Localization of adaptive variants in human genomes using averaged one-dependence estimation

Sugden, Lauren Alpert; Atkinson, Elizabeth G.; Fischer, Annie P; Rong, Stephen; Henn, Brenna M.; Ramachandran, Sohini

doi:10.1101/229070

Cited by 20 publications

(34 citation statements)

References 142 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our model, similar to others (e.g., Lin et al, 2011;Schrider and Kern, 2016b;Sugden et al, 2018), not only assigns class labels, but also provides a probability for each of the K classes. A properly-calibrated classifier should be one in which the probability of observing a given class is the actual fraction of times that the classifier chooses this class.…”

Section: Calibrating Class Probabilitiesmentioning

confidence: 99%

“…Several methods have recently been developed that incorporate information from multiple summary statistics to locate positively-selected genomic regions (Lin et al, 2011;Ronen et al, 2013;Pybus et al, 2015;Schrider and Kern, 2016b;Sheehan and Song, 2016;Kern and Schrider, 2018;Sugden et al, 2018). Most existing supervised learning approaches for detecting sweeps use combinations of summary statistics calculated in genomic windows of simulated chromosomes to train classifiers using methods such as support vector machines, random forests, neural networks, and boosting.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Localizing and classifying adaptive targets with trend filtered regression

Mughal

DeGiorgio

2018

Preprint

View full text Add to dashboard Cite

Identifying genomic locations of natural selection from sequence data is an ongoing challenge in population genetics. Current methods utilizing information combined from several summary statistics typically assume no correlation of summary statistics regardless of the genomic location from which they are calculated. However, due to linkage disequilibrium, summary statistics calculated at nearby genomic positions are highly correlated. We introduce an approach termed Trendsetter that accounts for the similarity of statistics calculated from adjacent genomic regions through trend filtering, while reducing the effects of multicollinearity through regularization. Our penalized regression framework has high power to detect sweeps, is capable of classifying sweep regions as either hard or soft, and can be applied to other selection scenarios as well. We find that Trendsetter is robust to both extensive missing data and strong background selection, and has comparable power to similar current approaches. Moreover, the model learned by Trendsetter can be viewed as a set of curves modeling the spatial distribution of summary statistics in the genome. Application to human genomic data revealed positively-selected regions previously discovered such as LCT in Europeans and EDAR in East Asians. We also identified a number of novel candidates and show that populations with greater relatedness share more sweep signals.

show abstract

Section: Calibrating Class Probabilitiesmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Localizing and classifying adaptive targets with trend filtered regression

Mughal

DeGiorgio

2018

Preprint

View full text Add to dashboard Cite

show abstract

“…[9][10][11]). 36 The hitch-hiking effect provides a key signature of selection in modern datasets [12]. [12,13].…”

Section: Introduction 29mentioning

confidence: 99%

“…However, a major disadvantage of 92 such approaches is that the amount of simulation necessary to obtain an accurate estimate grows 93 dramatically with the dimensionality of the observed data (for a discussion, see e.g. [36]); similar 94 issues arise in the process of training machine learning methods (e.g. [22]), requiring considerations 95 to prevent overfitting and avoid excessive simulation.…”

Section: Introduction 29mentioning

confidence: 99%

An approximate full-likelihood method for inferring selection and allele frequency trajectories from DNA sequence data

Stern¹,

Wilton

Nielsen³

2019

Preprint

View full text Add to dashboard Cite

1Most current methods for detecting natural selection from DNA sequence data are limited in that 2 they are either based on summary statistics or a composite likelihood, and as a consequence, do not 3 make full use of the information available in DNA sequence data. We here present a new importance 4 sampling approach for approximating the full likelihood function for the selection coefficient. The 5 method treats the ancestral recombination graph (ARG) as a latent variable that is integrated out 6 using previously published Markov Chain Monte Carlo (MCMC) methods. The method can be used 7 for detecting selection, estimating selection coefficients, testing models of changes in the strength of 8 selection, estimating the time of the start of a selective sweep, and for inferring the allele frequency 9 trajectory of a selected or neutral allele. We perform extensive simulations to evaluate the method 10 and show that it uniformly improves power to detect selection compared to current popular methods 11 such as nSL and SDS, under various demographic models and can provide reliable inferences of allele 12 frequency trajectories under many conditions. We also explore the potential of our method to detect 13 extremely recent changes in the strength of selection. We use the method to infer the past allele 14 frequency trajectory for a lactase persistence SNP (MCM6) in Europeans. We also study a set of 11 15 pigmentation-associated variants. Several genes show evidence of strong selection particularly within 16 the last 5,000 years, including ASIP, KITLG, and TYR. However, selection on OCA2/HERC2 seems 17 to be much older and, in contrast to previous claims, we find no evidence of selection on TYRP1. 18Author summary 19 Current methods to study natural selection using modern population genomic data are limited in their 20 power and flexibility. Here, we present a new method to infer natural selection that builds on recent 21 methodological advances in estimating genome-wide genealogies. By using importance sampling 22 we are able to efficiently estimate the likelihood function of the selection coefficient. We show our 23 method improves power to test for selection over competing methods across a diverse range of 24 scenarios, and also accurately infers the selection coefficient. We also demonstrate a novel capability 25 of our model, using it to infer the allele's frequency over time. We validate these results with a 26 study of a lactase persistence SNP in Europeans, and also study a set of 11 pigmentation-associated 27 variants. 28 1/46 106 Furthermore, the new method is, to our knowledge, the first that is capable of inferring the allele 107 frequency trajectories for models with recombination and selection using only modern data. We are 108 able to accomplish this task using the aforementioned Markovian structure of both coalescence and 109 the trajectory, forming a HMM over these two hidden states and solving for the posterior marginals 110 of each hidden allele frequency state over time. Recently, Edge & Coop propose...

show abstract

“…Instead, we recommend that the reader take our chi-squared-distributed P-values with a grain of salt, and merely use them as a way to prioritize regions for more extensive downstream modeling and validation (for example, using methods like those described in refs. [51][52][53]).…”

Section: Selection Of Candidate Regionsmentioning

confidence: 99%

Identifying loci under positive selection in complex population histories

Refoyo-Martínez

Fonseca

Halldórsdóttir

et al. 2018

Preprint

View full text Add to dashboard Cite

Detailed modeling of a species' history is of prime importance for understanding how natural selection operates over time. Most methods designed to detect positive selection along sequenced genomes, however, use simplified representations of past histories as null models of genetic drift. Here, we present the first method that can detect signatures of strong local adaptation across the genome using arbitrarily complex admixture graphs, which are typically used to describe the history of past divergence and admixture events among any number of populations. The method-called Graph-aware Retrieval of Selective Sweeps (GRoSS )-has good power to detect loci in the genome with strong evidence for past selective sweeps and can also identify which branch of the graph was most affected by the sweep. As evidence of its utility, we apply the method to bovine, codfish and human population genomic data containing multiple population panels related in complex ways. We find new candidate genes for important adaptive functions, including immunity and metabolism in under-studied human populations, as well as muscle mass, milk production and tameness in specific bovine breeds. We are also able to pinpoint the emergence of large regions of differentiation due to inversions in the history of Atlantic codfish.

show abstract

Localization of adaptive variants in human genomes using averaged one-dependence estimation

Cited by 20 publications

References 142 publications

Localizing and classifying adaptive targets with trend filtered regression

Localizing and classifying adaptive targets with trend filtered regression

An approximate full-likelihood method for inferring selection and allele frequency trajectories from DNA sequence data

Identifying loci under positive selection in complex population histories

Contact Info

Product

Resources

About