Calibrated prediction intervals for polygenic scores across diverse contexts.

Polygenic scores (PGS) have emerged as the tool of choice for genomic prediction in a wide range of fields. We show that PGS performance varies broadly across contexts and biobanks. Contexts such as age, sex and income can impact PGS accuracy with similar magnitudes as genetic ancestry. Here we introduce an approach (CalPred) that models all contexts jointly to produce prediction intervals that vary across contexts to achieve calibration (include the trait with 90% probability), whereas existing methods are miscalibrated. In analyses of 72 traits across large and diverse biobanks (All of Us and UK Biobank), we find that prediction intervals required adjustment by up to 80% for quantitative traits. For disease traits, PGS-based predictions were miscalibrated across socioeconomic contexts such as annual household income levels, further highlighting the need of accounting for context information in PGS-based prediction across diverse populations.

Discovering non-additive heritability using additive GWAS summary statistics.

LD score regression (LDSC) is a method to estimate narrow-sense heritability from genome-wide association study (GWAS) summary statistics alone, making it a fast and popular approach. In this work, we present interaction-LD score (i-LDSC) regression: an extension of the original LDSC framework that accounts for interactions between genetic variants. By studying a wide range of generative models in simulations, and by re-analyzing 25 well-studied quantitative phenotypes from 349,468 individuals in the UK Biobank and up to 159,095 individuals in BioBank Japan, we show that the inclusion of a cis-interaction score (i.e. interactions between a focal variant and proximal variants) recovers genetic variance that is not captured by LDSC. For each of the 25 traits analyzed in the UK Biobank and BioBank Japan, i-LDSC detects additional variation contributed by genetic interactions. The i-LDSC software and its application to these biobanks represent a step towards resolving further genetic contributions of sources of non-additive genetic effects to complex trait variation.

Tradeoffs in Modeling Context Dependency in Complex Trait Genetics.

Genetic effects on complex traits may depend on context, such as age, sex, environmental exposures or social settings. However, it is often unclear if the extent of context dependency, or Gene-by-Environment interaction (GxE), merits more involved models than the additive model typically used to analyze data from genome-wide association studies (GWAS). Here, we suggest considering the utility of GxE models in GWAS as a tradeoff between bias and variance parameters. In particular, We derive a decision rule for choosing between competing models for the estimation of allelic effects. The rule weighs the increased estimation noise when context is considered against the potential bias when context dependency is ignored. In the empirical example of GxSex in human physiology, the increased noise of context-specific estimation often outweighs the bias reduction, rendering GxE models less useful when variants are considered independently. However, we argue that for complex traits, the joint consideration of context dependency across many variants mitigates both noise and bias. As a result, polygenic GxE models can improve both estimation and trait prediction. Finally, we exemplify (using GxDiet effects on longevity in fruit flies) how analyses based on independently ascertained "top hits" alone can be misleading, and that considering polygenic patterns of GxE can improve interpretation.

and the Persistence of Hereditarian Fallacies.

Clark (2023) considers the similarity in socioeconomic status between relatives, drawing on records spanning four centuries in England. The paper adapts a classic quantitative genetics model in order to argue the fit of the model to the data suggests that: (1) variation in socioeconomic status is largely determined by additive genetic variation; (2) contemporary English people "remain correlated in outcomes with their lineage relatives in exactly the same way as in preindustrial England"; and (3) social mobility has remained static over this time period due to strong assortative mating on a "social genotype." These conclusions are based on a misconstrual of model parameters, which conflates genetic and non-genetic transmission (e.g. of wealth) within families. As we show, there is strong confounding of genetic and non-genetic sources of similarity in these data. Inconsistent with claims (2) and (3), we show that familial correlations in status are variable-generally decreasing-through the time period analyzed. Lastly, we find that statistical artifacts substantially bias estimates of familial correlations in the paper. Overall, Clark (2023) provides no information about the relative contribution of genetic and non-genetic factors to social status.

Calibrated prediction intervals for polygenic scores across diverse contexts.

