No abstract

Knowledge of genome-wide genealogies for thousands of individuals would simplify most evolutionary analyses for humans and other species, but has remained computationally infeasible. We developed a method, Relate, scaling to > 10,000 sequences while simultaneously estimating branch lengths, mutational ages, and variable historical population sizes, as well as allowing for data errors. Application to 1000 Genomes Project haplotypes produces joint genealogical histories for 26 human populations. Highly diverged lineages are present in all groups, but most frequent in Africa. Outside Africa, these mainly reflect ancient introgression from groups related to Neanderthals and Denisovans, while African signals instead reflect unknown events, unique to that continent. Our approach allows more powerful inferences of natural selection than previously possible. We identify multiple novel regions under strong positive selection, and multi-allelic traits including hair colour, BMI, and blood pressure, showing strong evidence of directional selection, varying among human groups.Large-scale genetic variation datasets are now available for a variety of species, including tens of thousands of humans. In principle, all information about a sample's genetic history is captured by their underlying genealogical history, which records the historical coalescence, recombination, and mutation events that produced the observed variation patterns. In practice, several key existing approaches (e.g., Refs. [1,2]) leverage an underlying coalescent model, because this provides a flexible modelling framework and is the limiting behaviour of a variety of finite-population models 3,4 . However, inference under the coalescent is complicated by the structure of the model, uncertainty over the correct genealogy conditional on observed data, and the large resulting space of possible sample histories 5 . Other approaches 6-11 use more heuristic approximations to the coalescent, sometimes reducing accuracy: regardless, all published existing methods scale to tens or a few hundred samples at most.As a result of these issues, the use of direct genealogy-based inference to detect recombination events, date mutations, and reveal evidence of positive selection has been limited to smaller datasets 1,2 , while for larger datasets approaches based on data summaries 12-14 or downsampling 15,16 have predominated. A diverse set of tools have detected genetic structure that is in good agreement with geopolitical separation over generations 17 .Admixtures of ancient populations have been identified and dated 18 . Other applications have found bottlenecks in population sizes that are consistent with anthropological evidence of initial human migration from the African continent 15,19-21 and evidence of subsequent introgression with archaic humans, such as Neanderthals 22 .We have developed a scalable method, Relate, to estimate genome-wide genealogies (see Figure 1; Methods;URLs for implementation). Relate separates two steps; firstly identifying a genealogical fra...

Background Accurately quantifying the risk of osteoporotic fracture is important for directing appropriate clinical interventions. While skeletal measures such as heel quantitative speed of sound (SOS) and dual-energy X-ray absorptiometry bone mineral density are able to predict the risk of osteoporotic fracture, the utility of such measurements is subject to the availability of equipment and human resources. Using data from 341,449 individuals of white British ancestry, we previously developed a genome-wide polygenic risk score (PRS), called gSOS, that captured 25.0% of the total variance in SOS. Here, we test whether gSOS can improve fracture risk prediction. Methods We examined the predictive power of gSOS in five genome-wide genotyped cohorts, including 90,172 individuals of European ancestry and 25,034 individuals of Asian ancestry. We calculated gSOS for each individual and tested for the association between gSOS and incident major osteoporotic fracture and hip fracture. We tested whether adding gSOS to the risk prediction models had added value over models using other commonly used clinical risk factors. Results A standard deviation decrease in gSOS was associated with an increased odds of incident major osteoporotic fracture in populations of European ancestry, with odds ratios ranging from 1.35 to 1.46 in four cohorts. It was also associated with a 1.26-fold (95% confidence interval (CI) 1.13–1.41) increased odds of incident major osteoporotic fracture in the Asian population. We demonstrated that gSOS was more predictive of incident major osteoporotic fracture (area under the receiver operating characteristic curve (AUROC) = 0.734; 95% CI 0.727–0.740) and incident hip fracture (AUROC = 0.798; 95% CI 0.791–0.805) than most traditional clinical risk factors, including prior fracture, use of corticosteroids, rheumatoid arthritis, and smoking. We also showed that adding gSOS to the Fracture Risk Assessment Tool (FRAX) could refine the risk prediction with a positive net reclassification index ranging from 0.024 to 0.072. Conclusions We generated and validated a PRS for SOS which was associated with the risk of fracture. This score was more strongly associated with the risk of fracture than many clinical risk factors and provided an improvement in risk prediction. gSOS should be explored as a tool to improve risk stratification to identify individuals at high risk of fracture.

Background Since screening programs identify only a small proportion of the population as eligible for an intervention, genomic prediction of heritable risk factors could decrease the number needing to be screened by removing individuals at low genetic risk. We therefore tested whether a