Polygenic risk score as a possible tool for identifying familial monogenic causes of complex diseases

Polygenic embryo screening: quo vadis?

Recently, the use of polygenic risk scores in embryo screening (PGT-P) has been introduced on the premise of reducing polygenic disease risk through embryo selection. However, it has been met with extensive critique: considered "technology-driven" rather than "evidence-based", concerns exist about its validity, utility, ethics, and societal effects. Its scientific foundations and criticisms thus need to be carefully considered. However, seeing as PGT-P is already offered in some settings, further questions need to be addressed, in order to give due diligence to various aspects of PGT-P. By examining the complexities of clinical introduction of PGT-P, we discuss whether PGT-P could be responsibly implemented in the first place, what elements need to be addressed if PGT-P is clinically implemented, and subsequently how counselling and decision-making of its users could be envisaged. By dissecting these elements, we provide an overview of important practical questions of PGT-P and emphasize elements of PGT-P that we think have yet to be given sufficient attention. These questions and elements are for example related to the potential target group, scope, and decision-making possibilities of PGT-P. The aspects we raise are crucial to consider by the scientific community and policy makers for the development of guidelines and/or an ethical framework for PGT-P.

Genetic modifiers of rare variants in monogenic developmental disorder loci

Rare damaging variants in a large number of genes are known to cause monogenic developmental disorders (DDs) and have also been shown to cause milder subclinical phenotypes in population cohorts. Here, we show that carrying multiple (2-5) rare damaging variants across 599 dominant DD genes has an additive adverse effect on numerous cognitive and socioeconomic traits in UK Biobank, which can be partially counterbalanced by a higher educational attainment polygenic score (EA-PGS). Phenotypic deviators from expected EA-PGS could be partly explained by the enrichment or depletion of rare DD variants. Among carriers of rare DD variants, those with a DD-related clinical diagnosis had a substantially lower EA-PGS and more severe phenotype than those without a clinical diagnosis. Our results suggest that the overall burden of both rare and common variants can modify the expressivity of a phenotype, which may then influence whether an individual reaches the threshold for clinical disease.

Polygenic Risk in Families With Spontaneous Coronary Artery Dissection

Importance

Spontaneous coronary artery dissection (SCAD) is a poorly understood cause of acute coronary syndrome that predominantly affects women. Evidence to date suggests a complex genetic architecture, while a family history is reported for a minority of cases.

Objective

To determine the contribution of rare and common genetic variants to SCAD risk in familial cases, the latter via the comparison of a polygenic risk score (PRS) with those with sporadic SCAD and healthy controls.

Design, Setting, and Participants

This genetic association study analyzed families with SCAD, individuals with sporadic SCAD, and healthy controls. Genotyping was undertaken for all participants. Participants were recruited between 2017 and 2021. A PRS for SCAD was calculated for all participants. The presence of rare variants in genes associated with connective tissue disorders (CTD) was also assessed. Individuals with SCAD were recruited via social media or from a single medical center. A previously published control database of older healthy individuals was used. Data were analyzed from January 2022 to October 2023.

Exposures

PRS for SCAD comprised of 7 single-nucleotide variants.

Main Outcomes and Measures

Disease status (familial SCAD, sporadic SCAD, or healthy control) associated with PRS.

Results

A total of 13 families with SCAD (27 affected and 12 unaffected individuals), 173 individuals with sporadic SCAD, and 1127 healthy controls were included. A total of 188 individuals with SCAD (94.0%) were female, including 25 of 27 with familial SCAD and 163 of 173 with sporadic SCAD; of 12 unaffected individuals from families with SCAD, 6 (50%) were female; and of 1127 healthy controls, 672 (59.6%) were female. Compared with healthy controls, the odds of being an affected family member or having sporadic SCAD was significantly associated with a SCAD PRS (where the odds ratio [OR] represents an increase in odds per 1-SD increase in PRS) (affected family member: OR, 2.14; 95% CI, 1.78-2.50; adjusted P = 1.96 × 10-4; sporadic SCAD: OR, 1.63; 95% CI, 1.37-1.89; adjusted P = 5.69 × 10-4). This association was not seen for unaffected family members (OR, 1.03; 95% CI, 0.46-1.61; adjusted P = .91) compared with controls. Further, those with familial SCAD were overrepresented in the top quintile of the control PRS distribution (OR, 3.70; 95% CI, 2.93-4.47; adjusted P = .001); those with sporadic SCAD showed a similar pattern (OR, 2.51; 95% CI, 1.98-3.04; adjusted P = .001). Affected individuals within a family did not share any rare deleterious variants in CTD-associated genes.

