Principles and methods for transferring polygenic risk scores across global populations

Managing differential performance of polygenic risk scores across groups: Real-world experience of the eMERGE Network

The differential performance of polygenic risk scores (PRSs) by group is one of the major ethical barriers to their clinical use. It is also one of the main practical challenges for any implementation effort. The social repercussions of how people are grouped in PRS research must be considered in communications with research participants, including return of results. Here, we outline the decisions faced and choices made by a large multi-site clinical implementation study returning PRSs to diverse participants in handling this issue of differential performance. Our approach to managing the complexities associated with the differential performance of PRSs serves as a case study that can help future implementers of PRSs to plot an anticipatory course in response to this issue.

Genome wide association study and genomic risk prediction of age related macular degeneration in Israel

The risk of developing age-related macular degeneration (AMD) is influenced by genetic background. In 2016, the International AMD Genomics Consortium (IAMDGC) identified 52 risk variants in 34 loci, and a polygenic risk score (PRS) from these variants was associated with AMD. The Israeli population has a unique genetic composition: Ashkenazi Jewish (AJ), Jewish non-Ashkenazi, and Arab sub-populations. We aimed to perform a genome-wide association study (GWAS) for AMD in Israel, and to evaluate PRSs for AMD. Our discovery set recruited 403 AMD patients and 256 controls at Hadassah Medical Center. We genotyped individuals via custom exome chip. We imputed non-typed variants using cosmopolitan and AJ reference panels. We recruited additional 155 cases and 69 controls for validation. To evaluate predictive power of PRSs for AMD, we used IAMDGC summary-statistics excluding our study and developed PRSs via clumping/thresholding or LDpred2. In our discovery set, 31/34 loci reported by IAMDGC were AMD-associated (P < 0.05). Of those, all effects were directionally consistent with IAMDGC and 11 loci had a P-value under Bonferroni-corrected threshold (0.05/34 = 0.0015). At a 5 × 10-5 threshold, we discovered four suggestive associations in FAM189A1, IGDCC4, C7orf50, and CNTNAP4. Only the FAM189A1 variant was AMD-associated in the replication cohort after Bonferroni-correction. A prediction model including LDpred2-based PRS + covariates had an AUC of 0.82 (95% CI 0.79-0.85) and performed better than covariates-only model (P = 5.1 × 10-9). Therefore, previously reported AMD-associated loci were nominally associated with AMD in Israel. A PRS developed based on a large international study is predictive in Israeli populations.

Optimizing clinico-genomic disease prediction across ancestries: a machine learning strategy with Pareto improvement

BACKGROUND

Accurate prediction of an individual's predisposition to diseases is vital for preventive medicine and early intervention. Various statistical and machine learning models have been developed for disease prediction using clinico-genomic data. However, the accuracy of clinico-genomic prediction of diseases may vary significantly across ancestry groups due to their unequal representation in clinical genomic datasets.

METHODS

We introduced a deep transfer learning approach to improve the performance of clinico-genomic prediction models for data-disadvantaged ancestry groups. We conducted machine learning experiments on multi-ancestral genomic datasets of lung cancer, prostate cancer, and Alzheimer's disease, as well as on synthetic datasets with built-in data inequality and distribution shifts across ancestry groups.

RESULTS

Deep transfer learning significantly improved disease prediction accuracy for data-disadvantaged populations in our multi-ancestral machine learning experiments. In contrast, transfer learning based on linear frameworks did not achieve comparable improvements for these data-disadvantaged populations.

CONCLUSIONS

This study shows that deep transfer learning can enhance fairness in multi-ancestral machine learning by improving prediction accuracy for data-disadvantaged populations without compromising prediction accuracy for other populations, thus providing a Pareto improvement towards equitable clinico-genomic prediction of diseases.

Prospective of Mitochondrial DNA Variations in Cancer on Genomic Medicine

Mitochondrial DNA (mtDNA) has emerged as a pivotal component in understanding the etiology and susceptibility of cancer. A recent study by Chen and colleagues delineated the germline genetic effect of mtDNA single-nucleotide polymorphisms (SNP) and haplogroups across pan-cancer risk. They identified a subset of mtSNPs and the corresponding risk score, as well as haplogroups A and M7 alongside their genetic interactions, conferring a protective effect against various cancers. These findings underscored the value of mtDNA variations as biomarkers for cancer etiology and as tools for cancer risk stratification. Future investigations are encouraged to integrate comprehensive omics data of genomics, transcriptomics, proteomics, and metabolomics, etc., from nuclear DNA with mtDNA variations, alongside single-cell and spatial technologies, to unravel the tumor mechanism and identify the drug targets. Moreover, the incorporation of polygenic risk score, that included mtDNA variations with both rare and common frequencies, and liquid biopsy-based biomarkers would enhance the predictive performance of cancer risk assessment and refine the risk stratification of population-based cancer screening. This commentary advocates for the validation across diverse populations to harness the full potential of mitochondrial genomics, and ultimately paves the prospective way for advancements in personalized cancer therapeutics and prevention strategies. See related article by Chen and colleagues, Cancer Epidemiol Biomarkers Prev 2024;33:381-8.

