Variable prediction accuracy of polygenic scores within an ancestry group

Fields as diverse as human genetics and sociology are increasingly using polygenic scores based on genome-wide association studies (GWAS) for phenotypic prediction. However, recent work has shown that polygenic scores have limited portability across groups of different genetic ancestries, restricting the contexts in which they can be used reliably and potentially creating serious inequities in future clinical applications. Using the UK Biobank data, we demonstrate that even within a single ancestry group (i.e., when there are negligible differences in linkage disequilibrium or in causal alleles frequencies), the prediction accuracy of polygenic scores can depend on characteristics such as the socio-economic status, age or sex of the individuals in which the GWAS and the prediction were conducted, as well as on the GWAS design. Our findings highlight both the complexities of interpreting polygenic scores and underappreciated obstacles to their broad use.

Whole Exome Sequencing Study of Parkinson Disease and Related Endophenotypes in the Italian Population

Parkinson Disease (PD) is a complex neurodegenerative disorder characterized by large genetic heterogeneity and missing heritability. Since the genetic background of PD can partly vary among ethnicities and neurological scales have been scarcely investigated in a PD setting, we performed an exploratory Whole Exome Sequencing (WES) analysis of 123 PD patients from mainland Italy, investigating scales assessing motor (UPDRS), cognitive (MoCA), and other non-motor symptoms (NMS). We performed variant prioritization, followed by targeted association testing of prioritized variants in 446 PD cases and 211 controls. Then we ran Exome-Wide Association Scans (EWAS) within sequenced PD cases (N = 113), testing both motor and non-motor PD endophenotypes, as well as their associations with Polygenic Risk Scores (PRS) influencing brain subcortical volumes. We identified a variant associated with PD, rs201330591 in GTF2H2 (5q13; alternative T allele: OR [CI] = 8.16[1.08; 61.52], FDR = 0.048), which was not replicated in an independent cohort of European ancestry (1,148 PD cases, 503 controls). In the EWAS, polygenic analyses revealed statistically significant multivariable associations of amygdala- [β(SE) = -0.039(0.013); FDR = 0.039] and caudate-PRS [0.043(0.013); 0.028] with motor symptoms. All subcortical PRSs in a multivariable model notably increased the variance explained in motor (adjusted-R2 = 38.6%), cognitive (32.2%) and other non-motor symptoms (28.9%), compared to baseline models (~20%). Although, the small sample size warrants further replications, these findings suggest shared genetic architecture between PD symptoms and subcortical structures, and provide interesting clues on PD genetic and neuroimaging features.

How Intelligence Affects Fertility 30 Years On: Retherford and Sewell Revisited - With Polygenic Scores and Numbers of Grandchildren

Using newly available polygenic scores for educational attainment and cognitive ability, this paper investigates the possible presence and causes of a negative association between IQ and fertility in the Wisconsin Longitudinal Study sample, an issue that Retherford and Sewell first addressed 30 years ago. The effect of the polygenic score on the sample's reproductive characteristics was indirect: a latent cognitive ability measure, comprised of both educational attainment and IQ, wholly mediated the relationship. Age at first birth mediated the negative effect of cognitive ability on sample fertility, which had a direct (positive) effect on the number of grandchildren. Significantly greater impacts of cognitive ability on the sample's fertility characteristics were found among the female subsample. This indicates that, in this sample, having a genetic disposition toward higher cognitive ability does not directly reduce number of offspring; instead, higher cognitive ability is a risk factor for prolonging reproductive debut, which, especially for women, reduces the fertility window and, thus, the number of children and grandchildren that can be produced. By estimating the effect of the sample's reproductive characteristics on the strength of polygenic selection, it was found that the genetic variance component of IQ should be declining at a rate between -.208 (95% CI [-.020, -.383]) and -.424 (95% CI [-.041, -.766]) points per decade, depending on whether GCTA-GREML or classical behavior genetic estimates of IQ heritability are used to correct for 'missing' heritability.