Conclusions and Relevance

Extreme aggregation of common genetic risk appears to play a significant role in familial clustering of SCAD as well as in sporadic case predisposition, although further study is required.

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.

Identifying Rare Genetic Determinants for Improved Polygenic Risk Prediction of Bone Mineral Density and Fracture Risk

Osteoporosis and fractures severely impact the elderly population. Polygenic risk scores for bone mineral density have demonstrated potential clinical utility. However, the value of rare genetic determinants in risk prediction has not been assessed. With whole-exome sequencing data from 436,824 UK Biobank participants, we assigned White British ancestry individuals into a training data set (n = 317,434) and a test data set (n = 74,825). In the training data set, we developed a common variant-based polygenic risk score for heel ultrasound speed of sound (SOS). Next, we performed burden testing to identify genes harboring rare determinants of bone mineral density, targeting influential rare variants with predicted high deleteriousness. We constructed a genetic risk score, called ggSOS, to incorporate influential rare variants in significant gene burden masks into the common variant-based polygenic risk score. We assessed the predictive performance of ggSOS in the White British test data set, as well as in populations of non-White British European (n = 18,885), African (n = 7165), East Asian (n = 2236), South Asian (n = 9829), and other admixed (n = 1481) ancestries. Twelve genes in pivotal regulatory pathways of bone homeostasis harbored influential rare variants associated with SOS (p < 5.5 × 10-7 ), including AHNAK, BMP5, CYP19A1, FAM20A, FBXW5, KDM5B, KREMEN1, LGR4, LRP5, SMAD6, SOST, and WNT1. Among 4013 (5.4%) individuals in the test data set carrying these variants, a one standard deviation decrease in ggSOS was associated with 1.35-fold (95% confidence interval [CI] 1.16-1.57) increased hazard of major osteoporotic fracture. However, compared with a common variant-based polygenic risk score (C-index = 0.641), ggSOS had only marginally improved prediction accuracy in identifying at-risk individuals (C-index = 0.644), with overlapping confidence intervals. Similarly, ggSOS did not demonstrate substantially improved predictive performance in non-European ancestry populations. In summary, modeling the effects of rare genetic determinants may assist polygenic prediction of fracture risk among carriers of influential rare variants. Nonetheless, improved clinical utility is not guaranteed for population-level risk screening. © 2023 The Authors. Journal of Bone and Mineral Research published by Wiley Periodicals LLC on behalf of American Society for Bone and Mineral Research (ASBMR).

Polygenic Risk Score and Rare Variant Burden Identified by Targeted Sequencing in a Group of Patients with Pigment Epithelial Detachment in Age-Related Macular Degeneration

A subset of ophthalmic imaging examination results from 334 patients were subjected to reanalysis to identify a specific group of patients with pigment epithelial detachment (PED) in at least one eye. Overall, we found a subgroup of 47 patients manifesting PED and studied their genotypes in comparison to those of patients with age-related macular degeneration without PED and healthy controls. We established a polygenic risk score that allowed the explanation of 16.3% of the variation within the disease. The highest predictive value was achieved for a model consisting of six non-coding variants: rs760306 (BEST1), rs148662546 (BEST1), rs11569560 (C3), rs74600252 (GUCA1B), rs2240688 (PROM1), and rs185507582 (TCF4). The risk of PED occurrence was found to be the highest in the first tercile, showing a 7.89-fold higher risk compared to the third tercile for AMD without PED (95% CI: 2.87; 21.71, p < 0.001) and a 7.22-fold higher risk compared to the healthy controls (95% CI: 2.60; 20.06, p < 0.001). In addition, we focused on rare variants in targeted genes. The rare variants' burden was compared among the groups, but no statistical significance was observed in the number of rare variants, predicted functional effects, or pathogenicity classification.

Development of risk prediction models for depression combining genetic and early life risk factors

Background

Both genetic and early life risk factors play important roles in the pathogenesis and progression of adult depression. However, the interplay between these risk factors and their added value to risk prediction models have not been fully elucidated.

