomedical research [58]. Additional strengths involve the utilization of electronic health records as a cost efficient and data-dense resource for studies. A major limitation of our study is statistical power. Nevertheless, this study is one of only a handful to assess the genetic architecture of POAG in African Americans.

Notes

Funding

This work was supported by National Institutes of Health U01 HG004798 and its ARRA supplements. The cost of publication was funded by Case Western Reserve University’ Institute for Computational Biology. NAR was supported by the National Institutes of Health (NIH) Quantitative Ocular Genomics Training Program Pre-doctoral Trainee (1T32EY021453). The dataset (s) used for the analyses described were obtained from Vanderbilt University Medical Center’s BioVU which is supported by institutional funding and the National Center for Research Resources, Grant UL1 RR024975–01 (now at the National Center for Advancing Translational Sciences, Grant 2 UL1 TR000445–06).

Availability of data and materials

Data are available in the database of Genotypes and Phenotypes (dbGaP) study accession phs000297.v1.p1.

About this supplement

This article has been published as part of BMC Medical Genomics Volume 11 Supplement 3, 2018: Selected articles from the 7th Translational Bioinformatics Conference (TBC 2017): medical genomics. The full contents of the supplement are available online at https://bmcmedgenomics.biomedcentral.com/articles/supplements/volume-11-supplement-3.

Authors’ contributions

DCC and NAR designed the study. NAR and EFE collected and prepared the data, and NAR and SL analyzed the data. NAR drafted the manuscript. DCC, SL, and EFE were major contributors in revising the manuscript critically for all important intellectual content. All authors gave approval to the final version of the manuscript and agreed to be accountable to all aspects of the work.

Ethics approval and consent to participate

The data in this study were de-identified in accordance with provisions of Title 45, Code of Federal Regulations, part 46 (45 CFR 46); therefore, this study was considered non-human subjects research by the Vanderbilt University Internal Review Board.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary material

12920_2018_392_MOESM1_ESM.docx (52 kb)
Additional file 1: Table S1. Genome-wide association study (GWAS)-identified index variants associated with ocular disease and related traits directly assayed by the Illumina Metabochip. Table S2. The 100 most significant results for the genetic association analysis of the Metabochip and African American primary open-angle glaucoma cases (n = 138) and controls (n = 1376). Table S3. Results (p < 0.0001) for RECESSIVE genetic models for Metabochip-wide tests of association in African American primary open-angle glaucoma cases (n = 138) and controls (n = 1376). Table S4. Results (p < 0.0001) for DOMINANT genetic models for Metabochip-wide tests of association in African American primary open-angle glaucoma cases (n = 138) and controls (n = 1376). (DOCX 52 kb)