Polygenic scores (PGS) have emerged as the tool of choice for genomic prediction in a wide range of fields from agriculture to personalized medicine. We analyze data from two large biobanks in the US (All of Us) and the UK (UK Biobank) to find widespread variability in PGS performance across contexts. Many contexts, including age, sex, and income, impact PGS accuracies with similar magnitudes as genetic ancestry. PGSs trained in single versus multi-ancestry cohorts show similar context-specificity in their accuracies. We introduce trait prediction intervals that are allowed to vary across contexts as a principled approach to account for context-specific PGS accuracy in genomic prediction. We model the impact of all contexts in a joint framework to enable PGS-based trait predictions that are well-calibrated (contain the trait value with 90% probability in all contexts), whereas methods that ignore context are mis-calibrated. We show that prediction intervals need to be adjusted for all considered traits ranging from 10% for diastolic blood pressure to 80% for waist circumference. Adjustment of prediction intervals depends on the dataset; for example, prediction intervals for education years need to be adjusted by 90% in All of Us versus 8% in UK Biobank. Our results provide a path forward towards utilization of PGS as a prediction tool across all individuals regardless of their contexts while highlighting the importance of comprehensive profile of context information in study design and data collection.

Amplification is the primary mode of gene-by-sex interaction in complex human traits.

Sex differences in complex traits are suspected to be in part due to widespread gene-by-sex interactions (GxSex), but empirical evidence has been elusive. Here, we infer the mixture of ways in which polygenic effects on physiological traits covary between males and females. We find that GxSex is pervasive but acts primarily through systematic sex differences in the magnitude of many genetic effects ("amplification") rather than in the identity of causal variants. Amplification patterns account for sex differences in trait variance. In some cases, testosterone may mediate amplification. Finally, we develop a population-genetic test linking GxSex to contemporary natural selection and find evidence of sexually antagonistic selection on variants affecting testosterone levels. Our results suggest that amplification of polygenic effects is a common mode of GxSex that may contribute to sex differences and fuel their evolution.

1,000 ancient genomes uncover 10,000 years of natural selection in Europe.

Ancient DNA has revolutionized our understanding of human population history. However, its potential to examine how rapid cultural evolution to new lifestyles may have driven biological adaptation has not been met, largely due to limited sample sizes. We assembled genome-wide data from 1,291 individuals from Europe over 10,000 years, providing a dataset that is large enough to resolve the timing of selection into the Neolithic, Bronze Age, and Historical periods. We identified 25 genetic loci with rapid changes in frequency during these periods, a majority of which were previously undetected. Signals specific to the Neolithic transition are associated with body weight, diet, and lipid metabolism-related phenotypes. They also include immune phenotypes, most notably a locus that confers immunity to Salmonella infection at a time when ancient Salmonella genomes have been shown to adapt to human hosts, thus providing a possible example of human-pathogen co-evolution. In the Bronze Age, selection signals are enriched near genes involved in pigmentation and immune-related traits, including at a key human protein interactor of SARS-CoV-2. Only in the Historical period do the selection candidates we detect largely mirror previously-reported signals, highlighting how the statistical power of previous studies was limited to the last few millennia. The Historical period also has multiple signals associated with vitamin D binding, providing evidence that lactase persistence may have been part of an oligogenic adaptation for efficient calcium uptake and challenging the theory that its adaptive value lies only in facilitating caloric supplementation during times of scarcity. Finally, we detect selection on complex traits in all three periods, including selection favoring variants that reduce body weight in the Neolithic. In the Historical period, we detect selection favoring variants that increase risk for cardiovascular disease plausibly reflecting selection for a more active inflammatory response that would have been adaptive in the face of increased infectious disease exposure. Our results provide an evolutionary rationale for the high prevalence of these deadly diseases in modern societies today and highlight the unique power of ancient DNA in elucidating biological change that accompanied the profound cultural transformations of recent human history.

