Trans-ancestry polygenic models for the prediction of LDL blood levels: An analysis of the UK Biobank and Taiwan Biobank

Background: Polygenic risk scores (PRSs) are proposed for us e in clinical and research settings for risk stratification. PRS predictions often show bi as toward the population of available genome- wide association studies, which is typically of European ancestry. This study aims to assess the performance differences of ancestry-specif ic PRS and test the implementation of multi- ancestry PRS to enhance the generalizability of low-density lipoprotein (LDL) cholesterol predictions in the East Asian population

We computed ancestry-specific and multi-ancestry PRS for LDL using data from the global lipid consortium while accounting for population-specific linkage disequilibrium patterns using PRS-CSx method. We first conducted an anc estry-wide analysis using the UK Biobank dataset (n=423,596) and then applied the same models to the Taiwan Biobank dataset (TWB, n=68,978). PRS performances were based on linear regression with adjustment for age, sex, and principal components. PRS strata were considered to assess the extent to which a PRS categorization can stratify individual s for LDL cholesterol levels in East Asian samples.

Population-specific PRS better predicted LDL levels within the target population but multi- ancestry PRS were more generalizable. In the TWB dataset, covariate-adjusted R 2 values were 9.3% for ancestry-specific PRS, 6.7% for multi-ancestry PRS, and 4.5% for European- specific PRS. Similar trends (8.6%, 7.8%, 6.2%) were observed in the smaller East Asian population of the UK Biobank (n=1,480). Consistent with the R 2 values, PRS stratification in East Asians (TWB) effectively captured a heterogenous variability in LDL blood cholesterol levels across PRS strata. The mean difference in LDL levels between the lowest and highest East Asian-specific PRS (EAS_PRS) deciles was 0.82, compared to 0.59 for European- specific PRS (EUR_PRS) and 0.76 for multi-ancestry PRS. Notably, the mean LDL values in the top decile of multi-ancestry PRS were comparable to those of EAS_PRS (3.543 vs. 3.541, P=0.86).

Our analysis of the PRS prediction model for LD L cholesterol further supports the issue of PRS generalizability across populations. Our targeted analysis of the East Asian (EAS) population revealed that integrating non-European genotyping data, accounting for population-specific linkage disequilibrium, and considering meta-analyses of non-European- . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 6, 2023. ; https://doi.org/10.1101/2023.08.03.23293320doi: medRxiv preprint based GWAS alongside powerful European-based GWAS can enhance the generalizability of LDL PRS.

Blood lipid levels are a significant, modifiable, and heritable risk factors for coronary artery disease (CAD), including low-density lipoprotein (LDL-C) 1. Previous studies have demonstrated that lipid levels have a moderate-to-high heritability variations, ranging from 20 to 60% 2. Numerous common variants have been discovered in recent genome-wide association studies (GWASs) associated with LDL, as well as many other traits 3. However, the majority of these variants are weakly asso ciated individually with a given trait or disease and have limited predictive power. The cumulative effects of several common variants have been suggested to contribute significantly to the ri sk stratification for clinical utility. Methods have been developed for analyzing data from th ese large-scale studies and detecting genetic variants and phenotype associations, i.e., one such method is the polygenic risk score (PRS). Several studies have evaluated the association between PRS and the risk of various conditions 4, including lipid traits 5, CAD 6, cancer 7,8, diabetes 9, and neurodevelopmental disorders. One of the major