Multi-organ imaging-derived polygenic indexes for brain and body health

The UK Biobank (UKB) imaging project is a crucial resource for biomedical research, but is limited to 100,000 participants due to cost and accessibility barriers. Here we used genetic data to predict heritable imaging-derived phenotypes (IDPs) for a larger cohort. We developed and evaluated 4,375 IDP genetic scores (IGS) derived from UKB brain and body images. When applied to UKB participants who were not imaged, IGS revealed links to numerous phenotypes and stratified participants at increased risk for both brain and somatic diseases. For example, IGS identified individuals at higher risk for Alzheimer's disease and multiple sclerosis, offering additional insights beyond traditional polygenic risk scores of these diseases. When applied to independent external cohorts, IGS also stratified those at high disease risk in the All of Us Research Program and the Alzheimer's Disease Neuroimaging Initiative study. Our results demonstrate that, while the UKB imaging cohort is largely healthy and may not be the most enriched for disease risk management, it holds immense potential for stratifying the risk of various brain and body diseases in broader external genetic cohorts.

Ancestry, ethnicity, and race: explaining inequalities in cardiometabolic disease

Population differences in cardiometabolic disease remain unexplained. Misleading assumptions over genetic explanations are partly due to terminology used to distinguish populations, specifically ancestry, race, and ethnicity. These terms differentially implicate environmental and biological causal pathways, which should inform their use. Genetic variation alone accounts for a limited fraction of population differences in cardiometabolic disease. Research effort should focus on societally driven, lifelong environmental determinants of population differences in disease. Rather than pursuing population stratifiers to personalize medicine, we advocate removing socioeconomic barriers to receipt of and adherence to healthcare interventions, which will have markedly greater impact on improving cardiometabolic outcomes. This requires multidisciplinary collaboration and public and policymaker engagement to address inequalities driven by society rather than biology per se.

Polygenic Risk for Type 2 Diabetes in African Americans

African Americans (AAs) have been underrepresented in polygenic risk score (PRS) studies. Here, we integrated genome-wide data from multiple observational studies on type 2 diabetes (T2D), encompassing a total of 101,987 AAs, to train and optimize an AA-focused T2D PRS (PRSAA), using a Bayesian polygenic modeling method. We further tested the score in three independent studies with a total of 7,275 AAs and compared the PRSAA with other published scores. Results show that a 1-SD increase in the PRSAA was associated with 40-60% increase in the odds of T2D (odds ratio [OR] 1.60, 95% CI 1.37-1.88; OR 1.40, 95% CI 1.16-1.70; and OR 1.45, 95% CI 1.30-1.62) across three testing cohorts. These models captured 1.0-2.6% of the variance (R2) in T2D on the liability scale. The positive predictive values for three calculated score thresholds (the top 2%, 5%, and 10%) ranged from 14 to 35%. The PRSAA, in general, performed similarly to existing T2D PRS. The need remains for larger data sets to continue to evaluate the utility of within-ancestry scores in the AA population.

ARTICLE HIGHLIGHTS

Screening embryos for polygenic disease risk: a review of epidemiological, clinical, and ethical considerations

BACKGROUND

The genetic composition of embryos generated by in vitro fertilization (IVF) can be examined with preimplantation genetic testing (PGT). Until recently, PGT was limited to detecting single-gene, high-risk pathogenic variants, large structural variants, and aneuploidy. Recent advances have made genome-wide genotyping of IVF embryos feasible and affordable, raising the possibility of screening embryos for their risk of polygenic diseases such as breast cancer, hypertension, diabetes, or schizophrenia. Despite a heated debate around this new technology, called polygenic embryo screening (PES; also PGT-P), it is already available to IVF patients in some countries. Several articles have studied epidemiological, clinical, and ethical perspectives on PES; however, a comprehensive, principled review of this emerging field is missing.

OBJECTIVE AND RATIONALE

This review has four main goals. First, given the interdisciplinary nature of PES studies, we aim to provide a self-contained educational background about PES to reproductive specialists interested in the subject. Second, we provide a comprehensive and critical review of arguments for and against the introduction of PES, crystallizing and prioritizing the key issues. We also cover the attitudes of IVF patients, clinicians, and the public towards PES. Third, we distinguish between possible future groups of PES patients, highlighting the benefits and harms pertaining to each group. Finally, our review, which is supported by ESHRE, is intended to aid healthcare professionals and policymakers in decision-making regarding whether to introduce PES in the clinic, and if so, how, and to whom.

SEARCH METHODS

We searched for PubMed-indexed articles published between 1/1/2003 and 1/3/2024 using the terms 'polygenic embryo screening', 'polygenic preimplantation', and 'PGT-P'. We limited the review to primary research papers in English whose main focus was PES for medical conditions. We also included papers that did not appear in the search but were deemed relevant.

