Links from this page to related variants, conditions, drugs and topics are not included for automated readers. Those connections are the Gene Knowledge Graph and are licensed as data: https://www.mygenelog.com/license

Ophthalmic

Myopia

Reviewed September 11, 2026

Nearsightedness, already affecting an estimated 23% of the world and projected to reach half of it by 2050. Common-variant genetics explains only a fraction of the risk; a splicing mechanism in the X-linked cone opsin genes explains much more in some people, and the same genes are why red-green colour blindness exists.

What this condition connects to

Prevalence
A 2016 meta-analysis of 145 studies and 2.1 million people estimated 22.9% global myopia prevalence in 2000 (1,406 million people) and 2.7% high myopia (163 million), projected to reach 49.8% and 9.8% respectively by 2050 (PMID:26875007). Red-green colour vision deficiency, from the same gene family, affects up to 8% of men and 0.5% of women of European descent (PMID:19927164).
Inheritance
Myopia itself is polygenic — many common variants of small effect, such as MIPEP and SNTB1, plus a splicing haplotype in OPN1LW with a much larger measured effect on refraction in one 2022 study. Red-green colour vision deficiency, from the same X-linked gene family, follows X-linked recessive inheritance.

Myopia (nearsightedness) is already one of the most common conditions covered on this site. A 2016 systematic review and meta-analysis of 145 studies covering 2.1 million people estimated 22.9% of the world's population was myopic in 2000 (1,406 million people), with 2.7% having high myopia (163 million). The same analysis projected those figures would reach 49.8% and 9.8% by 2050 — 4,758 million and 938 million people.

Two ordinary GWAS hits, and one much stranger mechanism

Genome-wide association studies have found dozens of common variants linked to myopia risk, each with a small individual effect — rs9318086 near MIPEP and rs6469937 near SNTB1 are two of them. Identified markers collectively explain only a modest share of the heritability, and none of them changes how myopia is diagnosed or corrected.

Tedja et al. 2018, the CREAM consortium, ran a much larger meta-analysis: 160,420 people at discovery, plus 95,505 more at replication, raising the number of established independent refractive-error signals from 37 to 161. Genetic correlation between the European and Asian discovery samples was high (over 0.78), meaning largely the same loci matter across ancestries. The paper's central finding was mechanistic rather than just a longer gene list: the associated genes point to a light-dependent signaling cascade running from the retina to the sclera, alongside the environmental "near work" story long assumed to drive myopia. Novel biology included genes for synaptic signaling between rod and cone photoreceptors and bipolar cells, anterior-segment eye morphology, and blood-vessel formation. This page's rs670352, near GJD2 — a longstanding, well-replicated myopia gene coding a retinal gap-junction protein — is one of the paper's own loci. Twelve of this page's variants come from this study.

Three earlier studies established the ground this later work built on. Solouki et al. 2010, in 5,328 Dutch people replicated in 10,280 more, found rs634990 at 15q14, near GJD2 — the odds of myopia versus hyperopia were 1.41 times higher for people carrying one copy of the risk allele and 1.83 times higher for two copies. In the same year, Hysi et al. 2010, in 4,270 TwinsUK participants replicated in 13,414 more across six cohorts, found a second locus at 15q25, near RASGRF1 — this page's rs939658 is that paper's own finding. Verhoeven et al. 2013 then combined 37,382 Europeans with 8,376 Asians and found 24 loci in total, including PRSS56, a gene tied to eye development — this page's rs1656404 is in that gene — while also confirming GJD2 and RASGRF1. A genetic risk score built from the associated variants showed a tenfold higher myopia risk for people in the highest-risk group.

A different, more mechanistically striking story comes from the OPN1LW and OPN1MW genes — the long- and middle-wavelength cone opsins, sitting in a tandem array at Xq28. A 2022 study of 413 men of European ancestry found that specific combinations of eight coding positions in OPN1LW's exon 3 — a haplotype, not a single variant — determine how often that exon gets skipped during splicing. Cones expressing an exon-3-skipped transcript are left nearly empty of photopigment. Mean refraction across the eleven haplotypes tested ranged from +0.51 diopters (haplotype MVAIS, the least myopic) to -3.14 diopters (haplotype MVVVA, the most myopic) — a roughly 3.6-diopter spread from haplotype alone. rs145009674, one of the eight defining positions, is included on this site as a way into that finding, not as a variant that alone explains it.

The hypothesis this generates is genuinely testable: cones next to each other with very different amounts of photopigment would send a distorted contrast signal, and that distortion — not blur itself — might be what drives the eye's axial elongation in myopia. The same research group is testing contrast-reducing spectacle lenses in a multi-center randomized trial of 256 participants; a 12-month interim analysis favored the contrast-reducing lenses over standard correction.