References

  1. 1.
    The Eye Diseases Prevalence Research Group. Causes and prevalence of visual impairment among adults in the United States. Arch Ophthalmol. 2004;122(4):477–85.  https://doi.org/10.1001/archopht.122.4.477.CrossRefGoogle Scholar
  2. 2.
    Stein JD, Kim DS, Niziol LM, Talwar N, Nan B, Musch DC, et al. Differences in rates of glaucoma among Asian Americans and other racial groups, and among various asian ethnic groups. Ophthalmology. 2011;118(6):1031–7.  https://doi.org/10.1016/j.ophtha.2010.10.024. PMC3109193CrossRefPubMedPubMedCentralGoogle Scholar
  3. 3.
    Friedman DS, Wolfs RC, O'Colmain BJ, Klein BE, Taylor HR, West S, et al. Prevalence of open-angle glaucoma among adults in the United States. Arch Ophthalmol. 2004;122(4):532–8.  https://doi.org/10.1001/archopht.122.4.532. PMC2798086CrossRefPubMedGoogle Scholar
  4. 4.
    Quigley HA, Broman AT. The number of people with glaucoma worldwide in 2010 and 2020. Br J Ophthalmol. 2006;90(3):262–7.  https://doi.org/10.1136/bjo.2005.081224. PMC1856963CrossRefPubMedPubMedCentralGoogle Scholar
  5. 5.
    Ladapo JA, Kymes SM, Ladapo JA, Nwosu VC, Pasquale LR. Projected clinical outcomes of glaucoma screening in African American individuals. Arch Ophthalmol. 2012;130(3):365–72.  https://doi.org/10.1001/archopthalmol.2011.1224.CrossRefPubMedGoogle Scholar
  6. 6.
    Leske M, Connell AS, Wu S, Hyman LG, Schachat AP. Risk factors for open-angle glaucoma: the Barbados eye study. Arch Ophthalmol. 1995;113(7):918–24.  https://doi.org/10.1001/archopht.1995.01100070092031.CrossRefPubMedGoogle Scholar
  7. 7.
    Pan C-W, Cheung CY, Aung T, Cheung C-M, Zheng Y-F, Wu R-Y, et al. Differential associations of myopia with major age-related eye diseases: the Singapore Indian eye study. Ophthalmology. 2013;120(2):284–91.  https://doi.org/10.1016/j.ophtha.2012.07.065.CrossRefPubMedGoogle Scholar
  8. 8.
    Chandrasekaran S, Cumming RG, Rochtchina E, Mitchell P. Associations between elevated intraocular pressure and Glaucoma, use of Glaucoma medications, and 5-year incident cataract: the Blue Mountains eye study. Ophthalmology. 2006;113(3):417–24.  https://doi.org/10.1016/j.ophtha.2005.10.050.CrossRefPubMedGoogle Scholar
  9. 9.
    Jiang X, Varma R, Wu S, Torres M, Azen SP, Francis BA, et al. Baseline risk factors that predict the development of open-angle Glaucoma in a population: the Los Angeles Latino eye study. Ophthalmology. 2012;119(11):2245–53.  https://doi.org/10.1016/j.ophtha.2012.05.030. PMC3474872CrossRefPubMedPubMedCentralGoogle Scholar
  10. 10.
    Budde WM. Heredity in primary open-angle glaucoma. Curr Opin Ophthalmol. 2000;11(2):101–6.CrossRefPubMedGoogle Scholar
  11. 11.
    Tielsch JM, Katz J, Sommer A, Quigley HA, Javitt JC. Family history and risk of primary open angle glaucoma: the Baltimore eye survey. Arch Ophthalmol. 1994;112(1):69–73.  https://doi.org/10.1001/archopht.1994.01090130079022.CrossRefPubMedGoogle Scholar
  12. 12.
    Freeman EE, Roy-Gagnon M-H, Descovich D, Massé H, Lesk MR. The heritability of glaucoma-related traits corneal hysteresis, central corneal thickness, intraocular pressure, and choroidal blood flow pulsatility. PLoS One. 2013;8(1):e55573.  https://doi.org/10.1371/journal.pone.0055573. PMC3559508CrossRefPubMedPubMedCentralGoogle Scholar
  13. 13.