issues concerning the tr anslational use of PRS is the strong dependency on population specificity. In fact, the perform ance of PRS can be significantly influenced by the linkage-disequlibrium (LD) across variants and allele frequencies that are specific to different populations 10. As a consequence, PRS has been mostly limited to European ancestry cohorts for which larger reference GWAS are available 11. In addition to LD and allele frequencies also gene-environment interactions might also be responsible for the different genetic susceptibilities toward a trait. Fo r instance the genetic liability of lipid levels is less understood in East Asian ancestry populations 11. Since individuals with East Asian ancestry account for more than a fifth of the global population, understanding genetic variation in East Asians is crucial to improve risk characterization and preventive interventions12. In the last few years the availability of large population-based cohorts and cross-ancestry GWAS enabled also the development of novel computational algorithms to improve the generalizability of PRS 13,14. A multi-ancestry, GWAS meta-analysis of lipid levels was conducted by the Global Lipid Genetics Co nsortium including 350,000 people of non- European ancestry, 150,000 East Asian individuals, and approximately 1.65 million people 5. The study also demonstrated that our understanding of the genetic component associated . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 6, 2023. ; https://doi.org/10.1101/2023.08.03.23293320doi: medRxiv preprint with lipid levels is significantly improved by increasing diversity rather than including additional European ancestry individuals. In this study, we derived ancestry-specific and cross-ancestry PRS to predict serum LDL level by first considering all populations and then focusing on East Asian individuals. Specifically, we derived six LDL-PRSs: four ances try-specific PRS (East Asian, South Asian, European, African) and two multi-ancestry PRSs (East Asian with European meta-analysis, and the four ancestry meta-analysis). The six PRSs were tested among nine population groups estimated from the UK Biobank (UKB, n=423,596). We focused on the East Asian ancestry group from the UKB and validated the PRS with participants from the Taiwan Biobank (TWB, n=68,978). We then tested the associations between PRS and LDL cholesterol changes among East Asian individuals in both biobanks.

Study subjects The analysis was performed using genetic and phenotypic data of the UKB and TWB. UKB is a population-based cohort study, with over 500,000 individuals aged 40 to 69 years at the time of recruitment 15. We excluded outliers with high genotype missing rates, putative sex chromosome aneuploidy, and discordant reported sex vs genotypic sex 16. We randomly excluded one from each pair of related individuals if the genetic relationship was closer than the second degree, defined as kinship coefficient > 0.0884 as calculated by the UKB. We applied a previous approach to divide UKB individuals into nine ancestry groups by projecting data onto the PCA space of 1000 Genomes Project 11. TWB is a Taiwanese-based cohort study, with 68,978 individuals aged from 30 to 75 across 750k SNPs 17. For more overlapping SNPs with PRS models, we imputed the TWB cohort. First, we filtered out SNPs based on the missing rate and Hardy- Weinberg equilibrium. Then, we imputed the genotype with a reference based on the whole genome sequencing data of 1,496 Taiwanese individuals. In total, we obtained 15 million SNPs for 69k Taiwanese individuals as our external validation set. UK Biobank ancestry grouping . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 6, 2023. ; https://doi.org/10.1101/2023.08.03.23293320doi: medRxiv preprint We assigned the samples to different countries using PC-projection as demonstrated in a previous study 11. In this previous study, the authors explored different methods to classify individuals into ancestry groups using principal component analysis (PCA) of genome-wide genotype data. They find that Euclidean distances in the PCA space are proportional to genetic differences between populations and recommend using this distance measure. They suggest using all principal components to capture population structure, as using only two or four is insufficient for distinguishing certain populations. They apply PCA-based distance to infer ancestry in datasets and propose two solutions: projecting PCs to reference populations or using internal data. They demonstrate that these solutions are effective for inferring ancestry and grouping genetically similar individuals. Here, we used this approach to define the nine ancestry groups based on UK Biobank data and birth country information, with some groups including individuals from neighboring countries (namely, East Asian: China; European: United Kingdom, Italy, and Poland; African: Nigeria, and Caribbean; South Asian: India; Middle East: Iran, and Ashkenazi Jewish). Additionally, we defined East Asian subpopulations by projecting samples in the 1000 Genomes Project PC space considering the five East Asian subpopulations as references. Construction of multi-ancestry polygenic score To evaluate the potential of PRS to predict increased LDL cholesterol levels in East Asian ancestry. We used the latest Global Lipid Genetics Consortium GWAS that was conducted in different populations to derive an ancestry-specific or multi- ancestry LDL PRS 5. We considered the summary statistics that did not include UK Biobank samples. Six PRSs were created: one for each ancestry (East Asian, South Asian, European, African), and two meta-analyses using multi-ancestry GWAS (one using East Asian and European ancestry, and the other using the four ancestries). PRSs weights were conducted using PRS-CSx 13 (accounting for population-specific allele frequencies and LD patterns ) and the 1000 Genomes Project as a reference panel that matched the ancestry of each discovery GWAS. The PRS-CSx method incorporates summary statistics from different GWAS and links the genetic effects across populations using a continuous shrinkage prior that is shared between them. This approach allows for more precise estimation of effect sizes by using information . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 6, 2023. ; https://doi.org/10.1101/2023.08.03.23293320doi: medRxiv preprint from the summary statistics and taking advantage of the variation in linkage disequilibrium across the discovery samples. By jointly modeling these multi- ancestry summary statistics, PRS-CSx might be able to better capture the underlying genetic effects and produce more accurate predictions. We developed the multi- ancestry PRS using the “--meta” option provided by the software. We tested each of the six PRSs in the nine population groups from the UKB. We then evaluated the six PRSs among the East Asian cohort of the TWB. We compared the PRS performance between individuals in TWB and two East Asian sub-populations from the UKB (namely Han Chinese South [CHS] and Kinh in Ho Chi Minh City, Vietnam [KHV]) from the 1000 Genomes Project. Notebly, that most of TWB individuals are clustered with the CHS group 18. We excluded the other three East Asian subpopulations due to sample size limitations. Assessment of PRS accuracy We assessed the prediction accuracy of the six PRSs in the nine estimated populations from the UKB and Taiwanese population from the TWB. We standardized PRSs to a mean of 0 and standard deviation of 1. Two models were used: (1) the full model which included PRS with sex, age, age 2 and the first four genetic principal components as covariates (formula: LDL ~ PRS + sex + age + age2 + PC1 + PC2 + PC3 + PC4), and (2) the reference model which only accounted for covariates (formula: LDL ~ sex + age + age 2 + PC1 + PC2 + PC3 + PC4). Linear regression was performed, and incremental R 2 was calculated following previous study19 as the difference between the adjusted R 2 of the full model (including PRS as an additional predictor) and the reference model. Mean LDL values across deciles of EAS_PRS, EUR_PRS, and multi-ancestry PRS were computed in all individuals of TWB to evaluate the range of phenotypic variability cover for these PRS.