Where near-work habits enter the genetics

Fan et al. 2016 (CREAM Consortium) analysed 50,351 people — 40,036 of European ancestry and 10,315 of Asian ancestry, across 34 studies — using a model that jointly tests an ordinary genetic effect on refractive error alongside how that effect might change with a person's education level, used here as a proxy for a lifetime of near-work: reading and close-up focus are thought to induce a lag in visual accommodation that promotes the eye elongation behind myopia.

The joint test found nine new loci for refractive error — six in the European analysis, three in the Asian analysis. Only the three Asian loci showed a real, statistically significant interaction with education specifically; the study states plainly that the six European loci reached significance through the same joint statistical test but did not actually show an education interaction on closer inspection — their effect on refractive error was constant regardless of educational history. Where an interaction was real, it was substantial: about four times larger on average in the Asian cohort than the European one. rs12511037, near AREG, is one of this page's ten variants from that Asian, education-interacting set.

The other nine variants on this page come from the same combined analysis, spanning both its newly discovered loci and refractive-error signals it reconfirmed from earlier research — not every one individually tested for or shown to carry an education interaction the way AREG was.

Why colour vision deficiency does not get its own page

OPN1LW and OPN1MW are also the genes behind red-green colour vision deficiency, affecting up to 8% of men and 0.5% of women of European descent — the X-linked recessive pattern is why men, with only one X chromosome, are affected so much more often. But the classic cause of colour blindness is a different kind of change entirely: whole-gene deletions or hybrid genes formed when the two near-identical genes recombine unequally with each other. That is a structural rearrangement, not a single-nucleotide variant, and it has no entry in the GWAS Catalog at all. This site's variant pages are built around single positions with a stable rsID, so colour vision deficiency is described here in prose, on the strength of the same gene family covered above, rather than forced onto a variant that does not represent it.

In the news

2026-04-21 · Genome-wide association study of myopia progression in Chinese adolescents and application of polygenic risk score prediction. Frontiers in Chemistry. 2026. DOI:10.3389/fchem.2026.1778732

Seven suggestive loci track how fast myopia worsens in Chinese teenagers, not just whether they get it

Most myopia GWAS ask who develops myopia at all; this study asked a narrower question -- among adolescents who already have it, what predicts how fast it gets worse. Following 294 Chinese adolescents for two years in a two-stage discovery (n=176) and replication (n=118) design, the study identified seven suggestive loci associated with myopia progression: FSTL5, SMARCA2, CCDC3, GALNT6/ACVR1B, CRY1, ULK2 and MYL4/EFCAB13-DT. These reached only a suggestive statistical threshold in this small cohort, not full genome-wide significance, so they should be read as an early, unreplicated-at-scale signal. The more immediately useful result may be practical: building a polygenic risk score from this East Asian discovery cohort predicted progression risk better in East Asian adolescents (R2=5.69%, OR=1.61, AUC=0.66) than a score built from European UK Biobank data (N=460,536) or a cross-ancestry blend -- a concrete demonstration that genetic risk tools trained on the 'wrong' ancestry underperform, even when the training dataset is far larger. This site's myopia page carries 117 variants; none of these seven progression-specific genes are currently among them, and given the small cohort and suggestive (not genome-wide-significant) threshold, this entry documents the finding as an early signal rather than an established one.

2026-04-20 · Multi-ancestry genome-wide association analyses of refractive error augment genetic discovery and polygenic prediction. Nature Genetics. 2026. DOI:10.1038/s41588-026-02576-0

Nearly 1 million refractive-error variants later, a polygenic score can now flag high myopia with 0.81 accuracy

Refractive error (myopia, hyperopia and astigmatism) affects over half the global population. This study ran the largest multi-ancestry GWAS of refractive error to date -- 1.76 million people combined across European (1,495,159), East Asian (121,172) and African (144,737) ancestries. The cross-ancestry meta-analysis found 932 associated variants, 241 of them newly reported, plus 4 East Asian-specific and 1 African-specific association -- genuine evidence that ancestry-specific analysis finds real signal a European-only study would miss. Fine-mapping identified 16 high-confidence causal variants, and gene prioritization highlighted 23 genes with roles in eye development. The study's most immediately useful contribution was building an enhanced polygenic predictor incorporating functional annotations, explaining 21.4% of refractive error variation and achieving an AUC of 0.806 for identifying high myopia specifically -- a level of predictive accuracy with real potential for identifying children at highest risk of severe, vision-threatening myopia before it progresses. This site's myopia page carries 117 variants; this entry documents the much larger, more recent multi-ancestry catalog underlying current polygenic prediction rather than listing all 241 newly reported variants individually.