    van Koolwijk LME, Despriet DDG, van Duijn CM, Pardo Cortes LM, Vingerling JR, Aulchenko YS, et al. Genetic contributions to Glaucoma: heritability of intraocular pressure, retinal nerve Fiber layer thickness, and optic disc morphology. Invest Ophthalmol Vis Sci. 2007;48(8):3669–76.  https://doi.org/10.1167/iovs.06-1519.CrossRefPubMedGoogle Scholar
  14. 14.
    Charlesworth J, Kramer PL, Dyer T, Diego V, Samples JR, Craig JE, et al. The path to open-angle glaucoma gene discovery: endophenotypic status of intraocular pressure, cup-to-disc ratio, and central corneal thickness. Invest Ophthalmol Vis Sci. 2010;51(7):3509–14.  https://doi.org/10.1167/iovs.09-4786. PMC2904007CrossRefPubMedPubMedCentralGoogle Scholar
  15. 15.
    Chang TC, Congdon NG, Wojciechowski R, Muñoz B, Gilbert D, Chen P, et al. Determinants and heritability of intraocular pressure and cup-to-disc ratio in a defined older population. Ophthalmology. 2005;112(7):1186–91.  https://doi.org/10.1016/j.ophtha.2005.03.006. PMC3124001CrossRefPubMedPubMedCentralGoogle Scholar
  16. 16.
    Findl O, Rainer G, Dallinger S, Dorner G, Polak K, Kiss B, et al. Assessment of optic disk blood flow in patients with open-angle glaucoma. Am J Ophthalmol. 2000;130(5):589–96.  https://doi.org/10.1016/S0002-9394(00)00636-X.CrossRefPubMedGoogle Scholar
  17. 17.
    Fontana L, Poinoosawmy D, Bunce CV, O’Brien C, Hitchings RA. Pulsatile ocular blood flow investigation in asymmetric normal tension glaucoma and normal subjects. Br J Ophthalmol. 1998;82(7):731–6.  https://doi.org/10.1136/bjo.82.7.731. PMC1722652CrossRefPubMedPubMedCentralGoogle Scholar
  18. 18.
    Stone EM, Fingert JH, Alward WLM, Nguyen TD, Polansky JR, Sunden SLF, et al. Identification of a gene that causes primary open angle Glaucoma. Science. 1997;275(5300):668–70.  https://doi.org/10.1126/science.275.5300.668.CrossRefPubMedGoogle Scholar
  19. 19.
    Rezaie T, Child A, Hitchings R, Brice G, Miller L, Coca-Prados M, et al. Adult-onset primary open-angle Glaucoma caused by mutations in Optineurin. Science. 2002;295(5557):1077–9.  https://doi.org/10.1126/science.1066901.CrossRefPubMedGoogle Scholar
  20. 20.
    Monemi S, Spaeth G, DaSilva A, Popinchalk S, Ilitchev E, Liebmann J, et al. Identification of a novel adult-onset primary open-angle glaucoma (POAG) gene on 5q22.1. Hum Mol Genet. 2005;14(6):725–33.  https://doi.org/10.1093/hmg/ddi068.CrossRefPubMedGoogle Scholar
  21. 21.
    Adam MF, Belmouden A, Binisti P, Brézin AP, Valtot F, Béchetoille A, et al. Recurrent mutations in a single exon encoding the evolutionarily conserved Olfactomedin-homology domain of TIGR in familial open-angle Glaucoma. Hum Mol Genet. 1997;6(12):2091–7.  https://doi.org/10.1093/hmg/6.12.2091.CrossRefPubMedGoogle Scholar
  22. 22.
    Suzuki Y, Shirato S, Taniguchi F, Ohara K, Nishimaki K, Ohta S. Mutations in the TIGR gene in familial primary open-angle Glaucoma in Japan. Am J Hum Genet. 1997;61(5):1202–4.  https://doi.org/10.1086/301612. PMC1716051CrossRefPubMedPubMedCentralGoogle Scholar
  23. 23.
    Nakano M, Ikeda Y, Tokuda Y, Fuwa M, Omi N, Ueno M, et al. Common variants in CDKN2B-AS1 associated with optic-nerve vulnerability of glaucoma identified by genome-wide association studies in Japanese. PLoS One. 2012;7(3):e33389.  https://doi.org/10.1371/journal.pone.0033389. PMC3299784CrossRefPubMedPubMedCentralGoogle Scholar
  24. 24.
    Osman W, Low SK, Takahashi A, Kubo M, Nakamura Y. A genome-wide association study in the Japanese population confirms 9p21 and 14q23 as susceptibility loci for primary open angle glaucoma. Hum Mol Genet. 2012;21(12):2836–42.  https://doi.org/10.1093/hmg/dds103.CrossRefPubMedGoogle Scholar