Concerted evolution reveals co-adapted amino acid substitutions in Na+K+-ATPase of frogs that prey on toxic toads.

Although gene duplication is an important source of evolutionary innovation, the functional divergence of duplicates can be opposed by ongoing gene conversion between them. Here, we report on the evolution of a tandem duplication of Na+,K+-ATPase subunit α1 (ATP1A1) shared by frogs in the genus Leptodactylus, a group of species that feeds on toxic toads. One ATP1A1 paralog evolved resistance to toad toxins although the other retained ancestral susceptibility. Within species, frequent non-allelic gene conversion homogenized most of the sequence between the two copies but was counteracted by strong selection on 12 amino acid substitutions that distinguish the two paralogs. Protein-engineering experiments show that two of these substitutions substantially increase toxin resistance, whereas the additional 10 mitigate their deleterious effects on ATPase activity. Our results reveal how examination of neo-functionalized gene duplicate evolution can help pinpoint key functional substitutions and interactions with the genetic backgrounds on which they arise.

Mutability of demographic noise in microbial range expansions.

Demographic noise, the change in the composition of a population due to random birth and death events, is an important driving force in evolution because it reduces the efficacy of natural selection. Demographic noise is typically thought to be set by the population size and the environment, but recent experiments with microbial range expansions have revealed substantial strain-level differences in demographic noise under the same growth conditions. Many genetic and phenotypic differences exist between strains; to what extent do single mutations change the strength of demographic noise? To investigate this question, we developed a high-throughput method for measuring demographic noise in colonies without the need for genetic manipulation. By applying this method to 191 randomly-selected single gene deletion strains from the E. coli Keio collection, we find that a typical single gene deletion mutation decreases demographic noise by 8% (maximal decrease: 81%). We find that the strength of demographic noise is an emergent trait at the population level that can be predicted by colony-level traits but not cell-level traits. The observed differences in demographic noise from single gene deletions can increase the establishment probability of beneficial mutations by almost an order of magnitude (compared to in the wild type). Our results show that single mutations can substantially alter adaptation through their effects on demographic noise and suggest that demographic noise can be an evolvable trait of a population.

The evolution of group differences in changing environments.

The selection pressures that have shaped the evolution of complex traits in humans remain largely unknown, and in some contexts highly contentious, perhaps above all where they concern mean trait differences among groups. To date, the discussion has focused on whether such group differences have any genetic basis, and if so, whether they are without fitness consequences and arose via random genetic drift, or whether they were driven by selection for different trait optima in different environments. Here, we highlight a plausible alternative: that many complex traits evolve under stabilizing selection in the face of shifting environmental effects. Under this scenario, there will be rapid evolution at the loci that contribute to trait variation, even when the trait optimum remains the same. These considerations underscore the strong assumptions about environmental effects that are required in ascribing trait differences among groups to genetic differences.

Genetic Adaptation in New York City Rats.

Brown rats (Rattus norvegicus) thrive in urban environments by navigating the anthropocentric environment and taking advantage of human resources and by-products. From the human perspective, rats are a chronic problem that causes billions of dollars in damage to agriculture, health, and infrastructure. Did genetic adaptation play a role in the spread of rats in cities? To approach this question, we collected whole-genome sequences from 29 brown rats from New York City (NYC) and scanned for genetic signatures of adaptation. We tested for 1) high-frequency, extended haplotypes that could indicate selective sweeps and 2) loci of extreme genetic differentiation between the NYC sample and a sample from the presumed ancestral range of brown rats in northeast China. We found candidate selective sweeps near or inside genes associated with metabolism, diet, the nervous system, and locomotory behavior. Patterns of differentiation between NYC and Chinese rats at putative sweep loci suggest that many sweeps began after the split from the ancestral population. Together, our results suggest several hypotheses on adaptation in rats living in proximity to humans.

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.

