Figure 3.
Association studies of polygenic scores with %HbF and acute VOP event rate in the SCCRIP cohort. (A) The number of participants (N) experiencing 0, 1, 2, or >2 VOP events according to age. There are in total 1897 data points. (B) The proportion of participants treated with hydroxyurea (HU) therapy by age. (C) Fetal hemoglobin levels (%HbF) of participants according to PGSHbF and age in years. The PGS score on the x axis shows the number of unfavorable (low HbF–associated) alleles carried by each individual, with 0, 1, or 2 alleles being possible for each SNP. Accordingly, for the 11-SNP PGSHbF, the potential PGS for an individual can range from 0 to 22. The lines denote the smoothed mean values with the lower and upper shaded areas indicating the 95% confidence intervals. P values were calculated by a Wald test to test the association between %HbF and PGSHbF (estimate, −0.15; standard error, 0.09), using generalized estimating equations that accounted for the correlation between %HbF measurements on the same participants after adjusting for sex, 5 PCs, −α3.7 deletion, exposure to HU, chronic transfusion therapy, and nonlinear (quadratic and cubic) terms of age. (D) Association of VOP event rate with the polygenic scores defined in Table 3. The scatterplots show the estimated VOP event rate (y axis) according to each PGS model. Each symbol represents an estimated value of the VOP event rate, with the fitted linear regression line being shown for each PGS. P values were calculated based on a GLMM to test for associations between VOP event rate and the indicated PGS, which was analyzed as a continuous variable, adjusting for the same covariates listed in Table 3. Est, β estimate; SE, standard error. (E) The distribution of participants with respect to the 21-SNP PGSHbF+COMT+5snps. (F) Boxplots of the estimated VOP event rate (y axis) according to the 3 strata of PGSHbF+COMT+5snps: bottom 5% (<19, n = 16), middle group (19-30, n = 290), and top 5% (≥30, n = 19). P values were calculated as described for Table 3. (G) The distribution of exposure to HU therapy (y axis) with respect to the 3 strata of PGSHbF+COMT+5snps. P values were calculated based on the GLMM with a binomial link function, with exposure to HU therapy as a dependent variable and the 3 strata as independent variables, adjusted for the same covariates described in Table 3 except for exposure to HU therapy.

Association studies of polygenic scores with %HbF and acute VOP event rate in the SCCRIP cohort. (A) The number of participants (N) experiencing 0, 1, 2, or >2 VOP events according to age. There are in total 1897 data points. (B) The proportion of participants treated with hydroxyurea (HU) therapy by age. (C) Fetal hemoglobin levels (%HbF) of participants according to PGSHbF and age in years. The PGS score on the x axis shows the number of unfavorable (low HbF–associated) alleles carried by each individual, with 0, 1, or 2 alleles being possible for each SNP. Accordingly, for the 11-SNP PGSHbF, the potential PGS for an individual can range from 0 to 22. The lines denote the smoothed mean values with the lower and upper shaded areas indicating the 95% confidence intervals. P values were calculated by a Wald test to test the association between %HbF and PGSHbF (estimate, −0.15; standard error, 0.09), using generalized estimating equations that accounted for the correlation between %HbF measurements on the same participants after adjusting for sex, 5 PCs, −α3.7 deletion, exposure to HU, chronic transfusion therapy, and nonlinear (quadratic and cubic) terms of age. (D) Association of VOP event rate with the polygenic scores defined in Table 3. The scatterplots show the estimated VOP event rate (y axis) according to each PGS model. Each symbol represents an estimated value of the VOP event rate, with the fitted linear regression line being shown for each PGS. P values were calculated based on a GLMM to test for associations between VOP event rate and the indicated PGS, which was analyzed as a continuous variable, adjusting for the same covariates listed in Table 3. Est, β estimate; SE, standard error. (E) The distribution of participants with respect to the 21-SNP PGSHbF+COMT+5snps. (F) Boxplots of the estimated VOP event rate (y axis) according to the 3 strata of PGSHbF+COMT+5snps: bottom 5% (<19, n = 16), middle group (19-30, n = 290), and top 5% (≥30, n = 19). P values were calculated as described for Table 3. (G) The distribution of exposure to HU therapy (y axis) with respect to the 3 strata of PGSHbF+COMT+5snps. P values were calculated based on the GLMM with a binomial link function, with exposure to HU therapy as a dependent variable and the 3 strata as independent variables, adjusted for the same covariates described in Table 3 except for exposure to HU therapy.

Close Modal
This Feature Is Available To Subscribers Only

or Create an Account

Close Modal
Close Modal