This variant position (rs1048661) was found to be strongly associated with exfoliation glaucoma (XFG) in icelandic and swedish individuals. The variant conferring more risk was the “G” allele — which is the reference variant (141R) — and so 141L would be interpreted as having a protective effect. In their study, they found the G allele had a frequency of 0.65 in controls and 0.83 cases of XFG.
They also report this genotype is associated with significant difference is gene expression. However, the method did not independently introduce this variant to constructs to measure its effect on expression alone; rather, they looked at expression data from individuals who happen to carry the genotype. This means the expression differences could be caused by other nearby linked genetic variants, rather than rs1048661 itself.
Because they report no significant deviation from Hardy-Weinberg equilibrium in cases and controls, their allele frequencies imply genotype frequencies in cases of GG=.69 / GT=.28 / TT=0.3 and in controls of GG=0.42 / GT=0.45 / TT = 0.12. Controls were not chosen as specifically excluding the disease.
To estimate how carrying this variant affects lifetime risk for XFG in these populations, we would like to compare average risk vs. risk dependent on genotype. Assuming a lifetime risk for XFG of 15%, we would estimate the non-XFG group to have genotype frequencies of GG=0.38 / GT=0.48 / TT=0.14. (i.e. for GG, 0.42 = 0.15 * 0.69 + 0.85 * 0.38, etc.) Using this, GG’s lifetime risk for disease is 0.24 = (.69*.15 / (.69 * .15 + .38 * .85)), GT’s risk = 0.09, TT’s risk = .04. From this perspective, carrying 141L homozygously is associated with being protective having 11% decreased attributable risk and GT protective with 6% decreased attributable risk (while homozygosity for the reference GG genotype is pathogenic with 9% increased attributable risk).