Development of a Genotype Assay for Age-Related Macular Degeneration: The EYE-RISK Consortium

84 Purpose: To develop a genotype assay to assess associations with common and rare AMD risk variants, to 85 calculate an overall genetic risk score (GRS), and to identify potential misdiagnoses with inherited macular 86 dystrophies that mimic AMD. 87 Design: Case-control study. 88 Participants: Individuals (N=4,740) from five European cohorts. 89 Methods: We designed single molecule molecular inversion probes (smMIPs) for target selection and used next 90 generation sequencing (NGS) to sequence eighty-seven single nucleotide polymorphisms (SNPs), coding and 91 splice-site regions of ten AMD-(related) genes (ARMS2, C3, C9, CD46, CFB, CFH, CFI, HTRA1, TIMP3, SLC16A8), 92 and three genes that cause inherited macular dystrophies (ABCA4, CTNNA1, PRPH2). GRS for common AMD risk 93 variants were calculated based on effect size and genotype of 52 AMD-associated variants. Frequency of rare 94 variants was compared between late AMD cases and control individuals with logistic regression analysis. 95 Main Outcome Measures: GRS, association of genetic variants with AMD, genotype-phenotype correlations. 96 Results: We observed high concordance rates between our platform and other genotyping platforms for the 69 97 successfully genotyped SNPs (96.77-97.28%) and for the rare variants (99.81%). We observed a higher GRS for 98 patients with late AMD compared to patients with early/intermediate AMD (p<0.001) and individuals without 99 AMD (p<0.001). A higher proportion of rare loss-of-function variants and variants with a Combined Annotation 100 Dependent Depletion score ≥20 in the CFH (50 [2.92%] vs 8 [1.02%], OR=2.88 [1.36-6.11], p=0.006), CFI (38 101 [2.22%] vs 4 [0.51%], OR=4.45 [1.58-12.50], p=0.005) and C3 (56 [3.27%] vs 4 [0.51%], OR=6.56 [2.37-18.17], 102 p=0.0003) genes was observed in late AMD cases compared to control individuals. In nine patients we 103 identified pathogenic variants in the PRPH2, ABCA4 and CTNNA1 genes, which allowed reclassification of these 104 patients as inherited macular dystrophy. 105 Conclusions: This study reports a high-throughput and comprehensive genotype assay for common and rare 106 AMD genetic variants. This test can identify individuals at intermediate to high genetic risk of late AMD, and 107 enables differential diagnosis of AMD mimicking dystrophies. Our study supports sequencing of CFH, CFI and C3 108 genes as they harbor rare high-risk loss-of-function variants. Carriers of these variants could be amendable for 109 new treatments for AMD that are currently under development. 110

(e.g. ARHGAP21, B3GALTL) have been associated with AMD. [4][5][6][7][8][9] The largest genome-wide association study 122 (GWAS) in AMD was published in 2016 and identified 52 independently associated genetic variants with AMD 123 distributed across 34 loci. 7 The majority of these variants were common genetic variants, while seven variants 124 were rare (minor allele frequency < 0.01) in the investigated population. Furthermore, a significantly higher 125 burden of rare variants in the CFH, CFI, TIMP3 and SLC16A8 genes was identified in AMD patients compared to 126 control individuals. In recent years, the role of rare genetic variants in AMD gained attention, as they can have 127 large effect sizes. Sequencing of candidate genes in case-control studies and in AMD families resulted in the 128 identification of rare variants in the CFH, CFI, C3 and C9 genes that could be linked to AMD.

246
Performance of the genotype assay

247
Out of the 87 SNPs, 69 SNPs were genotyped successfully, while 11 SNPs were excluded due to low 248 coverage ( Figure S1 and

288
In Figure 2 we demonstrated how the GRS can be used to report the AMD risk to individuals, using a 289 small family as an example.  Figure 3 shows  unclear. 28 We identified one individual carrying this particular variant. The overall GRS of this individual was 372 1.39. Although the phenotype of this individual did not match with a butterfly-shaped pigment dystrophy, we 373 did observe an egg-yolk lesion in one eye, which is also observed in patients with Best vitelliform macular 374 dystrophy ( Figure 3I and

440
In the GWAS of the IAMDGC study the authors identified a burden of rare variants for the CFH, CFI, 441 SLC16A8 and TIMP3 genes. 7 In our study we did not observe a higher occurrence of rare variants in the 442 SLC16A8 and TIMP3 genes. This could potentially be attributed to the smaller sample size compared to the 443 GWAS of the IAMDGC study. Furthermore, two exons of the SCL16A8 gene showed a lower coverage on our 444 genotype platform, therefore, we potentially could have missed rare variants in these regions.

447
The ABCA4, CTNNA1 and PRPH2 genes were included in this study to identify potential misdiagnoses.

448
Genotype and phenotype data of our study revealed nine potential misdiagnoses of inherited macular

499
When one or more rare variants are identified in a patient we believe it is important to take into 500 account the functional effect of the rare variant. For some variants the functional effect has been tested 501 previously and it has been reported that some rare variants confer high risk of AMD, whereas other rare 502 variants do not influence the protein or are even protective for AMD. 19 For the majority of the rare variants the 503 functional effect is currently unknown. When rare variants in the CFH or CFI genes are identified we would 504 recommend to perform an ELISA assay to determine FH levels or FI levels, respectively. Not all rare variants 505 cause lower protein levels. Some rare variants present with normal protein levels, whereas the functionality 506 has been reduced. 44 In these cases functional assays such as a C3b degradation assay can be performed ( Figure   507 4     Précis 1 This study reports a genetic test for age-related macular degeneration, which can identify individuals at high 2 risk for late age-related macular degeneration, carriers of rare high-risk variants, and potential misdiagnoses 3 with inherited macular dystrophies.