Methods

Leveraging a meta-analysis of major depressive disorder genome-wide association studies (N = 45,591 cases and 97,674 controls), we developed and optimized a polygenic risk score for depression using LDpred in a model selection dataset from the UK Biobank (N = 130,092 European ancestry individuals). In a UK Biobank test dataset (N = 278,730 European ancestry individuals), we tested whether the polygenic risk score and early life risk factors were associated with each other and compared their associations with depression phenotypes. Finally, we conducted joint predictive modeling to combine this polygenic risk score with early life risk factors by stepwise regression, and assessed the model performance in identifying individuals at high risk of depression.

Results

In the UK Biobank test dataset, the polygenic risk score for depression was moderately associated with multiple early life risk factors. For instance, a one standard deviation increase in the polygenic risk score was associated with 1.16-fold increased odds of frequent domestic violence (95% CI: 1.14-1.19) and 1.09-fold increased odds of not having access to medical care as a child (95% CI: 1.05-1.14). However, the polygenic risk score was more strongly associated with depression phenotypes than most early life risk factors. A joint predictive model integrating the polygenic risk score, early life risk factors, age and sex achieved an AUROC of 0.6766 for predicting strictly defined major depressive disorder, while a model without the polygenic risk score and a model without any early life risk factors had an AUROC of 0.6593 and 0.6318, respectively.

Conclusion

We have developed a polygenic risk score to partly capture the genetic liability to depression. Although genetic and early life risk factors can be correlated, joint predictive models improved risk stratification despite limited improvement in magnitude, and may be explored as tools to better identify individuals at high risk of depression.

Non-BRCA1/BRCA2 high-risk familial breast cancers are not associated with a high prevalence of BRCAness

BACKGROUND

Familial breast cancer is in most cases unexplained due to the lack of identifiable pathogenic variants in the BRCA1 and BRCA2 genes. The somatic mutational landscape and in particular the extent of BRCA-like tumour features (BRCAness) in these familial breast cancers where germline BRCA1 or BRCA2 mutations have not been identified is to a large extent unknown.

METHODS

We performed whole-genome sequencing on matched tumour and normal samples from high-risk non-BRCA1/BRCA2 breast cancer families to understand the germline and somatic mutational landscape and mutational signatures. We measured BRCAness using HRDetect. As a comparator, we also analysed samples from BRCA1 and BRCA2 germline mutation carriers.

RESULTS

We noted for non-BRCA1/BRCA2 tumours, only a small proportion displayed high HRDetect scores and were characterized by concomitant promoter hypermethylation or in one case a RAD51D splice variant previously reported as having unknown significance to potentially explain their BRCAness. Another small proportion showed no features of BRCAness but had mutationally active tumours. The remaining tumours lacked features of BRCAness and were mutationally quiescent.

CONCLUSIONS

A limited fraction of high-risk familial non-BRCA1/BRCA2 breast cancer patients is expected to benefit from treatment strategies against homologue repair deficient cancer cells.

A Machine Learning Model to Aid Detection of Familial Hypercholesterolemia

Background

People with monogenic familial hypercholesterolemia (FH) are at an increased risk of premature coronary heart disease and death. With a prevalence of 1:250, FH is relatively common; but currently there is no population screening strategy in place and most carriers are identified late in life, delaying timely and cost-effective interventions.

Objectives

The purpose of this study was to derive an algorithm to identify people with suspected monogenic FH for subsequent confirmatory genomic testing and cascade screening.

Methods

A least absolute shrinkage and selection operator logistic regression model was used to identify predictors that accurately identified people with FH in 139,779 unrelated participants of the UK Biobank. Candidate predictors included information on medical and family history, anthropometric measures, blood biomarkers, and a low-density lipoprotein cholesterol (LDL-C) polygenic score (PGS). Model derivation and evaluation were performed in independent training and testing data.

Results

A total of 488 FH variant carriers were identified using whole-exome sequencing of the low-density lipoprotein receptor, apolipoprotein B, apolipoprotein E, proprotein convertase subtilisin/kexin type 9 genes. A 14-variable algorithm for FH was derived, with an area under the curve of 0.77 (95% CI: 0.71-0.83), where the top 5 most important variables included triglyceride, LDL-C, apolipoprotein A1 concentrations, self-reported statin use, and LDL-C PGS. Excluding the PGS as a candidate feature resulted in a 9-variable model with a comparable area under the curve: 0.76 (95% CI: 0.71-0.82). Both multivariable models (w/wo the PGS) outperformed screening-prioritization based on LDL-C adjusted for statin use.