Study populations In the UK Biobank, the estimated ethnic groups of the United Kingdom (UK) and China had significantly different study participant characteristics (Table 1). In comparison to people in the United Kingdom (UK), Chinese participants had lower LDL concentrations (mean, SD: 3.42 mmol/L, 0.77), lower TC levels (mean, SD: 5.54 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 6, 2023. ; https://doi.org/10.1101/2023.08.03.23293320doi: medRxiv preprint mmol/L, 1.03), and similar HDL levels (mean, SD: 1.46 mmol/L, 0.38). They were also younger (mean age, SD: 52.3, 7.71). The Chinese participants had a lower percentage of men than the UK (38.8% vs. 45.9%). Participants from China had a significantly lower body mass index (BMI) (mean, SD: 24.07 kg/m2, 3.4) than UK participants (p-value <2.2 x 10-16). . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 6, 2023. ; https://doi.org/10.1101/2023.08.03.23293320doi: medRxiv preprint Table 1 Study participant characteristics stratified by estimated ethnicity in UK Biobank. For continuous variables, p-values from the Welch t-statistic tests are reported, while for categorical and binary variables, p-values from Pearson’s Chi-squared tests are reported. HC, hypercholesterolemia; HDL, high-density lipoprotein cholesterol; LDL, low-density lipoprotein cholesterol; SD, standard deviation. Participants, N Males, N (%) Age, Mean (SD) HC Cases, N (%) HC Controls, N BMI, Mean (SD) LDL, Mean (SD) HDL, Mean (SD) TC, Mea n (SD) UK Biobank United Kingdom 423596 194259 (45.9) 56.81 (8.02) 110166 (26.01) 313430 (73.99) 27.4 (4.76) 3.57 (0.87) 1.45 (0.38) 5.71 (1.14) Italy 6451 2882 (44.7) 54.5 (8.41) 1624 (25.17) 4827 (74.83) 27.35 (4.94) 3.56 (0.86) 1.45 (0.38) 5.68 (1.12) India 6303 3413 (54.1) 53.42 (8.41) 1135 (18.01) 5168 (81.99) 27.42 (4.5) 3.35 (0.85) 1.25 (0.32) 5.31 (1.12) Poland 4095 1544 (37.7) 54.4 (7.53) 1088 (26.57) 3007 (73.43) 27.39 (4.96) 3.59 (0.85) 1.49 (0.4) 5.76 (1.13) Nigeria 3802 1744 (45.9) 51.95 (8.14) 551 (14.49) 3251 (85.51) 29.82 (5.31) 3.21 (0.84) 1.43 (0.35) 5.17 (1.09) Caribbean 2492 898 (36) 52.52 (8.13) 396 (15.89) 2096 (84.11) 29.49 (5.56) 3.28 (0.83) 1.47 (0.38) 5.29 (1.09) Ashkenazi 2359 1067 (45.2) 58.09 (7.1) 613 (25.99) 1746 (74.01) 27.13 (4.54) 3.55 (0.9) 1.44 (0.39) 5.68 (1.2) China 1480 545 (36.8) 52.33 (7.71) 263 (17.77) 1217 (82.23) 24.07 (3.4) 3.42 (0.77) 1.46 (0.38) 5.54 (1.03) Iran 1145 680 (59.4) 51.99 (7.98) 234 (20.44) 911 (79.56) 27.98 (4.55) 3.43 (0.86) 1.28 (0.33) 5.4 (1.11) P-Value: United Kingdom/China <1.1 x 10-16 <2.2 x 10-16 <2.9 x 10-16 <2.9 x 10-16 <2.2 x 10-16 <1.8 x 10-13 0.53 <1.6 x 10-10 Taiwan Biobank Taiwan 68,978 21,495 (31.2) 51.0 (10.9) 8,196 (13.5) 60,782 (86.5) 24.25 (3.8) 3.16 (0.82) 1.43 (0.35) 5.12 (0.93) . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 6, 2023. ; https://doi.org/10.1101/2023.08.03.23293320doi: medRxiv preprint In the TWB, the percentage of men is 31.2% which is lower than the percentage of Chinese participants in the UK Biobank, while the age distribution (mean, SD: 51.0, 10.9) is similar. In addition, TWB individuals had lower levels of lipid traits, including LDL (mean, SD: 3.16 mmol/L, 0.82), HDL (mean, SD: 1.43 mmol/L, 0.35), and TC (mean, SD: 5.12 mmol/L, 0.93), but higher BMI (mean, SD: 24.25 kg/m2, 3.80). Evaluation of the PRS in the nine-estimated populations from the UK Biobank We assessed the performance of ancestry-specific PRS for the LDL levels across the nine estimated populations in the UKB (Figure 1). As expected, the LDL PRS derived from European GWAS (EUR_PRS) was associated with the best performance in different European populations (namely, United Kingdom, Poland, and Italy) and in Middle East populations (namely, Ashkenazi Jews and Iranians). Similarly, the LDL PRS derived from African GWAS (AFR_PRS) showed the best performance in the population of African origin (Nigeria and Caribbean). The LDL PRS derived from the East Asian GWAS (EAS_PRS) was the best performing in the Chinese population. Surprisingly, when we tested EUR_PRS and PRS derived from the South Asian GWAS (SAS_PRS) in the India participants, EUR_PRS performed better than SAS_PRS. Concerning the multi-ancestry PRS, we tested a PRS derived from a meta-analysis of European and East Asian GWASs (EUR_EAS_PRS), and a global PRS derived from a meta analysis of the four ancestries (EUR_EAS_SAS_AFR_PRS). The multi- ancestry PRS showed comparable prediction to ancestry-specific PRS and seems to be more generalizable across populations, specifically for European, Middle East, and SAS populations. For instance, for the United Kingdom population, the adjusted R2% using EUR_PRS (8.62%) was similar to that using EUR_EAS_SAS_AFR_PRS (8.56%). For the AFR and EAS populations, ancestry-specific PRS performed better than multi-ancestry PRS. For instance, for the Chinese population, the adjusted R2% using EAS_PRS (6.35%) was higher than that using EUR_EAS_SAS_AFR_PRS (5.55%). . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 6, 2023. ; https://doi.org/10.1101/2023.08.03.23293320doi: medRxiv preprint Figure 1. Performance of ancestry-specific and multi-ancestry PRS models for LDL across nine sub-population in the UKB. Evaluation of the PRS in the Taiwan biobank Within the TWB, we evaluated the different ancestry-specific and multi-ancestry PRSs for the LDL levels (Figure 2). Similar to our findings in the UKB China participants, the EAS_PRS (adjusted R 2%=9.3%) also demonstrated better performance than EUR_PRS (adjusted R 2%=4.5%) in the TWB individuals and had even a better performance compared to multi-ancestry PRS (adjusted R 2%=6.7%). We also compared the performance of PRS between TWB individuals and the East Asian sub-populations from the UKB. We found that EAS_PRS has a comparable performance specially between populations from TWB (adjusted R 2%=6.5%) and CHS (adjusted R2=6.1%) from UKB. . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 6, 2023. ; https://doi.org/10.1101/2023.08.03.23293320doi: medRxiv preprint Figure 2. Performance of ancestry-specific and multi-ancestry PRS models for LDL on East Asian populations. Association between different PRS strata and LDL values We analyzed the mean of LDL levels in individuals from TWB based on their EAS_PRS, EUR_PRS, and multi-ancestry PRS deciles. We compared the difference in mean LDL levels between the lowest and highest deciles of EAS_PRS, EUR_PRS, and Multi PRS. Our findings showed that in East Asians, EAS_PRS explained a wider range of phenotypic variability compared to EUR_PRS. Specifically, the difference in mean LDL levels between the lowest and highest EAS_PRS deciles was 0.82, while for EUR_PRS it was 0.59 (Figure 3). The mean difference in LDL levels between the lowest and highest multi-ancestry PRS deciles was 0.76. However, the mean LDL levels in the highest deciles in both EAS_PRS and multi-ancestry PRS were the same (LDL mean (mmol/L) = 3.54, P=0.86). . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 6, 2023. ; https://doi.org/10.1101/2023.08.03.23293320doi: medRxiv preprint Figure 3. LDL mean values across deciles of EAS, EUR and Multi PRSs of East Asians (TWB) PRS.