  25. 25.
    Wiggs JL, Yaspan BL, Hauser MA, Kang JH, Allingham RR, Olson LM, et al. Common variants at 9p21 and 8q22 are associated with increased susceptibility to optic nerve degeneration in Glaucoma. PLoS Genet. 2012;8(4):e1002654.  https://doi.org/10.1371/journal.pgen.1002654. PMC3342074CrossRefPubMedPubMedCentralGoogle Scholar
  26. 26.
    Thorleifsson G, Walters GB, Hewitt AW, Masson G, Helgason A, DeWan A, et al. Common variants near CAV1 and CAV2 are associated with primary open-angle glaucoma. Nat Genet. 2010;42(10):906–9.  https://doi.org/10.1038/ng.661. PMC3222888CrossRefPubMedPubMedCentralGoogle Scholar
  27. 27.
    Li Z, Allingham RR, Nakano M, Jia L, Chen Y, Ikeda Y, et al. A common variant near TGFBR3 is associated with primary open angle glaucoma. Hum Mol Genet. 2015;24(13):3880–92.  https://doi.org/10.1093/hmg/ddv128. PMC4459396CrossRefPubMedPubMedCentralGoogle Scholar
  28. 28.
    Liu W, Liu Y, Challa P, Herndon LW, Wiggs JL, Girkin CA, et al. Low prevalence of myocilin mutatons in an African American population wiht primary open-angle glaucoma. Mol Vision. 2012;18:2241–6. PMC3429360Google Scholar
  29. 29.
    Matise TC, Ambite JL, Buyske S, Carlson CS, Cole SA, Crawford DC, et al. The next PAGE in understanding complex traits: design for the analysis of population architecture using genetics and epidemiology (PAGE) study. Am J Epidemiol. 2011;174(7):849–59.  https://doi.org/10.1093/aje/kwr160. PMC3176830CrossRefPubMedPubMedCentralGoogle Scholar
  30. 30.
    Buyske S, Wu Y, Carty CL, Cheng I, Assimes TL, Dumitrescu L, et al. Evaluation of the Metabochip genotyping Array in African Americans and implications for fine mapping of GWAS-identified Loci: the PAGE study. PLoS One. 2012;7(4):e35651.  https://doi.org/10.1371/journal.pone.0035651. PMC3335090CrossRefPubMedPubMedCentralGoogle Scholar
  31. 31.
    Roden DM, Pulley JM, Basford MA, Bernard GR, Clayton EW, Balser JR, et al. Development of a large-scale De-identified DNA biobank to enable personalized medicine. Clin Pharmacol Ther. 2008;84(3):362–9.  https://doi.org/10.1038/clpt.2008.89. PMC3763939CrossRefPubMedPubMedCentralGoogle Scholar
  32. 32.
    Pulley J, Clayton E, Bernard GR, Roden DM, Masys DR. Principles of Human Subjects Protections Applied in an Opt-Out, De-identified Biobank. Clin Transl Sci. 2010;3(1):42–8.  https://doi.org/10.1111/j.1752-8062.2010.00175.x. PMC3075971CrossRefPubMedPubMedCentralGoogle Scholar
  33. 33.
    Crawford DC, Goodloe R, Farber-Eger E, Boston J, Pendergrass SA, Haines JL, et al. Leveraging epidemiologic and clinical collections for genomic studies of complex traits. Hum Hered. 2015;79(3–4):137–46.  https://doi.org/10.1159/000381805. PMC4528966CrossRefPubMedPubMedCentralGoogle Scholar
  34. 34.
    Restrepo NA, Farber-Eger E, Goodloe R, Haines JL, Crawford DC. Extracting primary open-angle glaucoma from electronic medical records for genetic association studies. PLoS One. 2015;10(6):e0127817.  https://doi.org/10.1371/journal.pone.0127817. PMC4465698CrossRefPubMedPubMedCentralGoogle Scholar
  35. 35.
    Voight BF, Kang HM, Ding J, Palmer CD, Sidore C, Chines PS, et al. The Metabochip, a custom genotyping array for genetic studies of metabolic, cardiovascular, and anthropometric traits. PLoS Genet. 2012;8(8):e1002793.  https://doi.org/10.1371/journal.pgen.1002793. PMC3410907CrossRefPubMedPubMedCentralGoogle Scholar