OUTCOMES

The main theoretical benefit of PES is a reduction in lifetime polygenic disease risk for children born after screening. The magnitude of the risk reduction has been predicted based on statistical modelling, simulations, and sibling pair analyses. Results based on all methods suggest that under the best-case scenario, large relative risk reductions are possible for one or more diseases. However, as these models abstract several practical limitations, the realized benefits may be smaller, particularly due to a limited number of embryos and unclear future accuracy of the risk estimates. PES may negatively impact patients and their future children, as well as society. The main personal harms are an unindicated IVF treatment, a possible reduction in IVF success rates, and patient confusion, incomplete counselling, and choice overload. The main possible societal harms include discarded embryos, an increasing demand for 'designer babies', overemphasis of the genetic determinants of disease, unequal access, and lower utility in people of non-European ancestries. Benefits and harms will vary across the main potential patient groups, comprising patients already requiring IVF, fertile people with a history of a severe polygenic disease, and fertile healthy people. In the United States, the attitudes of IVF patients and the public towards PES seem positive, while healthcare professionals are cautious, sceptical about clinical utility, and concerned about patient counselling.

WIDER IMPLICATIONS

The theoretical potential of PES to reduce risk across multiple polygenic diseases requires further research into its benefits and harms. Given the large number of practical limitations and possible harms, particularly unnecessary IVF treatments and discarded viable embryos, PES should be offered only within a research context before further clarity is achieved regarding its balance of benefits and harms. The gap in attitudes between healthcare professionals and the public needs to be narrowed by expanding public and patient education and providing resources for informative and unbiased genetic counselling.

Generalizability of polygenic prediction models: how is the R2 defined on test data?

BACKGROUND

Polygenic risk scores (PRS) quantify an individual's genetic predisposition for different traits and are expected to play an increasingly important role in personalized medicine. A crucial challenge in clinical practice is the generalizability and transferability of PRS models to populations with different ancestries. When assessing the generalizability of PRS models for continuous traits, the R 2 is a commonly used measure to evaluate prediction accuracy. While the R 2 is a well-defined goodness-of-fit measure for statistical linear models, there exist different definitions for its application on test data, which complicates interpretation and comparison of results.

METHODS

Based on large-scale genotype data from the UK Biobank, we compare three definitions of the R 2 on test data for evaluating the generalizability of PRS models to different populations. Polygenic models for several phenotypes, including height, BMI and lipoprotein A, are derived based on training data with European ancestry using state-of-the-art regression methods and are evaluated on various test populations with different ancestries.

RESULTS

Our analysis shows that the choice of the R 2  definition can lead to considerably different results on test data, making the comparison of R 2  values from the literature problematic. While the definition as the squared correlation between predicted and observed phenotypes solely addresses the discriminative performance and always yields values between 0 and 1, definitions of the R 2 based on the mean squared prediction error (MSPE) with reference to intercept-only models assess both discrimination and calibration. These MSPE-based definitions can yield negative values indicating miscalibrated predictions for out-of-target populations. We argue that the choice of the most appropriate definition depends on the aim of PRS analysis - whether it primarily serves for risk stratification or also for individual phenotype prediction. Moreover, both correlation-based and MSPE-based definitions of R 2 can provide valuable complementary information.

CONCLUSIONS

Awareness of the different definitions of the R 2 on test data is necessary to facilitate the reporting and interpretation of results on PRS generalizability. It is recommended to explicitly state which definition was used when reporting R 2 values on test data. Further research is warranted to develop and evaluate well-calibrated polygenic models for diverse populations.

Genetic association and transferability for urinary albumin-creatinine ratio as a marker of kidney disease in four Sub-Saharan African populations and non-continental individuals of African ancestry

Background

Genome-wide association studies (GWAS) have predominantly focused on populations of European and Asian ancestry, limiting our understanding of genetic factors influencing kidney disease in Sub-Saharan African (SSA) populations. This study presents the largest GWAS for urinary albumin-to-creatinine ratio (UACR) in SSA individuals, including 8,970 participants living in different African regions and an additional 9,705 non-resident individuals of African ancestry from the UK Biobank and African American cohorts.

Methods

Urine biomarkers and genotype data were obtained from two SSA cohorts (AWI-Gen and ARK), and two non-resident African-ancestry studies (UK Biobank and CKD-Gen Consortium). Association testing and meta-analyses were conducted, with subsequent fine-mapping, conditional analyses, and replication studies. Polygenic scores (PGS) were assessed for transferability across populations.

Results

Two genome-wide significant (P < 5 × 10-8) UACR-associated loci were identified, one in the BMP6 region on chromosome 6, in the meta-analysis of resident African individuals, and another in the HBB region on chromosome 11 in the meta-analysis of non-resident SSA individuals, as well as the combined meta-analysis of all studies. Replication of previous significant results confirmed associations in known UACR-associated regions, including THB53, GATM, and ARL15. PGS estimated using previous studies from European ancestry, African ancestry, and multi-ancestry cohorts exhibited limited transferability of PGS across populations, with less than 1% of observed variance explained.

Conclusion

This study contributes novel insights into the genetic architecture of kidney disease in SSA populations, emphasizing the need for conducting genetic research in diverse cohorts. The identified loci provide a foundation for future investigations into the genetic susceptibility to chronic kidney disease in underrepresented African populations Additionally, there is a need to develop integrated scores using multi-omics data and risk factors specific to the African context to improve the accuracy of predicting disease outcomes.

