Including diverse and admixed populations in genetic epidemiology research

Genetic distance and ancestry proportion modify the association between maternal genetic risk score of type 2 diabetes and fetal growth

BACKGROUND

Maternal genetic risk of type 2 diabetes (T2D) has been associated with fetal growth, but the influence of genetic ancestry is not yet fully understood. We aimed to investigate the influence of genetic distance (GD) and genetic ancestry proportion (GAP) on the association of maternal genetic risk score of T2D (GRST2D) with fetal weight and birthweight.

METHODS

Multi-ancestral pregnant women (n = 1,837) from the NICHD Fetal Growth Studies - Singletons cohort were included in the current analyses. Fetal weight (in grams, g) was estimated from ultrasound measurements of fetal biometry, and birthweight (g) was measured at delivery. GRST2D was calculated using T2D-associated variants identified in the latest trans-ancestral genome-wide association study and was categorized into quartiles. GD and GAP were estimated using genotype data of four reference populations. GD was categorized into closest, middle, and farthest tertiles, and GAP was categorized as highest, medium, and lowest. Linear regression analyses were performed to test the association of GRST2D with fetal weight and birthweight, adjusted for covariates, in each GD and GAP category.

RESULTS

Among women with the closest GD from African and Amerindigenous ancestries, the fourth and third GRST2D quartile was significantly associated with 5.18 to 7.48 g (weeks 17-20) and 6.83 to 25.44 g (weeks 19-27) larger fetal weight compared to the first quartile, respectively. Among women with middle GD from European ancestry, the fourth GRST2D quartile was significantly associated with 5.73 to 21.21 g (weeks 18-26) larger fetal weight. Furthermore, among women with middle GD from European and African ancestries, the fourth and second GRST2D quartiles were significantly associated with 117.04 g (95% CI = 23.88-210.20, p = 0.014) and 95.05 g (95% CI = 4.73-185.36, p = 0.039) larger birthweight compared to the first quartile, respectively. The absence of significant association among women with the closest GD from East Asian ancestry was complemented by a positive significant association among women with the highest East Asian GAP.

CONCLUSIONS

The association between maternal GRST2D and fetal growth began in early-second trimester and was influenced by GD and GAP. The results suggest the use of genetic GD and GAP could improve the generalizability of GRS.

The combined effect of lifestyle factors and polygenic scores on age at onset in Parkinson's disease

The objective of this study was to investigate the association between a Parkinson's disease (PD)-specific polygenic score (PGS) and protective lifestyle factors on age at onset (AAO) in PD. We included data from 4367 patients with idiopathic PD, 159 patients with GBA1-PD, and 3090 healthy controls of European ancestry from AMP-PD, PPMI, and Fox Insight cohorts. The association between PGS and lifestyle factors on AAO was assessed with linear and Cox proportional hazards models. The PGS showed a negative association with AAO (β = - 1.07, p = 6 × 10-7) in patients with idiopathic PD. The use of one, two, or three of the protective lifestyle factors showed a reduction in the hazard ratio by 21% (p = 0.0001), 44% (p < 2 × 10-16), and 55% (p < 2 × 10-16), compared to no use. An additive effect of aspirin (β = 7.62, p = 9 × 10-7) and PGS (β = - 1.58, p = 0.0149) was found for AAO without an interaction (p = 0.9993) in the linear regressions, and similar effects were seen for tobacco. In contrast, no association between aspirin intake and AAO was found in GBA1-PD (p > 0.05). In our cohort, coffee, tobacco, aspirin, and PGS are independent predictors of PD AAO. Additionally, lifestyle factors seem to have a greater influence on AAO than common genetic risk variants with aspirin presenting the largest effect.