  36. 36.
    Crawford DC, Goodloe R, Brown-Gentry K, Wilson S, Robberson J, Gillani NB, et al. Characterization of the Metabochip in diverse populations from the International HapMap Project in the Epidemiologic Architecture for Genes Linked to Environment (EAGLE) Project. Pac Symp Biocomput. 2013;18:188–99. PMC3584704Google Scholar
  37. 37.
    Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D. Principal components analysis corrects for stratification in genome-wide association studies. Nat Genet. 2006;38(8):904–9.CrossRefGoogle Scholar
  38. 38.
    Patterson N, Price AL, Reich D. Population structure and Eigenanalysis. PLoS Genet. 2006;2(12):e190.  https://doi.org/10.1371/journal.pgen.0020190.CrossRefPubMedPubMedCentralGoogle Scholar
  39. 39.
    Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, et al. PLINK: a tool set for whole-genome association and population-based linkage analysis. Am J Hum Genet. 2007;81(3):559–75.CrossRefPubMedPubMedCentralGoogle Scholar
  40. 40.
    Johnson AD, Handsaker RE, Pulit S, Nizzari MM, ODonnell CJ, de Bakker PI. SNAP: a web-based tool for identification and annotation of proxy SNPs using HapMap. Bioinformatics. 2008;24(24):2938–9.  https://doi.org/10.1093/bioinformatics/btn564. PMC2720775CrossRefPubMedPubMedCentralGoogle Scholar
  41. 41.
    Gauderman WJ. Sample size requirements for association studies of gene-gene interaction. Am J Epidemiol. 2002;155(5):478–84.CrossRefPubMedGoogle Scholar
  42. 42.
    Baran Y, Pasaniuc B, Sankararaman S, Torgerson DG, Gignoux C, Eng C, et al. Fast and accurate inference of local ancestry in Latino populations. Bioinformatics. 2012;28(10):1359–67.  https://doi.org/10.1093/bioinformatics/bts144. PMC3348558CrossRefPubMedPubMedCentralGoogle Scholar
  43. 43.
    R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2013.Google Scholar
  44. 44.
    Raj A, Stephens M, Pritchard JK. fastSTRUCTURE: Variational inference of population structure in large SNP data sets. Genetics. 2014;197(2):573–89.  https://doi.org/10.1534/genetics.114.164350. PMC4063916CrossRefPubMedPubMedCentralGoogle Scholar
  45. 45.
    Francis RM. pophelper: an R package and web app to analyse and visualize population structure. Mol Ecol Resour. 2017;17(1):27–32.  https://doi.org/10.1111/1755-0998.12509.CrossRefPubMedGoogle Scholar
  46. 46.
    Parra EJ, Marcini A, Akey J, Martinson J, Batzer MA, Cooper R, et al. Estimating African American admixture proportions by use of population-specific alleles. Am J Hum Genet. 1998;63(6):1839–51.  https://doi.org/10.1086/302148. PMC1377655CrossRefPubMedPubMedCentralGoogle Scholar
  47. 47.
    Baharian S, Barakatt M, Gignoux CR, Shringarpure S, Errington J, Blot WJ, et al. The great migration and African-American genomic diversity. PLoS Genet. 2016;12(5):e1006059.  https://doi.org/10.1371/journal.pgen.1006059. PMC4883799CrossRefPubMedPubMedCentralGoogle Scholar
  48. 48.
    Willer CJ, Sanna S, Jackson AU, Scuteri A, Bonnycastle LL, Clarke R, et al. Newly identified loci that influence lipid concentrations and risk of coronary artery disease. Nat Genet. 2008;40(2):161–9.  https://doi.org/10.1038/ng.76. PMC5206900CrossRefPubMedPubMedCentralGoogle Scholar
  49. 49.