2026-03-01 · Journal of Translational Medicine 2026, PMID:41888893

From a GWAS hit to a working gene therapy: boosting CPNE1 slowed myopia progression in animal eyes

Verified directly against the paper's own abstract via Europe PMC (PMID:41888893, published March 2026). Researchers meta-analyzed two large myopia GWAS (41,468 cases, 817,202 controls), found 35 genome-wide significant loci, and used integrative analysis to prioritize CPNE1 as a candidate gene -- Mendelian randomization supported a protective causal role, and CPNE1 was found downregulated in myopic sclera tissue. The study then went from genetic finding to working intervention: in an animal model, AAV-mediated CPNE1 overexpression (a gene therapy approach) attenuated axial elongation and refractive error progression -- the two physical measures that define worsening myopia. Mechanistically, CPNE1 acted through TGF-beta1 and collagen pathways affecting scleral fibrotic remodeling. This is a different locus from any of this page's own linked variants, but demonstrates a real GWAS-to-therapy pipeline for the same disease this page covers.

2026-04-20 · Nature Genetics

Largest myopia genetics study yet spans three ancestries, finds 241 new risk variants

A genome-wide study of refractive error spanning 1,495,159 people of European, 121,172 of East Asian, and 144,737 of African ancestry found 932 associated variants in total, 241 of them new (plus a few ancestry-specific signals). Fine-mapping narrowed this to 16 high-confidence causal variants across 23 genes tied to eye development. A polygenic score built from the results explained 21.4% of refractive-error variation and predicted high myopia with an AUC of 0.806 -- strong enough to flag which children's myopia is likely to progress furthest, not just who has it. PMID:42009823.

2026-04-20 · Cheng FF, et al., Nature Genetics 2026, PMID:42009823

Multi-ancestry genome-wide association analyses of refractive error augment genetic discovery and polygenic prediction

Refractive errors affect over half of the world's population. This meta-analysis covered people of European (1,495,159), East Asian (121,172) and African (144,737) ancestry and identified 932 associated variants, 241 of them previously unknown, four specific to East Asian ancestry and one to African ancestry. Fine-mapping pinpointed 16 high-confidence putative causal variants, and 23 prioritised genes are involved in eye development. A polygenic predictor built from the results explained 21.4% of the variation in refractive error and reached an area under the curve of 0.806 for predicting high myopia — a research result, not a test in clinical use. 245 of the study's positions are already on this site, 219 of them on the myopia page, including rs634990 (near GJD2), rs11602008 (LRRC4C) and rs10500355 (RBFOX1), which this site's cataract page also holds. PMID:42009823.

Clinical detail

What is actually diagnosed and treated here

Myopia is diagnosed with a standard eye exam and refraction, not by genotype. It is corrected with glasses, contact lenses, or refractive surgery; a range of interventions (including specialty contact lenses and low-dose atropine) are used specifically to slow its progression in children. None of that depends on any variant on this page.

Colour vision deficiency is diagnosed with standardised colour plates or anomaloscopy, and its management is chiefly practical: appropriate counselling, including career counselling, since some occupations (aviation, some electrical work, some military and public-safety roles) have colour-vision requirements. Visual aids exist but do not restore normal colour discrimination.

The OPN1LW exon-3 haplotype work is a mechanism finding, still being tested in a clinical trial — it has not translated into a diagnostic test or a treatment recommendation outside that trial.

Which ancestries this evidence comes from

The studies behind these variants recruited participants from different ancestries — a result found in one population doesn't always transfer to another. Based on 126 of 390 linked studies with a resolved discovery ancestry.

European · 0.5% East Asian · 31.8% Not yet resolved · 67.7%

Sources

Databases, guidelines and references

Papers, with their authors

Quoting this page

Licensed CC BY-NC 4.0: free to read, share and adapt for non-commercial use, naming MyGeneLog™ and linking the exact page. Commercial use, including training or serving a commercial AI system, needs a licence — https://www.mygenelog.com/license

Myopia. MyGeneLog™. https://www.mygenelog.com/conditions/myopia

Questions about Myopia

Can these variants diagnose myopia or predict who will get it?

No. Myopia is diagnosed with a standard eye exam and refraction. These variants come from genetic association studies and are not used diagnostically.

Why does this page cover colour vision deficiency if there is no variant for it?

Colour vision deficiency is caused by whole-gene deletions and hybrid genes in the same OPN1LW/OPN1MW gene family this page's myopia variant comes from — a structural rearrangement rather than a single-nucleotide variant, and it has no GWAS Catalog entry at all. It is described here in prose rather than forced onto a variant that does not represent it.

Does the OPN1LW finding mean myopia can be prevented with special glasses?

Not yet established. A multi-center randomized trial of contrast-reducing lenses is under way and an interim 12-month analysis favored them over standard correction, but this has not translated into a general recommendation.

How common is myopia expected to become?

A 2016 meta-analysis projected global prevalence would rise from 22.9% in 2000 to 49.8% by 2050 — from 1,406 million to 4,758 million people.

Free to reuse, non-commercially. This page's text is original writing from freely-available research, licensed CC BY-NC 4.0 — share and adapt it with attribution to MyGeneLog™; commercial use is licensed. It's general research-derived information, not medical advice or a diagnosis — see Terms of Use.