Recent advances in polygenic scores: translation, equitability, methods and FAIR tools

Polygenic scores (PGS) can be used for risk stratification by quantifying individuals' genetic predisposition to disease, and many potentially clinically useful applications have been proposed. Here, we review the latest potential benefits of PGS in the clinic and challenges to implementation. PGS could augment risk stratification through combined use with traditional risk factors (demographics, disease-specific risk factors, family history, etc.), to support diagnostic pathways, to predict groups with therapeutic benefits, and to increase the efficiency of clinical trials. However, there exist challenges to maximizing the clinical utility of PGS, including FAIR (Findable, Accessible, Interoperable, and Reusable) use and standardized sharing of the genomic data needed to develop and recalculate PGS, the equitable performance of PGS across populations and ancestries, the generation of robust and reproducible PGS calculations, and the responsible communication and interpretation of results. We outline how these challenges may be overcome analytically and with more diverse data as well as highlight sustained community efforts to achieve equitable, impactful, and responsible use of PGS in healthcare.

Power of inclusion: Enhancing polygenic prediction with admixed individuals

Admixed individuals offer unique opportunities for addressing limited transferability in polygenic scores (PGSs), given the substantial trans-ancestry genetic correlation in many complex traits. However, they are rarely considered in PGS training, given the challenges in representing ancestry-matched linkage-disequilibrium reference panels for admixed individuals. Here we present inclusive PGS (iPGS), which captures ancestry-shared genetic effects by finding the exact solution for penalized regression on individual-level data and is thus naturally applicable to admixed individuals. We validate our approach in a simulation study across 33 configurations with varying heritability, polygenicity, and ancestry composition in the training set. When iPGS is applied to n = 237,055 ancestry-diverse individuals in the UK Biobank, it shows the greatest improvements in Africans by 48.9% on average across 60 quantitative traits and up to 50-fold improvements for some traits (neutrophil count, R2 = 0.058) over the baseline model trained on the same number of European individuals. When we allowed iPGS to use n = 284,661 individuals, we observed an average improvement of 60.8% for African, 11.6% for South Asian, 7.3% for non-British White, 4.8% for White British, and 17.8% for the other individuals. We further developed iPGS+refit to jointly model the ancestry-shared and -dependent genetic effects when heterogeneous genetic associations were present. For neutrophil count, for example, iPGS+refit showed the highest predictive performance in the African group (R2 = 0.115), which exceeds the best predictive performance for the White British group (R2 = 0.090 in the iPGS model), even though only 1.49% of individuals used in the iPGS training are of African ancestry. Our results indicate the power of including diverse individuals for developing more equitable PGS models.

Clinical utility of polygenic risk scores: a critical 2023 appraisal

Since their first appearance in the context of schizophrenia and bipolar disorder in 2009, polygenic risk scores (PRSs) have been described for a large number of common complex diseases. However, the clinical utility of PRSs in disease risk assessment or therapeutic decision making is likely limited because PRSs usually only account for the heritable component of a trait and ignore the etiological role of environment and lifestyle. We surveyed the current state of PRSs for various diseases, including breast cancer, diabetes, prostate cancer, coronary artery disease, and Parkinson disease, with an extra focus upon the potential improvement of clinical scores by their combination with PRSs. We observed that the diagnostic and prognostic performance of PRSs alone is consistently low, as expected. Moreover, combining a PRS with a clinical score at best led to moderate improvement of the power of either risk marker. Despite the large number of PRSs reported in the scientific literature, prospective studies of their clinical utility, particularly of the PRS-associated improvement of standard screening or therapeutic procedures, are still rare. In conclusion, the benefit to individual patients or the health care system in general of PRS-based extensions of existing diagnostic or treatment regimens is still difficult to judge.

Interaction of Mitochondrial Polygenic Score and Lifestyle Factors in LRRK2 p.Gly2019Ser Parkinsonism

BACKGROUND

A mitochondrial polygenic score (MGS) is composed of genes related to mitochondrial function and found to be associated with Parkinson's disease (PD) risk.

OBJECTIVE

To investigate the impact of the MGS and lifestyle/environment on age at onset (AAO) in LRRK2 p.Gly2019Ser parkinsonism (LRRK2-PD) and idiopathic PD (iPD).