    Kathiresan S, Melander O, Guiducci C, Surti A, Burtt NP, Rieder MJ, et al. Six new loci associated with blood low-density lipoprotein cholesterol, high-density lipoprotein cholesterol or triglycerides in humans. Nat Genet. 2008;40(2):189–97.  https://doi.org/10.1038/ng.75. PMC2682493CrossRefPubMedPubMedCentralGoogle Scholar
  50. 50.
    Beecham GW, Hamilton K, Naj AC, Martin ER, Huentelman M, Myers AJ, et al. Genome-wide association meta-analysis of neuropathologic features of Alzheimer’s disease and related dementias. PLoS Genet. 2014;10(9):e1004606.  https://doi.org/10.1371/journal.pgen.1004606. PMC4154667CrossRefPubMedPubMedCentralGoogle Scholar
  51. 51.
    Pasquale LR, Loomis SJ, Kang JH, Yaspan BL, Abdrabou W, Budenz DL, et al. CDKN2B-AS1 genotype–glaucoma feature correlations in primary open-angle glaucoma patients from the United States. Am J Ophthalmol. 2013;155(2):342–53.e5.  https://doi.org/10.1016/j.ajo.2012.07.023. PMC3544983CrossRefPubMedGoogle Scholar
  52. 52.
    Williams SE, Carmichael TR, Allingham RR, Hauser M, Ramsay M. The genetics of POAG in black South Africans: a candidate gene association study. Sci Rep. 2015;5:8378.  https://doi.org/10.1038/srep08378. PMC4323640CrossRefPubMedPubMedCentralGoogle Scholar
  53. 53.
    Cao D, Jiao X, Liu X, Hennis A, Leske MC, Nemesure B, et al. CDKN2B polymorphism is associated with primary open-angle glaucoma (POAG) in the Afro-Caribbean population of Barbados, West Indies. PLoS One. 2012;7(6):e39278.  https://doi.org/10.1371/journal.pone.0039278. PMC3384655CrossRefPubMedPubMedCentralGoogle Scholar
  54. 54.
    Jandrot-Perrus M, Busfield S, Lagrue A-H, Xiong X, Debili N, Chickering T, et al. Cloning, characterization, and functional studies of human and mouse glycoprotein VI: a platelet-specific collagen receptor from the immunoglobulin superfamily. Blood. 2000;96(5):1798-807.Google Scholar
  55. 55.
    Lonsdale J, Thomas J, Salvatore M, Phillips R, Lo E, Shad S, et al. The Genotype-Tissue Expression (GTEx) project. Nat Genet. 2013;45(6):580-85.Google Scholar
  56. 56.
    Yu NY-L, Hallström BM, Fagerberg L, Ponten F, Kawaji H, Carninci P, Forrest ARR. The FANTOM Consortium, Yoshihide Hayashizaki, Mathias Uhlén, Carsten O. Daub. Complementing tissue characterization by integrating transcriptome profiling from the Human Protein Atlas and from the FANTOM5 consortium. Nucleic Acids Res. 2015;43(14):6787-98.CrossRefPubMedPubMedCentralGoogle Scholar
  57. 57.
    Arjunan P, Gnanaprakasam JP, Ananth S, Romej MA, Rajalakshmi V-K, Prasad PD, Martin PM, Gurusamy M, Thangaraju M, Bhutia YD, Ganapathy V. Increased Retinal Expression of the Pro-Angiogenic Receptor GPR91 via BMP6 in a Mouse Model of Juvenile Hemochromatosis. Investigative Opthalmology & Visual Science. 2016;57(4):1612.CrossRefGoogle Scholar
  58. 58.
    Popejoy AB, Fullerton SM. Genomics is failing on diversity. Nature 2016;538(7624):161-164CrossRefPubMedPubMedCentralGoogle Scholar

Copyright information

© The Author(s). 2018

Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Authors and Affiliations

  • Nicole A. Restrepo
    • 1
  • Sarah M. Laper
    • 2
  • Eric Farber-Eger
    • 3
  • Dana C. Crawford
    • 1
  1. 1.Department of Population and Quantitative Health Sciences, Institute for Computational BiologyCase Western Reserve UniversityClevelandUSA
  2. 2.Eastern Virginia Medical SchoolNorfolkUSA
  3. 3.Vanderbilt Institute for Clinical and Translational ResearchVanderbilt University Medical CenterNashvilleUSA

Personalised recommendations