Investigating the sources of variable impact of pathogenic variants in monogenic metabolic conditions

Over three percent of people carry a dominant pathogenic variant, yet only a fraction of carriers develop disease. Disease phenotypes from carriers of variants in the same gene range from mild to severe. Here, we investigate underlying mechanisms for this heterogeneity: variable variant effect sizes, carrier polygenic backgrounds, and modulation of carrier effect by genetic background (marginal epistasis). We leveraged exomes and clinical phenotypes from the UK Biobank and the Mt. Sinai BioMe Biobank to identify carriers of pathogenic variants affecting cardiometabolic traits. We employed recently developed methods to study these cohorts, observing strong statistical support and clinical translational potential for all three mechanisms of variable carrier penetrance and disease severity. For example, scores from our recent model of variant pathogenicity were tightly correlated with phenotype amongst clinical variant carriers, they predicted effects of variants of unknown significance, and they distinguished gain- from loss-of-function variants. We also found that polygenic scores predicted phenotypes amongst pathogenic carriers and that epistatic effects can exceed main carrier effects by an order of magnitude.

An ensemble penalized regression method for multi-ancestry polygenic risk prediction

Great efforts are being made to develop advanced polygenic risk scores (PRS) to improve the prediction of complex traits and diseases. However, most existing PRS are primarily trained on European ancestry populations, limiting their transferability to non-European populations. In this article, we propose a novel method for generating multi-ancestry Polygenic Risk scOres based on enSemble of PEnalized Regression models (PROSPER). PROSPER integrates genome-wide association studies (GWAS) summary statistics from diverse populations to develop ancestry-specific PRS with improved predictive power for minority populations. The method uses a combination ofL 1 (lasso) andL 2 (ridge) penalty functions, a parsimonious specification of the penalty parameters across populations, and an ensemble step to combine PRS generated across different penalty parameters. We evaluate the performance of PROSPER and other existing methods on large-scale simulated and real datasets, including those from 23andMe Inc., the Global Lipids Genetics Consortium, and All of Us. Results show that PROSPER can substantially improve multi-ancestry polygenic prediction compared to alternative methods across a wide variety of genetic architectures. In real data analyses, for example, PROSPER increased out-of-sample prediction R2 for continuous traits by an average of 70% compared to a state-of-the-art Bayesian method (PRS-CSx) in the African ancestry population. Further, PROSPER is computationally highly scalable for the analysis of large SNP contents and many diverse populations.

Genetic Association and Transferability for Urinary Albumin-Creatinine Ratio as a Marker of Kidney Disease in four Sub-Saharan African Populations and non-continental Individuals of African Ancestry

BACKGROUND

Genome-wide association studies (GWAS) have predominantly focused on populations of European and Asian ancestry, limiting our understanding of genetic factors influencing kidney disease in Sub-Saharan African (SSA) populations. This study presents the largest GWAS for urinary albumin-to-creatinine ratio (UACR) in SSA individuals, including 8,970 participants living in different African regions and an additional 9,705 non-resident individuals of African ancestry from the UK Biobank and African American cohorts.

METHODS

Urine biomarkers and genotype data were obtained from two SSA cohorts (AWI-Gen and ARK), and two non-resident African-ancestry studies (UK Biobank and CKD-Gen Consortium). Association testing and meta-analyses were conducted, with subsequent fine-mapping, conditional analyses, and replication studies. Polygenic scores (PGS) were assessed for transferability across populations.

RESULTS

Two genome-wide significant (P<5x10-8) UACR-associated loci were identified, one in the BMP6 region on chromosome 6, in the meta-analysis of resident African individuals, and another in the HBB region on chromosome 11 in the meta-analysis of non-resident SSA individuals, as well as the combined meta-analysis of all studies. Replication of previous significant results confirmed associations in known UACR-associated regions, including THB53, GATM, and ARL15. PGS estimated using previous studies from European ancestry, African ancestry, and multi-ancestry cohorts exhibited limited transferability of PGS across populations, with less than 1% of observed variance explained.

CONCLUSION

This study contributes novel insights into the genetic architecture of kidney disease in SSA populations, emphasizing the need for conducting genetic research in diverse cohorts. The identified loci provide a foundation for future investigations into the genetic susceptibility to chronic kidney disease in underrepresented African populations Additionally, there is a need to develop integrated scores using multi-omics data and risk factors specific to the African context to improve the accuracy of predicting disease outcomes.

METHODS

Urine biomarkers and genotype data were obtained from two SSA cohorts (AWI-Gen and ARK), and two non-resident African-ancestry studies (UK Biobank and CKD-Gen Consortium). Association testing and meta-analyses were conducted, with subsequent fine-mapping, conditional analyses, and replication studies. Polygenic scores (PGS) were assessed for transferability across populations.

RESULTS

Two genome-wide significant (P<5x10-8) UACR-associated loci were identified, one in the BMP6 region on chromosome 6, in the meta-analysis of resident African individuals, and another in the HBB region on chromosome 11 in the meta-analysis of non-resident SSA individuals, as well as the combined meta-analysis of all studies. Replication of previous significant results confirmed associations in known UACR-associated regions, including THB53, GATM, and ARL15. PGS estimated using previous studies from European ancestry, African ancestry, and multi-ancestry cohorts exhibited limited transferability of PGS across populations, with less than 1% of observed variance explained.