METHODS

We included N = 486 patients with LRRK2-PD and N = 9259 with iPD from the Accelerating Medicines Partnership® Parkinson's Disease Knowledge Platform (AMP-PD), Fox Insight, and a Tunisian Arab-Berber founder population. Genotyping data were used to perform the MGS analysis. Additionally, lifestyle/environmental data were obtained from the PD Risk Factor Questionnaire (PD-RFQ). Linear regression models were used to assess the relationship between MGS, lifestyle/environment, and AAO.

RESULTS

Our derived MGS was significantly higher in PD cases compared with controls (P = 1.1 × 10-8 ). We observed that higher MGS was significantly associated with earlier AAO in LRRK2-PD (P = 0.047, β = -1.40) and there was the same trend with a smaller effect size in iPD (P = 0.231, β = 0.22). There was a correlation between MGS and AAO in LRRK2-PD patients of European descent (P = 0.049, r = -0.12) that was visibly less pronounced in Tunisians (P = 0.449, r = -0.05). We found that the MGS interacted with caffeinated soda consumption (P = 0.003, β = -5.65) in LRRK2-PD and with tobacco use (P = 0.010, β = 1.32) in iPD. Thus, patients with a high MGS had an earlier AAO only if they consumed caffeinated soda or were non-smokers.

CONCLUSIONS

The MGS was more strongly associated with earlier AAO in LRRK2-PD compared with iPD. Caffeinated soda consumption or tobacco use interacted with MGS to predict AAO. Our study suggests gene-environment interactions as modifiers of AAO in LRRK2-PD. © 2023 The Authors. Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.

The combined effect of lifestyle factors and polygenic scores on age at onset in Parkinson's disease

Objective

To investigate the association between a Parkinson's disease (PD)-specific polygenic score (PGS) and protective lifestyle factors on age at onset (AAO) in PD.

Methods

We included data from 4375 patients with idiopathic PD, 167 patients with GBA1-PD, and 3091 healthy controls of European ancestry from AMP-PD, PPMI, and Fox Insight cohorts. The PGS was calculated based on a previously proposed composition of 1805 variants. The association between PGS and lifestyle factors (i.e., coffee, tobacco, and aspirin) on AAO was assessed with linear and Cox proportional hazards models.

Results

The PGS showed a negative association with AAO (β=-1.07, p=6×10-7). The use of one, two, or three of the protective lifestyle factors showed a reduction in the hazard ratio by 21% (p=0.0001), 45% (p<2×10-16), and 55% (p<2×10-16), respectively, compared to no use. An additive effect of aspirin (β=7.61, p=8×10-7) and PGS (β=-1.63, p=0.0112) was found for AAO without an interaction (p=0.9789) in the linear regressions, and similar effects were seen for tobacco. Aspirin is shown to be a better predictor of AAO (R2=0.1740) compared to coffee and tobacco use (R2=0.0243, R2=0.0295) or the PGS (R2=0.0141). In contrast, no association between aspirin and AAO was found in GBA1-PD (p>0.05).

Interpretation

In our cohort, coffee, tobacco, aspirin, and PGS are independent predictors of PD AAO. Additionally, lifestyle factors seem to have a greater influence on AAO than common genetic risk variants with aspirin presenting the largest effect. External validation of our findings is needed.

Efforts to Grow Genomic Research in Ancestrally Diverse and Admixed Populations

Recent initiatives by the research community to characterize the genomic and molecular landscapes of tumors in ancestrally diverse and admixed populations, including the publication by Ding and colleagues in this issue of Cancer Research, represent important efforts to improve our understanding of the entire spectrum of cancer genomic variation with potential clinical consequences. Ding and colleagues confirmed a similar prevalence of mutations in established breast cancer driver genes including PIK3CA, TP53, GATA3, MAP3K1, CDH1, CBFB, PTEN, and RUNX1 and recurrent amplifications in breast cancer drivers including MYC, FGFR1, CCND1, and ERBB2 in tumors from Hispanic/Latina women as compared with non-Hispanic White women. Importantly, they also identified Catalogue of Somatic Mutations in Cancer (COSMIC) signature 16 in a significant fraction of tumors from Hispanic/Latina women and a novel recurrent amplification on 17q11.2. This study highlights the potential for inclusion of participants from diverse populations to accelerate discoveries and advance equity in genomic medicine, as well as the need for even larger collaborative initiatives. See related article by Ding et al., p. 2600.

