Polygenic risk score prediction of multiple sclerosis in individuals of South Asian ancestry

Abstract

Polygenic risk scores aggregate an individual's burden of risk alleles to estimate the overall genetic risk for a specific trait or disease. Polygenic risk scores derived from genome-wide association studies of European populations perform poorly for other ancestral groups. Given the potential for future clinical utility, underperformance of polygenic risk scores in South Asian populations has the potential to reinforce health inequalities. To determine whether European-derived polygenic risk scores underperform at multiple sclerosis prediction in a South Asian-ancestry population compared with a European-ancestry cohort, we used data from two longitudinal genetic cohort studies: Genes & Health (2015-present), a study of ∼50 000 British-Bangladeshi and British-Pakistani individuals, and UK Biobank (2006-present), which is comprised of ∼500 000 predominantly White British individuals. We compared individuals with and without multiple sclerosis in both studies (Genes & Health: N _Cases = 42, N _Control = 40 490; UK Biobank: N _Cases = 2091, N _Control = 374 866). Polygenic risk scores were calculated using clumping and thresholding with risk allele effect sizes obtained from the largest multiple sclerosis genome-wide association study to date. Scores were calculated with and without the major histocompatibility complex region, the most influential locus in determining multiple sclerosis risk. Polygenic risk score prediction was evaluated using Nagelkerke's pseudo-R ² metric adjusted for case ascertainment, age, sex and the first four genetic principal components. We found that, as expected, European-derived polygenic risk scores perform poorly in the Genes & Health cohort, explaining 1.1% (including the major histocompatibility complex) and 1.5% (excluding the major histocompatibility complex) of disease risk. In contrast, multiple sclerosis polygenic risk scores explained 4.8% (including the major histocompatibility complex) and 2.8% (excluding the major histocompatibility complex) of disease risk in European-ancestry UK Biobank participants. These findings suggest that polygenic risk score prediction of multiple sclerosis based on European genome-wide association study results is less accurate in a South Asian population. Genetic studies of ancestrally diverse populations are required to ensure that polygenic risk scores can be useful across ancestries.

Keywords: ethnicity; genetics; multiple sclerosis.

Graphical abstract

Figure 1

Flow diagram of individual quality control in G&H. From an initial 44 396 individuals with genotype data, 40 532 individuals were retained for analysis comprising 42 multiple sclerosis cases and 40 490 controls

Figure 2

Multiple sclerosis PRS performance in the G&H cohort of South Asian-ancestry individuals. (A and B) Density plots showing the distribution of PRS for PRS with (A) and without (B) the MHC locus in multiple sclerosis cases and controls. (C) Odds ratio quartile plots for individual PRS scores. ORs were calculated relative to the lower quartile. (D) Receiver operating characteristic (ROC) curves for the MHC PRS model, non-MHC model and the null model, with corresponding AUC scores. (E and F) Calibration plots showing the absolute multiple sclerosis disease probabilities (prevalence) for each PRS quartile versus mean fitted probabilities within each quartile from the PRS models. Plots shown for MHC PRS model, non-MHC PRS model, null model and the observed multiple sclerosis risk in each quartile. Odds ratios and AUC values are derived from multivariable logistic regression models

Figure 3

Estimates of PRS performance in EUR UKB participants and SAS G&H participants. Each point represents the estimated liability explained by the optimal PRS, with 95% confidence intervals for the sub-samples of UKB. The vertical lines indicate the performance of each score in G&H. PRS containing the MHC are coloured in purple, and those without coloured in orange. Estimates reflect Nagelkerke’s pseudo-R² statistic adjusted for disease prevalence, which is derived from multivariable logistic regression models. ‘UKB all’ refers to scores calculated in all EUR-ancestry UKB participants. To control for effects of sample size, we resampled subsets of UKB to have equivalent case and control numbers to G&H (42 cases, 40 490 controls). ‘UKB subset’ refers to estimates derived from 1000 replicates of this random sampling procedure, with empirical 95% confidence intervals