Complex diseases like cancer and heart disease are caused by the interplay of many factors: the variants of genes we inherit, the lifestyles we lead and the environments we inhabit, plus the interaction of all these factors. In fact, almost every trait, even how many years we will spend studying, is influenced both by our environment and our genes. To identify some of the genetic factors at play, scientists perform analyses known as genome-wide association studies, or GWAS for short. In these studies, the genomes from many different people are scanned to look for genetic differences associated with differences in traits. By summing up all the small genetic differences, so-called "polygenic scores" can be calculated. When there is a large genetic component to a trait, polygenic scores can be useful predictive tools. But there is a catch: polygenic scores make less accurate predictions for individuals of a different ancestry than those involved in the GWAS, which limits the use of these tools around the world. Mostafavi, Harpak et al. set out to understand if there are other factors in addition to ancestry that could influence the performance of polygenic scores. Using data from the UK Biobank, an international health resource that pairs genomic data and clinical information, Mostafavi, Harpak et al. examined polygenic scores among individuals that share a single, common ancestry. These polygenic scores were used to predict three traits (blood pressure, body mass index and educational attainment) in individuals and the predictions were then compared to the actual trait values to see how accurate they were. The analysis revealed that even within a group of people with similar ancestry, the accuracy of polygenic scores can vary, depending on characteristics such as the sex, age or socioeconomic status of the individuals. This analysis emphasises how variable GWAS and their predictive value can be even within seemingly similar population groups. It further highlights both the complexities of interpreting polygenic scores and underappreciated obstacles to their broad use in medical and social sciences.

Reduced signal for polygenic adaptation of height in UK Biobank.

Several recent papers have reported strong signals of selection on European polygenic height scores. These analyses used height effect estimates from the GIANT consortium and replication studies. Here, we describe a new analysis based on the the UK Biobank (UKB), a large, independent dataset. We find that the signals of selection using UKB effect estimates are strongly attenuated or absent. We also provide evidence that previous analyses were confounded by population stratification. Therefore, the conclusion of strong polygenic adaptation now lacks support. Moreover, these discrepancies highlight (1) that methods for correcting for population stratification in GWAS may not always be sufficient for polygenic trait analyses, and (2) that claims of differences in polygenic scores between populations should be treated with caution until these issues are better understood. Editorial note: This article has been through an editorial process in which the authors decide how to respond to the issues raised during peer review. The Reviewing Editor's assessment is that all the issues have been addressed (see decision letter).

Evidence for Weak Selective Constraint on Human Gene Expression.

Gene expression variation is a major contributor to phenotypic variation in human complex traits. Selection on complex traits may therefore be reflected in constraint on gene expression. Here, we explore the effects of stabilizing selection on cis-regulatory genetic variation in humans. We analyze patterns of expression variation at copy number variants and find evidence for selection against large increases in gene expression. Using allele-specific expression (ASE) data, we further show evidence of selection against smaller-effect variants. We estimate that, across all genes, singletons in a sample of 122 individuals have ∼2.2× greater effects on expression variation than the average variant across allele frequencies. Despite their increased effect size relative to common variants, we estimate that singletons in the sample studied explain, on average, only 5% of the heritability of gene expression from cis-regulatory variants. Finally, we show that genes depleted for loss-of-function variants are also depleted for cis-eQTLs and have low levels of allelic imbalance, confirming tighter constraint on the expression levels of these genes. We conclude that constraint on gene expression is present, but has relatively weak effects on most cis-regulatory variants, thus permitting high levels of gene-regulatory genetic variation.

Clonal interference can cause wavelet-like oscillations of multilocus linkage disequilibrium.

Within-host adaptation of pathogens such as human immunodeficiency virus (HIV) often occurs at more than two loci. Multiple beneficial mutations may arise simultaneously on different genetic backgrounds and interfere, affecting each other's fixation trajectories. Here, we explore how these evolutionary dynamics are mirrored in multilocus linkage disequilibrium (MLD), a measure of multi-way associations between alleles. In the parameter regime corresponding to HIV, we show that deterministic early infection models induce MLD to oscillate over time in a wavelet-like fashion. We find that the frequency of these oscillations is proportional to the rate of adaptation. This signature is robust to drift, but can be eroded by high variation in fitness effects of beneficial mutations. Our findings suggest that MLD oscillations could be used as a signature of interference among multiple equally advantageous mutations and may aid the interpretation of MLD in data.