CONCLUSION

This study contributes novel insights into the genetic architecture of kidney function in SSA populations, emphasizing the need for conducting genetic research in diverse cohorts. The identified loci provide a foundation for future investigations into the genetic susceptibility to chronic kidney disease in underrepresented African populations.

Evaluating and improving health equity and fairness of polygenic scores

Polygenic scores (PGSs) are quantitative metrics for predicting phenotypic values, such as human height or disease status. Some PGS methods require only summary statistics of a relevant genome-wide association study (GWAS) for their score. One such method is Lassosum, which inherits the model selection advantages of Lasso to select a meaningful subset of the GWAS single-nucleotide polymorphisms as predictors from their association statistics. However, even efficient scores like Lassosum, when derived from European-based GWASs, are poor predictors of phenotype for subjects of non-European ancestry; that is, they have limited portability to other ancestries. To increase the portability of Lassosum, when GWAS information and estimates of linkage disequilibrium are available for both ancestries, we propose Joint-Lassosum (JLS). In the simulation settings we explore, JLS provides more accurate PGSs compared to other methods, especially when measured in terms of fairness. In analyses of UK Biobank data, JLS was computationally more efficient but slightly less accurate than a Bayesian comparator, SDPRX. Like all PGS methods, JLS requires selection of predictors, which are determined by data-driven tuning parameters. We describe a new approach to selecting tuning parameters and note its relevance for model selection for any PGS. We also draw connections to the literature on algorithmic fairness and discuss how JLS can help mitigate fairness-related harms that might result from the use of PGSs in clinical settings. While no PGS method is likely to be universally portable, due to the diversity of human populations and unequal information content of GWASs for different ancestries, JLS is an effective approach for enhancing portability and reducing predictive bias.

MUSSEL: Enhanced Bayesian polygenic risk prediction leveraging information across multiple ancestry groups

Polygenic risk scores (PRSs) are now showing promising predictive performance on a wide variety of complex traits and diseases, but there exists a substantial performance gap across populations. We propose MUSSEL, a method for ancestry-specific polygenic prediction that borrows information in summary statistics from genome-wide association studies (GWASs) across multiple ancestry groups via Bayesian hierarchical modeling and ensemble learning. In our simulation studies and data analyses across four distinct studies, totaling 5.7 million participants with a substantial ancestral diversity, MUSSEL shows promising performance compared to alternatives. For example, MUSSEL has an average gain in prediction R2 across 11 continuous traits of 40.2% and 49.3% compared to PRS-CSx and CT-SLEB, respectively, in the African ancestry population. The best-performing method, however, varies by GWAS sample size, target ancestry, trait architecture, and linkage disequilibrium reference samples; thus, ultimately a combination of methods may be needed to generate the most robust PRSs across diverse populations.

Admix-kit: an integrated toolkit and pipeline for genetic analyses of admixed populations

SUMMARY

Admixed populations, with their unique and diverse genetic backgrounds, are often underrepresented in genetic studies. This oversight not only limits our understanding but also exacerbates existing health disparities. One major barrier has been the lack of efficient tools tailored for the special challenges of genetic studies of admixed populations. Here, we present admix-kit, an integrated toolkit and pipeline for genetic analyses of admixed populations. Admix-kit implements a suite of methods to facilitate genotype and phenotype simulation, association testing, genetic architecture inference, and polygenic scoring in admixed populations.

AVAILABILITY AND IMPLEMENTATION

Admix-kit package is open-source and available at https://github.com/KangchengHou/admix-kit. Additionally, users can use the pipeline designed for admixed genotype simulation available at https://github.com/UW-GAC/admix-kit_workflow.

Genomic Insights for Personalized Care: Motivating At-Risk Individuals Toward Evidence-Based Health Practices

Lung cancer and tobacco use pose significant global health challenges and require a comprehensive translational roadmap for improved prevention strategies. We propose the GREAT care paradigm ( G enomic Informed Care for Motivating High R isk Individuals E ligible for Evidence-b a sed Prevention), which employs polygenic risk scores (PRSs) to stratify disease risk and personalize interventions, such as lung cancer screening and tobacco treatment. We developed PRSs using large-scale multi-ancestry genome-wide association studies and adjusted for genetic ancestry for standardized risk stratification across diverse populations. We applied our PRSs to over 340,000 individuals of diverse ethnic background and found significant odds ratios for lung cancer and difficulty quitting smoking. These findings enable the evaluation of PRS-based interventions in ongoing trials aimed at motivating health behavior changes in high-risk patients. This pioneering approach enhances primary care with genomic insights, promising improved outcomes in cancer prevention and tobacco treatment, and is currently under assessment in clinical trials.