What your genes can (and can't) tell you about BMI and diabetes

Body mass index (BMI) is commonly used as a proxy for adiposity in epidemiological and public health studies. However, BMI may suffer from issues of misreporting and, because it fluctuates over the life course, its association with morbidities such as diabetes is difficult to measure. We examined the associations between actual BMI, genetic propensity for high BMI, and diabetes to better understand whether a BMI polygenic score (PGS) explained more variation in diabetes than self-reported BMI. We used a sample of non-Hispanic white adults from the longitudinal Health and Retirement Study (1992-2016). Structural equation models were used to determine how much variation in BMI could be explained by a BMI PGS. Then, we used logistic regression models (n = 12,086) to study prevalent diabetes at baseline and Cox regression models (n = 11,129) to examine incident diabetes with up to 24 years of follow-up. We observed that while both actual BMI and the BMI PGS were significantly associated with diabetes, actual BMI had a stronger association than its genetic counterpart and resulted in better model performance. Moreover, actual BMI explained more variation in baseline and incident diabetes than its genetic counterpart which may suggest that actual BMI captures more than just adiposity as intended.

Investigating genetic links between grapheme-colour synaesthesia and neuropsychiatric traits

Synaesthesia is a neurological phenomenon affecting perception, where triggering stimuli (e.g. letters and numbers) elicit unusual secondary sensory experiences (e.g. colours). Family-based studies point to a role for genetic factors in the development of this trait. However, the contributions of common genomic variation to synaesthesia have not yet been investigated. Here, we present the SynGenes cohort, the largest genotyped collection of unrelated people with grapheme-colour synaesthesia (n = 723). Synaesthesia has been associated with a range of other neuropsychological traits, including enhanced memory and mental imagery, as well as greater sensory sensitivity. Motivated by the prior literature on putative trait overlaps, we investigated polygenic scores derived from published genome-wide scans of schizophrenia and autism spectrum disorder (ASD), comparing our SynGenes cohort to 2181 non-synaesthetic controls. We found a very slight association between schizophrenia polygenic scores and synaesthesia (Nagelkerke's R2 = 0.0047, empirical p = 0.0027) and no significant association for scores related to ASD (Nagelkerke's R2 = 0.00092, empirical p = 0.54) or body mass index (R2 = 0.00058, empirical p = 0.60), included as a negative control. As sample sizes for studying common genomic variation continue to increase, genetic investigations of the kind reported here may yield novel insights into the shared biology between synaesthesia and other traits, to complement findings from neuropsychology and brain imaging. This article is part of a discussion meeting issue 'Bridging senses: novel insights from synaesthesia'.

Screening Human Embryos for Polygenic Traits Has Limited Utility

The increasing proportion of variance in human complex traits explained by polygenic scores, along with progress in preimplantation genetic diagnosis, suggests the possibility of screening embryos for traits such as height or cognitive ability. However, the expected outcomes of embryo screening are unclear, which undermines discussion of associated ethical concerns. Here, we use theory, simulations, and real data to evaluate the potential gain of embryo screening, defined as the difference in trait value between the top-scoring embryo and the average embryo. The gain increases very slowly with the number of embryos but more rapidly with the variance explained by the score. Given current technology, the average gain due to screening would be ≈2.5 cm for height and ≈2.5 IQ points for cognitive ability. These mean values are accompanied by wide prediction intervals, and indeed, in large nuclear families, the majority of children top-scoring for height are not the tallest.

A Polygenic Score for Higher Educational Attainment is Associated with Larger Brains

People who score higher on intelligence tests tend to have larger brains. Twin studies suggest the same genetic factors influence both brain size and intelligence. This has led to the hypothesis that genetics influence intelligence partly by contributing to the development of larger brains. We tested this hypothesis using four large imaging genetics studies (combined N = 7965) with polygenic scores derived from a genome-wide association study (GWAS) of educational attainment, a correlate of intelligence. We conducted meta-analysis to test associations among participants' genetics, total brain volume (i.e., brain size), and cognitive test performance. Consistent with previous findings, participants with higher polygenic scores achieved higher scores on cognitive tests, as did participants with larger brains. Participants with higher polygenic scores also had larger brains. We found some evidence that brain size partly mediated associations between participants' education polygenic scores and their cognitive test performance. Effect sizes were larger in the population-based samples than in the convenience-based samples. Recruitment and retention of population-representative samples should be a priority for neuroscience research. Findings suggest promise for studies integrating GWAS discoveries with brain imaging to understand neurobiology linking genetics with cognitive performance.

The Genome-Wide Study of Human Social Behavior and Its Application in Sociology

Recent years have seen a push for the integration of modern genomic methodologies with sociological inquiry. The inclusion of genomic approaches promises to help address long-standing issues in sociology (e.g., selection effects), as well as open up new avenues for future research. This article reviews the substantive findings of behavior genetic/genomic research, both from the recent past (e.g., twin/adoption studies, candidate gene studies) and from contemporary genomic analyses. The article primarily focuses on modern genomic methods available to sociologists (e.g., polygenic score analysis) and their various applications for answering sociological questions. The article concludes by considering a number of areas to which genomic researchers and sociologists should pay close attention if a consilience between genomic methods and sociological research is to be fully realized.