Conclusions

Detecting individuals with FH can be improved by considering additional predictors. This would reduce the sequencing burden in a 2-stage population screening strategy for FH.

Genetic determinants of polygenic prediction accuracy within a population

Genomic risk prediction is on the emerging path toward personalized medicine. However, the accuracy of polygenic prediction varies strongly in different individuals. Based on up to 352,277 European ancestry participants in the UK Biobank, we constructed polygenic risk scores for 15 physiological and biochemical quantitative traits. We identified a total of 185 polygenic prediction variability quantitative trait loci for 11 traits by Levene's test among 254,376 unrelated individuals. We validated the effects of prediction variability quantitative trait loci using an independent test set of 58,927 individuals. For instance, a score aggregating 51 prediction variability quantitative trait locus variants for triglycerides had the strongest Spearman correlation of 0.185 (P-value <1.0 × 10-300) with the squared prediction errors. We found a strong enrichment of complex genetic effects conferred by prediction variability quantitative trait loci compared to risk loci identified in genome-wide association studies, including 89 prediction variability quantitative trait loci exhibiting dominance effects. Incorporation of dominance effects into polygenic risk scores significantly improved polygenic prediction for triglycerides, low-density lipoprotein cholesterol, vitamin D, and platelet. In conclusion, we have discovered and profiled genetic determinants of polygenic prediction variability for 11 quantitative biomarkers. These findings may assist interpretation of genomic risk prediction in various contexts and encourage novel approaches for constructing polygenic risk scores with complex genetic effects.

Capturing additional genetic risk from family history for improved polygenic risk prediction

Family history of complex traits may reflect transmitted rare pathogenic variants, intra-familial shared exposures to environmental and lifestyle factors, as well as a common genetic predisposition. We developed a latent factor model to quantify trait heritability in excess of that captured by a common variant-based polygenic risk score, but inferable from family history. For 941 children in the Avon Longitudinal Study of Parents and Children cohort, a joint predictor combining a polygenic risk score for height and mid-parental height was able to explain ^~55% of the total variance in sex-adjusted adult height z-scores, close to the estimated heritability. Marginal yet consistent risk prediction improvements were also achieved among ^~400,000 European ancestry participants for 11 complex diseases in the UK Biobank. Our work showcases a paradigm for risk calculation, and supports incorporation of family history into polygenic risk score-based genetic risk prediction models.

Common risk variants for epilepsy are enriched in families previously targeted for rare monogenic variant discovery

BACKGROUND

The epilepsies are highly heritable conditions that commonly follow complex inheritance. While monogenic causes have been identified in rare familial epilepsies, most familial epilepsies remain unsolved. We aimed to determine (1) whether common genetic variation contributes to familial epilepsy risk, and (2) whether that genetic risk is enriched in familial compared with non-familial (sporadic) epilepsies.

METHODS

Using common variants derived from the largest epilepsy genome-wide association study, we calculated polygenic risk scores (PRS) for patients with familial epilepsy (n = 1,818 from 1,181 families), their unaffected relatives (n = 771), sporadic patients (n = 1,182), and population controls (n = 15,929). We also calculated separate PRS for genetic generalised epilepsy (GGE) and focal epilepsy. Statistical analyses used mixed-effects regression models to account for familial relatedness, sex, and ancestry.

FINDINGS

Patients with familial epilepsies had higher epilepsy PRS compared to population controls (OR 1·20, p_adj = 5×10^-9), sporadic patients (OR 1·11, p_adj = 0.008), and their own unaffected relatives (OR 1·12, p_adj = 0.01). The top 1% of the PRS distribution was enriched 3.8-fold for individuals with familial epilepsy when compared to the lowest decile (p_adj = 5×10^-11). Familial PRS enrichment was consistent across epilepsy type; overall, polygenic risk was greatest for the GGE clinical group. There was no significant PRS difference in familial cases with established rare variant genetic etiologies compared to unsolved familial cases.

INTERPRETATION

The aggregate effects of common genetic variants, measured as polygenic risk scores, play an important role in explaining why some families develop epilepsy, why specific family members are affected while their relatives are not, and why families manifest specific epilepsy types. Polygenic risk contributes to the complex inheritance of the epilepsies, including in individuals with a known genetic etiology.

FUNDING

National Health and Medical Research Council of Australia, National Institutes of Health, American Academy of Neurology, Thomas B and Jeannette E Laws McCabe Fund, Mirowski Family Foundation.