Differences in polygenic score distributions in European ancestry populations: implications for breast cancer risk prediction

The 313-variant polygenic risk score (PRS313) provides a promising tool for breast cancer risk prediction. However, evaluation of the PRS313 across different European populations which could influence risk estimation has not been performed. Here, we explored the distribution of PRS313 across European populations using genotype data from 94,072 females without breast cancer, of European-ancestry from 21 countries participating in the Breast Cancer Association Consortium (BCAC) and 225,105 female participants from the UK Biobank. The mean PRS313 differed markedly across European countries, being highest in south-eastern Europe and lowest in north-western Europe. Using the overall European PRS313 distribution to categorise individuals leads to overestimation and underestimation of risk in some individuals from south-eastern and north-western countries, respectively. Adjustment for principal components explained most of the observed heterogeneity in mean PRS. Country-specific PRS distributions may be used to calibrate risk categories in individuals from different countries.

The Potential of Genetics in Identifying Women at Lower Risk of Breast Cancer

Importance

Genetic information is not being used to identify women at lower risk of breast cancer or other diseases in clinical practice. With the new US Preventive Services Task Force guidelines lowering the age for mammogram screening for all, there is a potential benefit in identifying women at lower risk of disease who may defer the start of mammographic screening. This genetic risk-based approach would help mitigate overscreening, associated costs, and anxiety.

Objective

To assess breast cancer incidence and age of onset among women at low genetic risk compared with women at average risk and evaluate the potential to delay mammography on the basis of genetic risk stratification.

Design, Setting, and Participants

This retrospective case-control study included 25 591 women from the Healthy Nevada Project sequenced by Helix between 2018 and 2022. Data extracted from electronic health records at the end of 2022 (mean length of electronic health record available was 12 years) were used for the analysis in 2023.

Main Outcomes and Measures

Breast cancer diagnosis was identified from electronic health records. Classification to the low-risk genetic group required (1) the absence of pathogenic variants or a variant of uncertain significance in BRCA1, BRCA2, PALB2, ATM, or CHEK2, and (2) a low polygenic risk score (bottom 10%) using a 313-single-nucleotide variant model.

Results

Of 25 591 women in the study (mean [SD] age was 53.8 [16.9] years), 2338 women (9.1%) were classified as having low risk for breast cancer; 410 women (1.6%) were classified as high risk; and 22 843 women (89.3%) as average risk. There was a significant reduction in breast cancer diagnosis among the low-risk group (hazard ratio, 0.39; 95% CI, 0.29-0.52; P < .001). By 45 years of age, 0.69% of women in the average-risk group were diagnosed with breast cancer, whereas women in the low-risk group reached this rate at 51 years. By 50 years of age, 1.41% of those in the average-risk group were diagnosed with breast cancer, whereas those in the low-risk group reached this rate at age 58 years. These findings suggest that deferring mammogram screening by 5 to 10 years for women at low risk of breast cancer aligns with new draft recommendations.

Conclusions and Relevance

The findings of this retrospective case-control study underscore the value of genetics in individualizing the onset of breast cancer screening. Improving breast cancer risk stratification by implementing both high-risk and low-risk strategies in screening can refine preventive measures and optimize health care resource allocation.

Highly parameterized polygenic scores tend to overfit to population stratification via random effects

Polygenic scores (PGSs), increasingly used in clinical settings, frequently include many genetic variants, with performance typically peaking at thousands of variants. Such highly parameterized PGSs often include variants that do not pass a genome-wide significance threshold. We propose a mathematical perspective that renders the effects of many of these non-significant variants random rather than causal, with the randomness capturing population structure. We devise methods to assess variant effect randomness and population stratification bias. Applying these methods to 141 traits from the UK Biobank, we find that, for many PGSs, the effects of non-significant variants are considerably random, with the extent of randomness associated with the degree of overfitting to population structure of the discovery cohort. Our findings explain why highly parameterized PGSs simultaneously have superior cohort-specific performance and limited generalizability, suggesting the critical need for variant randomness tests in PGS evaluation. Supporting code and a dashboard are available at https://github.com/songlab-cal/StratPGS.

Multi-Omic Blood Biomarkers as Dynamic Risk Predictors in Late-Onset Alzheimer's Disease

Late-onset Alzheimer's disease is the leading cause of dementia worldwide, accounting for a growing burden of morbidity and mortality. Diagnosing Alzheimer's disease before symptoms are established is clinically challenging, but would provide therapeutic windows for disease-modifying interventions. Blood biomarkers, including genetics, proteins and metabolites, are emerging as powerful predictors of Alzheimer's disease at various timepoints within the disease course, including at the preclinical stage. In this review, we discuss recent advances in such blood biomarkers for determining disease risk. We highlight how leveraging polygenic risk scores, based on genome-wide association studies, can help stratify individuals along their risk profile. We summarize studies analyzing protein biomarkers, as well as report on recent proteomic- and metabolomic-based prediction models. Finally, we discuss how a combination of multi-omic blood biomarkers can potentially be used in memory clinics for diagnosis and to assess the dynamic risk an individual has for developing Alzheimer's disease dementia.

Genome-wide Polygenic Risk Scores Predict Risk of Glioma and Molecular Subtypes

Background

Polygenic risk scores (PRS) aggregate the contribution of many risk variants to provide a personalized genetic susceptibility profile. Since sample sizes of glioma genome-wide association studies (GWAS) remain modest, there is a need to find efficient ways of capturing genetic risk factors using available germline data.