Phenotypic Annotation: Using Polygenic Scores to Translate Discoveries From Genome-Wide Association Studies From the Top Down

Genome-wide association studies (GWASs) have identified specific genetic variants associated with complex human traits and behaviors, such as educational attainment, mental disorders, and personality. However, small effect sizes for individual variants, uncertainty regarding the biological function of discovered genotypes, and potential "outside-the-skin" environmental mechanisms leave a translational gulf between GWAS results and scientific understanding that will improve human health and well-being. We propose a set of social, behavioral, and brain-science research activities that map discovered genotypes to neural, developmental, and social mechanisms and call this research program phenotypic annotation. Phenotypic annotation involves (a) elaborating the nomological network surrounding discovered genotypes, (b) shifting focus from individual genes to whole genomes, and (c) testing how discovered genotypes affect life-span development. Phenotypic-annotation research is already advancing the understanding of GWAS discoveries for educational attainment and schizophrenia. We review examples and discuss methodological considerations for psychologists taking up the phenotypic-annotation approach.

Interpreting polygenic scores, polygenic adaptation, and human phenotypic differences

Recent analyses of polygenic scores have opened new discussions concerning the genetic basis and evolutionary significance of differences among populations in distributions of phenotypes. Here, we highlight limitations in research on polygenic scores, polygenic adaptation and population differences. We show how genetic contributions to traits, as estimated by polygenic scores, combine with environmental contributions so that differences among populations in trait distributions need not reflect corresponding differences in genetic propensity. Under a null model in which phenotypes are selectively neutral, genetic propensity differences contributing to phenotypic differences among populations are predicted to be small. We illustrate this null hypothesis in relation to health disparities between African Americans and European Americans, discussing alternative hypotheses with selective and environmental effects. Close attention to the limitations of research on polygenic phenomena is important for the interpretation of their relationship to human population differences.

Genetic instrumental variable regression: Explaining socioeconomic and health outcomes in nonexperimental data

We propose genetic instrumental variable (GIV) regression—a method that controls for pleiotropic effects of genes on two variables. GIV regression is broadly applicable to study outcomes for which polygenic scores from large-scale genome-wide association studies are available. We explore the performance of GIV regression in the presence of pleiotropy across a range of scenarios and find that it yields more accurate estimates than alternative approaches such as ordinary least-squares regression or Mendelian randomization. When GIV regression is combined with proper controls for purely environmental sources of bias (e.g., using control variables and sibling fixed effects), it improves our understanding of the causal relationships between genetically correlated variables.

Identifying causal effects in nonexperimental data is an enduring challenge. One proposed solution that recently gained popularity is the idea to use genes as instrumental variables [i.e., Mendelian randomization (MR)]. However, this approach is problematic because many variables of interest are genetically correlated, which implies the possibility that many genes could affect both the exposure and the outcome directly or via unobserved confounding factors. Thus, pleiotropic effects of genes are themselves a source of bias in nonexperimental data that would also undermine the ability of MR to correct for endogeneity bias from nongenetic sources. Here, we propose an alternative approach, genetic instrumental variable (GIV) regression, that provides estimates for the effect of an exposure on an outcome in the presence of pleiotropy. As a valuable byproduct, GIV regression also provides accurate estimates of the chip heritability of the outcome variable. GIV regression uses polygenic scores (PGSs) for the outcome of interest which can be constructed from genome-wide association study (GWAS) results. By splitting the GWAS sample for the outcome into nonoverlapping subsamples, we obtain multiple indicators of the outcome PGSs that can be used as instruments for each other and, in combination with other methods such as sibling fixed effects, can address endogeneity bias from both pleiotropy and the environment. In two empirical applications, we demonstrate that our approach produces reasonable estimates of the chip heritability of educational attainment (EA) and show that standard regression and MR provide upwardly biased estimates of the effect of body height on EA.