In this study aimed at predicting LDL in two EAS populations (from UKB and TWB) using the latest Global Lipids Genetics Consortium GWAS. Our findings indicate that: 1) ancestry-specific PRS yield better performance in predicting LDL levels, and 2) multi-ancestry PRSs together with computational approaches integrating population specific LD-pattern can be used to enhance the generalizability of PRSs. In particular, the multi-ancestry PRSs showed that even relatively small proportions of non-European samples can significantly improve prediction in non-EUR populations. Our work emphasizes the importance of conducting global GWAS that include diverse populations to enhance the generalizability of PRSs, even when the availability of diverse population samples is limited. Our study further suggest that statistical genetics approaches can be used to take advantage of the already available global GWAS data, even when the number of non-European samples is limited. One example, the latest Global Lipids Genetics Consortium GWAS includes individuals across five genetic ancestry groups: admixed African or African (6.0% of the sample), East Asian (8.9%), European (79.8%), Hispanic (2.9%), and South Asian (2.5%) 5. Recently published Bayesian PRS approaches demonstrated an improvement in the accuracy of PRSs in non- European populations by utilizing common genetic effects across ancestries 13,14. . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 6, 2023. ; https://doi.org/10.1101/2023.08.03.23293320doi: medRxiv preprint Another recent study, the authors conducted a benchmarking analysis to compare several PRS methods for multi-ancestry analysis in the UKBB dataset, which included lipid traits and EAS data using also global lipid consortium GWAS data 20. The findings of this study provided insights on the use of statistical methods to improve prediction performance in non-Europeans. The applicability of the findings on the portability of PRS from multi-ancestry meta- analyses to other traits needs to be taken into account, considering multiple factors21. These factors include the heritability of the trait 22, genetic correlation 23, causal variants allele frequencies 24, gene-environment interactions 25, and the inclusion of multi-ancestry populations in GWAS 26,27. In a recent study, they estimated the cross-ancestry genetic correlation for cholesterol and observed a significant genetic heterogeneity between ancestries for total and LDL cholesterol 22. While many traits exhibit a significant shared genetic correlation across ancestries, indicating potential transferability of multi-ancestry PRS 28, some traits have specific genetic variations that are more commonly found in particular ancestral groups 29,30. To ensure the effective use of PRS in diverse populations, it is crucial to conduct comprehensive investigations considering these factors and include a representative range of ancestries in future GWAS studies 31. Moreover, a recent study emphasizes the necessity of moving away from discrete genetic ancestry clusters and embracing the continuum of genetic ancestries when analyzing and interpreting PGS 10. By accounting for individual variation and considering the diverse genetic backgrounds within populations, more accurate PGS assessments can be achieved. By leveraging the available diverse GWAS data, we can improve the generalizability of PRSs, and ultimately enhance our ability to predict complex disease risk across diverse populations. As such, our study provides valuable insights into the development and implementation of PRSs for predicting lipid traits in East Asian populations, and highlights the need for continued efforts to increase diversity in genetic research while also working on bioinformatics approaches to meta-analyze the association signal across different populations. . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 6, 2023. ; https://doi.org/10.1101/2023.08.03.23293320doi: medRxiv preprint