Methods

We developed a novel PRS (PRS-CS) that uses continuous shrinkage priors to model the joint effects of over 1 million polymorphisms on disease risk and compared it to an approach (PRS-CT) that selects a limited set of independent variants that reach genome-wide significance (P<5×10-8). PRS models were trained using GWAS results stratified by histological (10,346 cases, 14,687 controls) and molecular subtype (2,632 cases, 2,445 controls), and validated in two independent cohorts.

Results

PRS-CS was consistently more predictive than PRS-CT across glioma subtypes with an average increase in explained variance (R2) of 21%. Improvements were particularly pronounced for glioblastoma tumors, with PRS-CS yielding larger effect sizes (odds ratio (OR)=1.93, P=2.0×10-54 vs. OR=1.83, P=9.4×10-50) and higher explained variance (R2=2.82% vs. R2=2.56%). Individuals in the 95th percentile of the PRS-CS distribution had a 3-fold higher lifetime absolute risk of IDH mutant (0.63%) and IDH wildtype (0.76%) glioma relative to individuals with average PRS. PRS-CS also showed high classification accuracy for IDH mutation status among cases (AUC=0.895).

Conclusions

Our novel genome-wide PRS may improve the identification of high-risk individuals and help distinguish between prognostic glioma subtypes, increasing the potential clinical utility of germline genetics in glioma patient management.

Multi-ancestry genetic analysis of gene regulation in coronary arteries prioritizes disease risk loci

Genome-wide association studies (GWASs) have identified hundreds of risk loci for coronary artery disease (CAD). However, non-European populations are underrepresented in GWASs, and the causal gene-regulatory mechanisms of these risk loci during atherosclerosis remain unclear. We incorporated local ancestry and haplotypes to identify quantitative trait loci for expression (eQTLs) and splicing (sQTLs) in coronary arteries from 138 ancestrally diverse Americans. Of 2,132 eQTL-associated genes (eGenes), 47% were previously unreported in coronary artery; 19% exhibited cell-type-specific expression. Colocalization revealed subgroups of eGenes unique to CAD and blood pressure GWAS. Fine-mapping highlighted additional eGenes, including TBX20 and IL5. We also identified sQTLs for 1,690 genes, among which TOR1AIP1 and ULK3 sQTLs demonstrated the importance of evaluating splicing to accurately identify disease-relevant isoform expression. Our work provides a patient-derived coronary artery eQTL resource and exemplifies the need for diverse study populations and multifaceted approaches to characterize gene regulation in disease processes.

Machine Learning to Advance Human Genome-Wide Association Studies

Machine learning, including deep learning, reinforcement learning, and generative artificial intelligence are revolutionising every area of our lives when data are made available. With the help of these methods, we can decipher information from larger datasets while addressing the complex nature of biological systems in a more efficient way. Although machine learning methods have been introduced to human genetic epidemiological research as early as 2004, those were never used to their full capacity. In this review, we outline some of the main applications of machine learning to assigning human genetic loci to health outcomes. We summarise widely used methods and discuss their advantages and challenges. We also identify several tools, such as Combi, GenNet, and GMSTool, specifically designed to integrate these methods for hypothesis-free analysis of genetic variation data. We elaborate on the additional value and limitations of these tools from a geneticist's perspective. Finally, we discuss the fast-moving field of foundation models and large multi-modal omics biobank initiatives.

Leveraging genetic correlations and multiple populations to improve genetic risk prediction for non-European populations

The disparity in genetic risk prediction accuracy between European and non-European individuals highlights a critical challenge in health inequality. To bridge this gap, we introduce JointPRS, a novel method that models multiple populations jointly to improve genetic risk predictions for non-European individuals. JointPRS has three key features. First, it encompasses all diverse populations to improve prediction accuracy, rather than relying solely on the target population with a singular auxiliary European group. Second, it autonomously estimates and leverages chromosome-wise cross-population genetic correlations to infer the effect sizes of genetic variants. Lastly, it provides an auto version that has comparable performance to the tuning version to accommodate the situation with no validation dataset. Through extensive simulations and real data applications to 22 quantitative traits and four binary traits in East Asian populations, nine quantitative traits and one binary trait in African populations, and four quantitative traits in South Asian populations, we demonstrate that JointPRS outperforms state-of-art methods, improving the prediction accuracy for both quantitative and binary traits in non-European populations.

The Application of Genetic Risk Scores in Rheumatic Diseases: A Perspective

Modest effect sizes have limited the clinical applicability of genetic associations with rheumatic diseases. Genetic risk scores (GRSs) have emerged as a promising solution to translate genetics into useful tools. In this review, we provide an overview of the recent literature on GRSs in rheumatic diseases. We describe six categories for which GRSs are used: (a) disease (outcome) prediction, (b) genetic commonalities between diseases, (c) disease differentiation, (d) interplay between genetics and environmental factors, (e) heritability and transferability, and (f) detecting causal relationships between traits. In our review of the literature, we identified current lacunas and opportunities for future work. First, the shortage of non-European genetic data restricts the application of many GRSs to European populations. Next, many GRSs are tested in settings enriched for cases that limit the transferability to real life. If intended for clinical application, GRSs are ideally tested in the relevant setting. Finally, there is much to elucidate regarding the co-occurrence of clinical traits to identify shared causal paths and elucidate relationships between the diseases. GRSs are useful instruments for this. Overall, the ever-continuing research on GRSs gives a hopeful outlook into the future of GRSs and indicates significant progress in their potential applications.