Predicting loneliness with polygenic scores of social, psychological and psychiatric traits

Loneliness is a heritable trait that accompanies multiple disorders. The association between loneliness and mental health indices may partly be due to inherited biological factors. We constructed polygenic scores for 27 traits related to behavior, cognition and mental health and tested their prediction for self-reported loneliness in a population-based sample of 8798 Dutch individuals. Polygenic scores for major depressive disorder (MDD), schizophrenia and bipolar disorder were significantly associated with loneliness. Of the Big Five personality dimensions, polygenic scores for neuroticism and conscientiousness also significantly predicted loneliness, as did the polygenic scores for subjective well-being, tiredness and self-rated health. When including all polygenic scores simultaneously into one model, only 2 major depression polygenic scores remained as significant predictors of loneliness. When controlling only for these 2 MDD polygenic scores, only neuroticism and schizophrenia remain significant. The total variation explained by all polygenic scores collectively was 1.7%. The association between the propensity to feel lonely and the susceptibility to psychiatric disorders thus pointed to a shared genetic etiology. The predictive power of polygenic scores will increase as the power of the genome-wide association studies on which they are based increases and may lead to clinically useful polygenic scores that can inform on the genetic predisposition to loneliness and mental health.

Geographic Clustering of Polygenic Scores at Different Stages of the Life Course

We interrogate state-level clustering of polygenic scores at different points in the life course and variation in the association of mean polygenic scores in a respondent's state of birth with corresponding phenotypes. The polygenic scores for height and smoking show the most state-level clustering (2 to 4 percent) with relatively little clustering observed for the other scores. However, even the small amounts of observed clustering are potentially meaningful. The state-mean polygenic score for educational attainment is strongly associated with an individual's educational attainment net of that person's polygenic score. The ecological clustering of polygenic scores may denote a new environmental factor in gene-environment research. We conclude by discussing possible mechanisms that underlie this association and the implications of our findings for social and genetic research.

Polygenic risk scores for schizophrenia and bipolar disorder associate with addiction

We use polygenic risk scores (PRSs) for schizophrenia (SCZ) and bipolar disorder (BPD) to predict smoking, and addiction to nicotine, alcohol or drugs in individuals not diagnosed with psychotic disorders. Using PRSs for 144 609 subjects, including 10 036 individuals admitted for in‐patient addiction treatment and 35 754 smokers, we find that diagnoses of various substance use disorders and smoking associate strongly with PRSs for SCZ (P = 5.3 × 10⁻⁵⁰–1.4 × 10⁻⁶) and BPD (P = 1.7 × 10⁻⁹–1.9 × 10⁻³), showing shared genetic etiology between psychosis and addiction. Using standardized scores for SCZ and BPD scaled to a unit increase doubling the risk of the corresponding disorder, the odds ratios for alcohol and substance use disorders range from 1.19 to 1.31 for the SCZ‐PRS, and from 1.07 to 1.29 for the BPD‐PRS. Furthermore, we show that as regular smoking becomes more stigmatized and less prevalent, these biological risk factors gain importance as determinants of the behavior.

Widespread covariation of early environmental exposures and trait-associated polygenic variation

Although gene-environment correlation is recognized and investigated by family studies and recently by SNP-heritability studies, the possibility that genetic effects on traits capture environmental risk factors or protective factors has been neglected by polygenic prediction models. We investigated covariation between trait-associated polygenic variation identified by genome-wide association studies (GWASs) and specific environmental exposures, controlling for overall genetic relatedness using a genomic relatedness matrix restricted maximum-likelihood model. In a UK-representative sample (n = 6,710), we find widespread covariation between offspring trait-associated polygenic variation and parental behavior and characteristics relevant to children's developmental outcomes-independently of population stratification. For instance, offspring genetic risk for schizophrenia was associated with paternal age (R² = 0.002; P = 1e-04), and offspring education-associated variation was associated with variance in breastfeeding (R² = 0.021; P = 7e-30), maternal smoking during pregnancy (R² = 0.008; P = 5e-13), parental smacking (R² = 0.01; P = 4e-15), household income (R² = 0.032; P = 1e-22), watching television (R² = 0.034; P = 5e-47), and maternal education (R² = 0.065; P = 3e-96). Education-associated polygenic variation also captured covariation between environmental exposures and children's inattention/hyperactivity, conduct problems, and educational achievement. The finding that genetic variation identified by trait GWASs partially captures environmental risk factors or protective factors has direct implications for risk prediction models and the interpretation of GWAS findings.