In our study we evaluated the performance of ancestry-specific and multi-ancestry PRSs for LDL, in various populations including East Asians from the UK Biobank and Taiwan Biobank. The findings corroborated that ancestry specific PRSs performed better than out of target population PRSs in the respective ancestry. In particular, the EAS_PRS had better performance in East Asian populations, while the EUR_PRS showed better performance in European and Middle East populations. The multi- ancestry PRS analysis showed that even a small proportion of non-European samples can significantly improve prediction in non-EUR populations. These findings provide valuable insights into the development of PRSs for diverse populations and the potential clinical applications of PRSs. On the one hand, our analysis suggests that incorporating cross-ancestry GWAS data and utilizing optimized computational algorithms to account for population-specific LD-patterns can improve the generalizability of PRS. On the other hand, these results further emphasize the necessity of enhancing genetic diversity in GWAS studies and establishing large- scale population-based cohorts to more accurately model the genetic liability of multifactorial traits, such as LDL cholesterol. List of abbreviations PRS: Polygenic risk scores UKB: UK TWB: Taiwan Biobank TC: total cholesterol LDL-C: low-density lipoprotein cholesterol GWAS: genome-wide association studies CHS: Han Chinese South KHV: Kinh in Ho Chi Minh City, Vietnam EUR: European EAS: East Asian SAS: South Asian . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 6, 2023. ; https://doi.org/10.1101/2023.08.03.23293320doi: medRxiv preprint AFR: African PC: Principal Component Declarations Ethics approval and consent to participate This study made use of anonymized data from UK Biobank and Taiwan Biobank. Written informed consent was provided by all individuals. The research ethics committee for the UK Biobank and Taiwan Biobank gave its approval to the protocol and the permission. Our investigation was carried out in accordance with approved UK Biobank data application number 52446 and Taiwan Biobank data application number TWBR10411-03. We hereby confirm that, all methods were carried out in accordance with relevant guidelines and regulations. Consent for publication: Not Applicable Availability of data and code Access to genome-wide genotyping data, and phenotypic data from the UK and Taiwan Biobank can be obtained through a successful project application process. Detailed information about the application process can be found at (http://www.ukbiobank.ac.uk/about-biobank-uk/ ) and (https://taiwanview.twbiobank.org.tw/data_appl). Certain restrictions apply to the availability of these data, as they were used under license for the current study. The codes related to the statistical analysis for this study have been deposited on GitLab at the following location (doi:10.17881/8wqn-x712). Competing Interests: No potential conflicts (financial, professional, or personal) for all authors relevant to the manuscript. Funding: PM received funding from the Luxembourg National Research Fund (FNR) INTER grant ‘ProtectMove’ (INTER/DFG/19/14429377). EH, DRB and PM were supported by the FNR INTER INTER/DFG/21/16394868. Authors’ contributions: EH, KL, YL and CM performed the statistical analysis and the bioinformatics. EH, KL, CC and CM conceived and designed the study. EH, KL, CC and CM drafted the initial manuscript. EH, KL, TH, RA, YL, DRB, PK, PM, CC . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 6, 2023. ; https://doi.org/10.1101/2023.08.03.23293320doi: medRxiv preprint and CM performed the critical expert revision. CC and CM supervised the study. All authors read and approved the final manuscript.

UK Biobank analyses were conducted via appl ication 52446 using a protocol approved by the Partners HealthCare Institutional Review Board and Taiwan Biobank using data application number TWBR10411-03.

1. Nelson, R. H. Hyperlipidemia as a Risk Factor for Cardiovascular Disease. Prim Care 40, 195 – 211 (2013). 2. Weiss, L. A., Pan, L., Abney, M. & Ober, C. The sex-specific genetic architecture of quantitative traits in humans. Nat Genet 38, 218– 222 (2006). 3. Sollis, E. et al. The NHGRI-EBI GWAS Catalog: knowledgebase and deposition resource. Nucleic Acids Res 51, D977– D985 (2022). 4. Khera, A. V. et al. Genome-wide polygenic scores for common diseases identify individuals with risk equivalent to monogenic mutations. Nat Genet 50, 1219– 1224 (2018). 5. Graham, S. E. et al. The power of genetic diversity in genome-wide association studies of lipids. Nature 600, 675– 679 (2021). 6. Fahed, A. C. et al. Polygenic background modifies penetrance of monogenic variants for tier 1 genomic conditions. Nat Commun 11, 3635 (2020). 7. Hassanin, E. et al. Breast and prostate cancer risk: The interplay of polygenic risk, rare pathogenic germline variants, and family history. Genetics in Medicine 24, 576– 585 (2022). . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 6, 2023. ; https://doi.org/10.1101/2023.08.03.23293320doi: medRxiv preprint 8. Hassanin, E. et al. Clinically relevant combined effect of polygenic background, rare pathogenic germline variants, and family history on colorectal cancer incidence. BMC Medical Genomics 16, 42 (2023). 9. Dornbos, P. et al. A combined polygenic score of 21,293 rare and 22 common variants improves diabetes diagnosis based on hemoglobin A1C levels. Nat Genet 54, 1609– 1614 (2022). 10. Ding, Y. et al. Polygenic scoring accuracy varies across the genetic ancestry continuum. Nature 618, 774– 781 (2023). 11. Privé , F. et al. Portability of 245 polygenic scores when derived from the UK Biobank and applied to 9 ancestry groups from the same cohort. The American Journal of Human Genetics 109, 12– 23 (2022). 12. Ge, T. et al. Development and validation of a trans-ancestry polygenic risk score for type 2 diabetes in diverse populations. Genome Medicine 14, 70 (2022). 13. Ruan, Y. et al. Improving polygenic prediction in ancestrally diverse populations. Nat Genet 54, 573– 580 (2022). 14. Hoggart, C. et al. BridgePRS: A powerful trans-ancestry Polygenic Risk Score method. 2023.02.17.528938 Preprint at https://doi.org/10.1101/2023.02.17.528938 (2023). 15. Bycroft, C. et al. The UK Biobank resource with deep phenotyping and genomic data. Nature 562, 203– 209 (2018). 16. Hassanin, E. et al. Assessing the role of polygenic background on the penetrance of monogenic forms in Parkinson’ s disease. 2021.06.06.21253270 Preprint at https://doi.org/10.1101/2021.06.06.21253270 (2021). . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 6, 2023. ; https://doi.org/10.1101/2023.08.03.23293320doi: medRxiv preprint 17. Wei, C.-Y. et al. Genetic profiles of 103,106 individuals in the Taiwan Biobank provide insights into the health and history of Han Chinese. NPJ Genom Med 6, 10 (2021). 18. Chen, C.-H. et al. Population structure of Han Chinese in the modern Taiwanese population based on 10,000 participants in the Taiwan Biobank project. Hum Mol Genet 25, 5321– 5331 (2016). 19. Huang, Q. Q. et al. Transferability of genetic loci and polygenic scores for cardiometabolic traits in British Pakistani and Bangladeshi individuals. Nat Commun 13, 4664 (2022). 20. Zhang, H. et al. Novel Methods for Multi-ancestry Polygenic Prediction and their Evaluations in 5.1 Million Individuals of Diverse Ancestry. 2022.03.24.485519 Preprint at https://doi.org/10.1101/2022.03.24.485519 (2023). 21. Majara, L. et al. Low and differential polygenic score generalizability among African populations due largely to genetic diversity. Human Genetics and Genomics Advances 4, 100184 (2023). 22. Momin, M. M., Zhou, X., Hyppö nen, E., Benyamin, B. & Lee, S. H. Cross- ancestry genetic architecture and prediction for cholesterol traits. 2023.01.31.23285307 Preprint at https://doi.org/10.1101/2023.01.31.23285307 (2023). 23. Shi, H. et al. Population-specific causal disease effect sizes in functionally important regions impacted by selection. Nat Commun 12, 1098 (2021). 24. Cavazos, T. B. & Witte, J. S. Inclusion of variants discovered from diverse populations improves polygenic risk score transferability. HGG Adv 2, 100017 (2020). . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 6, 2023. ; https://doi.org/10.1101/2023.08.03.23293320doi: medRxiv preprint 25. Peterson, R. E. et al. Genome-wide association studies in ancestrally diverse populations: opportunities, methods, pitfalls, and recommendations. Cell 179, 589– 603 (2019). 26. Fatumo, S. et al. A roadmap to increase diversity in genomic studies. Nat Med 28, 243– 250 (2022). 27. Yengo, L. et al. A saturated map of common genetic variants associated with human height. Nature 610, 704– 712 (2022). 28. Ho, W.-K. et al. European polygenic risk score for prediction of breast cancer shows similar performance in Asian women. Nat Commun 11, 3833 (2020). 29. El-Boraie, A. et al. TRANSFERABILITY OF ANCESTRY-SPECIFIC AND CROSS-ANCESTRY CYP2A6 ACTIVITY GENETIC RISK SCORES IN AFRICAN AND EUROPEAN POPULATIONS. Clin Pharmacol Ther 110, 975– 985 (2021). 30. Kamiza, A. B. et al. Transferability of genetic risk scores in African populations. Nat Med 28, 1163– 1166 (2022). 31. Duncan, L. et al. Analysis of polygenic risk score usage and performance in diverse human populations. Nat Commun 10, 3328 (2019). . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted August 6, 2023. ; https://doi.org/10.1101/2023.08.03.23293320doi: medRxiv preprint