Admix-kit: An Integrated Toolkit and Pipeline for Genetic Analyses of Admixed Populations

Admixed populations, with their unique and diverse genetic backgrounds, are often underrepresented in genetic studies. This oversight not only limits our understanding but also exacerbates existing health disparities. One major barrier has been the lack of efficient tools tailored for the special challenges of genetic study of admixed populations. Here, we present admix-kit, an integrated toolkit and pipeline for genetic analyses of admixed populations. Admix-kit implements a suite of methods to facilitate genotype and phenotype simulation, association testing, genetic architecture inference, and polygenic scoring in admixed populations.

Ensembled best subset selection using summary statistics for polygenic risk prediction

Polygenic risk scores (PRS) enhance population risk stratification and advance personalized medicine, yet existing methods face a tradeoff between predictive power and computational efficiency. We introduce ALL-Sum, a fast and scalable PRS method that combines an efficient summary statistic-based L 0 L 2 penalized regression algorithm with an ensembling step that aggregates estimates from different tuning parameters for improved prediction performance. In extensive large-scale simulations across a wide range of polygenicity and genome-wide association studies (GWAS) sample sizes, ALL-Sum consistently outperforms popular alternative methods in terms of prediction accuracy, runtime, and memory usage. We analyze 27 published GWAS summary statistics for 11 complex traits from 9 reputable data sources, including the Global Lipids Genetics Consortium, Breast Cancer Association Consortium, and FinnGen, evaluated using individual-level UKBB data. ALL-Sum achieves the highest accuracy for most traits, particularly for GWAS with large sample sizes. We provide ALL-Sum as a user-friendly command-line software with pre-computed reference data for streamlined user-end analysis.

Evaluating and Improving Health Equity and Fairness of Polygenic Scores

Polygenic scores (PGS) are quantitative metrics for predicting phenotypic values, such as human height or disease status. Some PGS methods require only summary statistics of a relevant genome-wide association study (GWAS) for their score. One such method is Lassosum, which inherits the model selection advantages of Lasso to select a meaningful subset of the GWAS single nucleotide polymorphisms as predictors from their association statistics. However, even efficient scores like Lassosum, when derived from European-based GWAS, are poor predictors of phenotype for subjects of non-European ancestry; that is, they have limited portability to other ancestries. To increase the portability of Lassosum, when GWAS information and estimates of linkage disequilibrium are available for both ancestries, we propose Joint-Lassosum. In the simulation settings we explore, Joint-Lassosum provides more accurate PGS compared with other methods, especially when measured in terms of fairness. Like all PGS methods, Joint-Lassosum requires selection of predictors, which are determined by data-driven tuning parameters. We describe a new approach to selecting tuning parameters and note its relevance for model selection for any PGS. We also draw connections to the literature on algorithmic fairness and discuss how Joint-Lassosum can help mitigate fairness-related harms that might result from the use of PGS scores in clinical settings. While no PGS method is likely to be universally portable, due to the diversity of human populations and unequal information content of GWAS for different ancestries, Joint-Lassosum is an effective approach for enhancing portability and reducing predictive bias.

Genome-wide association study and genomic risk prediction of age-related macular degeneration in Israel

Purpose

The risk of developing age-related macular degeneration(AMD) is influenced by genetic background. In 2016, International AMD Genomics Consortium(IAMDGC) identified 52 risk variants in 34 loci, and a polygenic risk score(PRS) based on these variants was associated with AMD. The Israeli population has a unique genetic composition: Ashkenazi Jewish(AJ), Jewish non-Ashkenazi, and Arab sub-populations. We aimed to perform a genome-wide association study(GWAS) for AMD in Israel, and to evaluate PRSs for AMD.

Methods

For our discovery set, we recruited 403 AMD patients and 256 controls at Hadassah Medical Center. We genotyped all individuals via custom exome chip. We imputed non-typed variants using cosmopolitan and AJ reference panels. We recruited additional 155 cases and 69 controls for validation. To evaluate predictive power of PRSs for AMD, we used IAMDGC summary statistics excluding our study and developed PRSs via either clumping/thresholding or LDpred2.

Results

In our discovery set, 31/34 loci previously reported by the IAMDGC were AMD associated with P<0.05. Of those, all effects were directionally consistent with the IAMDGC and 11 loci had a p-value under Bonferroni-corrected threshold(0.05/34=0.0015). At a threshold of 5x10 -5 , we discovered four suggestive associations in FAM189A1 , IGDCC4 , C7orf50 , and CNTNAP4 . However, only the FAM189A1 variant was AMD associated in the replication cohort after Bonferroni-correction. A prediction model including LDpred2-based PRS and other covariates had an AUC of 0.82(95%CI:0.79-0.85) and performed better than a covariates-only model(P=5.1x10 -9 ).

Conclusions

Previously reported AMD-associated loci were nominally associated with AMD in Israel. A PRS developed based on a large international study is predictive in Israeli populations.