Frequent nonallelic gene conversion on the human lineage and its effect on the divergence of gene duplicates.

Gene conversion is the copying of a genetic sequence from a "donor" region to an "acceptor." In nonallelic gene conversion (NAGC), the donor and the acceptor are at distinct genetic loci. Despite the role NAGC plays in various genetic diseases and the concerted evolution of gene families, the parameters that govern NAGC are not well characterized. Here, we survey duplicate gene families and identify converted tracts in 46% of them. These conversions reflect a large GC bias of NAGC. We develop a sequence evolution model that leverages substantially more information in duplicate sequences than used by previous methods and use it to estimate the parameters that govern NAGC in humans: a mean converted tract length of 250 bp and a probability of [Formula: see text] per generation for a nucleotide to be converted (an order of magnitude higher than the point mutation rate). Despite this high baseline rate, we show that NAGC slows down as duplicate sequences diverge-until an eventual "escape" of the sequences from its influence. As a result, NAGC has a small average effect on the sequence divergence of duplicates. This work improves our understanding of the NAGC mechanism and the role that it plays in the evolution of gene duplicates.

Between-region genetic divergence reflects the mode and tempo of tumor evolution.

Given the implications of tumor dynamics for precision medicine, there is a need to systematically characterize the mode of evolution across diverse solid tumor types. In particular, methods to infer the role of natural selection within established human tumors are lacking. By simulating spatial tumor growth under different evolutionary modes and examining patterns of between-region subclonal genetic divergence from multiregion sequencing (MRS) data, we demonstrate that it is feasible to distinguish tumors driven by strong positive subclonal selection from those evolving neutrally or under weak selection, as the latter fail to dramatically alter subclonal composition. We developed a classifier based on measures of between-region subclonal genetic divergence and projected patient data into model space, finding different modes of evolution both within and between solid tumor types. Our findings have broad implications for how human tumors progress, how they accumulate intratumoral heterogeneity, and ultimately how they may be more effectively treated.

Mutation Rate Variation is a Primary Determinant of the Distribution of Allele Frequencies in Humans.

The site frequency spectrum (SFS) has long been used to study demographic history and natural selection. Here, we extend this summary by examining the SFS conditional on the alleles found at the same site in other species. We refer to this extension as the "phylogenetically-conditioned SFS" or cSFS. Using recent large-sample data from the Exome Aggregation Consortium (ExAC), combined with primate genome sequences, we find that human variants that occurred independently in closely related primate lineages are at higher frequencies in humans than variants with parallel substitutions in more distant primates. We show that this effect is largely due to sites with elevated mutation rates causing significant departures from the widely-used infinite sites mutation model. Our analysis also suggests substantial variation in mutation rates even among mutations involving the same nucleotide changes. In summary, we show that variable mutation rates are key determinants of the SFS in humans.

Thousands of novel translated open reading frames in humans inferred by ribosome footprint profiling.

Accurate annotation of protein coding regions is essential for understanding how genetic information is translated into function. We describe riboHMM, a new method that uses ribosome footprint data to accurately infer translated sequences. Applying riboHMM to human lymphoblastoid cell lines, we identified 7273 novel coding sequences, including 2442 translated upstream open reading frames. We observed an enrichment of footprints at inferred initiation sites after drug-induced arrest of translation initiation, validating many of the novel coding sequences. The novel proteins exhibit significant selective constraint in the inferred reading frames, suggesting that many are functional. Moreover, ~40% of bicistronic transcripts showed negative correlation in the translation levels of their two coding sequences, suggesting a potential regulatory role for these novel regions. Despite known limitations of mass spectrometry to detect protein expressed at low level, we estimated a 14% validation rate. Our work significantly expands the set of known coding regions in humans.