The complex genetic architecture of Alzheimer's disease: novel insights and future directions

BACKGROUND

Alzheimer's disease (AD) is a complex multifactorial neurodegenerative disorder and the most common form of dementia. AD is highly heritable, with heritability estimates of ∼70% from twin studies. Progressively larger genome-wide association studies (GWAS) have continued to expand our knowledge of AD/dementia genetic architecture. Until recently these efforts had identified 39 disease susceptibility loci in European ancestry populations.

RECENT DEVELOPMENTS

Two new AD/dementia GWAS have dramatically expanded the sample sizes and the number of disease susceptibility loci. The first increased total sample size to 1,126,563-with an effective sample size of 332,376-by predominantly including new biobank and population-based dementia datasets. The second, expands on an earlier GWAS from the International Genomics of Alzheimer's Project (IGAP) by increasing the number of clinically-defined AD cases/controls in addition to incorporating biobank dementia datasets, resulting in a total sample size to 788,989 and an effective sample size of 382,472. Collectively both GWAS identified 90 independent variants across 75 AD/dementia susceptibility loci, including 42 novel loci. Pathway analyses indicate the susceptibility loci are enriched for genes involved in amyloid plaque and neurofibrillary tangle formation, cholesterol metabolism, endocytosis/phagocytosis, and the innate immune system. Gene prioritization efforts for the novel loci identified 62 candidate causal genes. Many of the candidate genes from known and newly discovered loci play key roles in macrophages and highlight phagocytic clearance of cholesterol-rich brain tissue debris by microglia (efferocytosis) as a core pathogenetic hub and putative therapeutic target for AD. WHERE NEXT?: While GWAS in European ancestry populations have substantially enhanced our understanding of AD genetic architecture, heritability estimates from population based GWAS cohorts are markedly smaller than those from twin studies. While this missing heritability is likely due to a combination of factors, it highlights that our understanding of AD genetic architecture and genetic risk mechanisms remains incomplete. These knowledge gaps result from several underexplored areas in AD research. First, rare variants remain understudied due to methodological issues in identifying them and the cost of generating sufficiently powered whole exome/genome sequencing datasets. Second, sample sizes of non-European ancestry populations in AD GWAS remain small. Third, GWAS of AD neuroimaging and cerebrospinal fluid endophenotypes remains limited due to low compliance and high costs associated with measuring amyloid-β and tau levels and other disease-relevant biomarkers. Studies generating sequencing data, including diverse populations, and incorporating blood-based AD biomarkers are set to substantially improve our knowledge of AD genetic architecture.

Clinical and genetic associations of deep learning-derived cardiac magnetic resonance-based left ventricular mass

Left ventricular mass is a risk marker for cardiovascular events, and may indicate an underlying cardiomyopathy. Cardiac magnetic resonance is the gold-standard for left ventricular mass estimation, but is challenging to obtain at scale. Here, we use deep learning to enable genome-wide association study of cardiac magnetic resonance-derived left ventricular mass indexed to body surface area within 43,230 UK Biobank participants. We identify 12 genome-wide associations (1 known at TTN and 11 novel for left ventricular mass), implicating genes previously associated with cardiac contractility and cardiomyopathy. Cardiac magnetic resonance-derived indexed left ventricular mass is associated with incident dilated and hypertrophic cardiomyopathies, and implantable cardioverter-defibrillator implant. An indexed left ventricular mass polygenic risk score ≥90th percentile is also associated with incident implantable cardioverter-defibrillator implant in separate UK Biobank (hazard ratio 1.22, 95% CI 1.05-1.44) and Mass General Brigham (hazard ratio 1.75, 95% CI 1.12-2.74) samples. Here, we perform a genome-wide association study of cardiac magnetic resonance-derived indexed left ventricular mass to identify 11 novel variants and demonstrate that cardiac magnetic resonance-derived and genetically predicted indexed left ventricular mass are associated with incident cardiomyopathy.