Genetic evidence for natural selection in humans in the contemporary United States

Recent findings from molecular genetics now make it possible to test directly for natural selection by analyzing whether genetic variants associated with various phenotypes have been under selection. I leverage these findings to construct polygenic scores that use individuals' genotypes to predict their body mass index, educational attainment (EA), glucose concentration, height, schizophrenia, total cholesterol, and (in females) age at menarche. I then examine associations between these scores and fitness to test whether natural selection has been occurring. My study sample includes individuals of European ancestry born between 1931 and 1953 who participated in the Health and Retirement Study, a representative study of the US population. My results imply that natural selection has been slowly favoring lower EA in both females and males, and are suggestive that natural selection may have favored a higher age at menarche in females. For EA, my estimates imply a rate of selection of about -1.5 mo of education per generation (which pales in comparison with the increases in EA observed in contemporary times). Although they cannot be projected over more than one generation, my results provide additional evidence that humans are still evolving-albeit slowly, especially compared with the rapid changes that have occurred over the past few generations due to cultural and environmental factors.

Assortative mating and differential fertility by phenotype and genotype across the 20th century

This study asks two related questions about the shifting landscape of marriage and reproduction in US society over the course of the last century with respect to a range of health and behavioral phenotypes and their associated genetic architecture: (i) Has assortment on measured genetic factors influencing reproductive and social fitness traits changed over the course of the 20th century? (ii) Has the genetic covariance between fitness (as measured by total fertility) and other traits changed over time? The answers to these questions inform our understanding of how the genetic landscape of American society has changed over the past century and have implications for population trends. We show that husbands and wives carry similar loadings for genetic factors related to education and height. However, the magnitude of this similarity is modest and has been fairly consistent over the course of the 20th century. This consistency is particularly notable in the case of education, for which phenotypic similarity among spouses has increased in recent years. Likewise, changing patterns of the number of children ever born by phenotype are not matched by shifts in genotype-fertility relationships over time. Taken together, these trends provide no evidence that social sorting is becoming increasingly genetic in nature or that dysgenic dynamics have accelerated.

Attention-deficit/hyperactivity disorder polygenic risk scores predict attention problems in a population-based sample of children

OBJECTIVE

Clinically, attention-deficit/hyperactivity disorder (ADHD) is characterized by hyperactivity, impulsivity, and inattention and is among the most common childhood disorders. These same traits that define ADHD are variable in the general population, and the clinical diagnosis may represent the extreme end of a continuous distribution of inattentive and hyperactive behaviors. This hypothesis can be tested by assessing the predictive value of polygenic risk scores derived from a discovery sample of ADHD patients in a target sample from the general population with continuous scores of inattention and hyperactivity. In addition, the genetic overlap between ADHD and continuous ADHD scores can be tested across rater and age.

METHOD

The Psychiatric Genomics Consortium has performed the largest genome-wide analysis (GWA) study of ADHD so far, including 5,621 clinical patients and 13,589 controls. The effects sizes of single nucleotide polymorphisms (SNPs) estimated in this meta-analysis were used to obtain individual polygenic risk scores in an independent population-based cohort of 2,437 children from the Netherlands Twin Register. The variance explained in Attention Problems (AP) scale scores by the polygenic risk scores was estimated by linear mixed modeling.

RESULTS

The ADHD polygenic risk scores significantly predicted both parent and teacher ratings of AP in preschool- and school-aged children.

CONCLUSION

These results indicate genetic overlap between a diagnosis of ADHD and AP scale scores across raters and age groups and provides evidence for a dimensional model of ADHD. Future GWA studies on ADHD can likely benefit from the inclusion of population-based cohorts and the analysis of continuous scores.

Polygenic scores associated with educational attainment in adults predict educational achievement and ADHD symptoms in children

The American Psychiatric Association estimates that 3 to 7 per cent of all school aged children are diagnosed with attention deficit hyperactivity disorder (ADHD). Even after correcting for general cognitive ability, numerous studies report a negative association between ADHD and educational achievement. With polygenic scores we examined whether genetic variants that have a positive influence on educational attainment have a protective effect against ADHD. The effect sizes from a large GWA meta-analysis of educational attainment in adults were used to calculate polygenic scores in an independent sample of 12-year-old children from the Netherlands Twin Register. Linear mixed models showed that the polygenic scores significantly predicted educational achievement, school performance, ADHD symptoms and attention problems in children. These results confirm the genetic overlap between ADHD and educational achievement, indicating that one way to gain insight into genetic variants responsible for variation in ADHD is to include data on educational achievement, which are available at a larger scale.