Evaluating Practical Approaches for Including MYOC Variants Alongside Common Variants for Genetics-Based Risk Stratification for Glaucoma

Abstract

Objective

Rare variants in the MYOC gene are associated with glaucoma risk, with p.Gln368Ter the most common pathogenic variant in Europeans. Genetics-based risk stratification may aid with early diagnosis for glaucoma but it is unclear how best to combine the p.Gln368Ter status with polygenic risk scores (PRS). Our study aimed to examine approaches for identifying p. Gln368Ter carriers using genotyping array data and the utility of integrating p.Gln368Ter status into glaucoma PRS.

Design

Retrospective cohort study.

Methods

We identified p.Gln368Ter carriers using directly genotyped and imputed data. Results were confirmed in a subset with sequencing data. We evaluated the combined effects of p.Gln368Ter status and PRS in stratified analyses by considering them as two separate factors and as an aggregate score.

Participants

A total of 58,452 participants from the Genetics of Glaucoma, the QSkin Sun and Health Study (QSKIN), and CARTaGENE projects, including 6015 with sequencing data.

Main Outcomes and Measures

The concordance of direct genotyping, compared with imputation and sequencing for p.Gln368Ter identification.

Results

Without appropriate quality control, substantial mis-calling may occur. Nevertheless, the p.Gln368Ter variant could be accurately genotyped in most cases by filtering individuals for call rate and heterozygosity. In 6015 individuals with sequencing data, direct genotyping exhibited perfect concordance with sequencing results. Filtered direct genotyping results showed high agreement with imputed results, with only 16 discrepancies among 57,468 individuals. When quality control is not possible (eg, heterozygosity filtering for an individual), we recommend comparing genotyped and imputed results to ensure accuracy. Incorporating p.Gln368Ter into PRS had additional effects on stratifying high–risk individuals, but did not improve risk prediction for the general population given the variant's rarity. The MYOC-enhanced PRS increased the proportion of p.Gln368Ter carriers classified as high risk from 32.31% to 75.38% in QSKIN and from 38.24% to 79.41% in CARTaGENE.

Conclusions

The p.Gln368Ter variant can be genotyped with high accuracy using array data, provided careful quality control measures are implemented. Incorporating p.Gln368Ter into glaucoma PRS improved risk stratification for carriers.