A Coalescent-Based Estimator of Admixture From DNA Sequences

Wang, Jinliang

doi:10.1534/genetics.105.054130

Cited by 17 publications

(10 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Drift affected significantly the estimation of p 1 , and even with 10 biallelic loci the posteriors for p 1 were rather wide with t i ¼ 0.01 (10 generations of drift for N e ¼ 1000). This is in agreement with Choisy et al (2004), Excoffier et al (2005), and Wang (2006) and confirms that methods that do not account for drift to estimate demographic parameters will tend to provide misleadingly precise values. Overall, the simulation study results show that ABC_ SUMSTAT provides good approximations to the full likelihood and is probably easier to use than ABC_ ALL_FREQ, despite the potential problem of choosing the summary statistics (but see Joyce and Marjoram 2008).…”

Section: Discussionsupporting

confidence: 82%

Approximate Bayesian Computation Without Summary Statistics: The Case of Admixture

et al. 2009

View full text Add to dashboard Cite

In recent years approximate Bayesian computation (ABC) methods have become popular in population genetics as an alternative to full-likelihood methods to make inferences under complex demographic models. Most ABC methods rely on the choice of a set of summary statistics to extract information from the data. In this article we tested the use of the full allelic distribution directly in an ABC framework. Although the ABC techniques are becoming more widely used, there is still uncertainty over how they perform in comparison with full-likelihood methods. We thus conducted a simulation study and provide a detailed examination of ABC in comparison with full likelihood in the case of a model of admixture. This model assumes that two parental populations mixed at a certain time in the past, creating a hybrid population, and that the three populations then evolve under pure drift. Several aspects of ABC methodology were investigated, such as the effect of the distance metric chosen to measure the similarity between simulated and observed data sets. Results show that in general ABC provides good approximations to the posterior distributions obtained with the full-likelihood method. This suggests that it is possible to apply ABC using allele frequencies to make inferences in cases where it is difficult to select a set of suitable summary statistics and when the complexity of the model or the size of the data set makes it computationally prohibitive to use full-likelihood methods. T HE genetic patterns observed today in most speciesare the result of complex histories, which include demographic events such as population admixture, expansions, and/or collapses. The detection and quantification of such events relies on the fact that different scenarios leave a specific genetic signature in present-day populations, as well as on knowledge from other sources (e.g. ecology, biogeography, archeology) to define plausible models to explain such patterns. Recent population genetic modeling has seen the development of a number of statistical approaches that aim at extracting as much information as possible from the full allelic distributions (Griffiths and Tavaré 1994;Wilson and Balding 1998;Beaumont 1999;Beerli and Felsenstein 2001;Chikhi et al. 2001;Storz et al. 2002). These approaches aim at computing the likelihood L(u), i.e., the probability P M (D j u) of generating the observed data D under some demographic model M, defined by a set of parameters u ¼ (u 1 , . . . , u k ). In Bayesian statistics, the posterior density is used to make inferences as it reflects the probability of the parameters given the data, and it is obtained through the relationship P(u j D) } L(u)P(u), where P(u) summarizes prior knowledge (or lack thereof) regarding u before the data are observed (Beaumont and Rannala 2004). For most demographic models there are no explicit likelihood functions or the likelihood cannot be derived analytically. Therefore, full-likelihood approaches rely on methods that explore the parameter space efficiently, such as...

show abstract

Section: Discussionsupporting

confidence: 82%

Approximate Bayesian Computation Without Summary Statistics: The Case of Admixture

et al. 2009

View full text Add to dashboard Cite

show abstract

“…The allele frequency in common ancestral population P 0 could be determined by either coalescence theory or uniform distribution with range from 0 to 122293031. However, joint frequency spectrum of given alleles in descendent populations (P 1′ , P 2′ , P h' , P 1 , P 2 and P h ) is difficult to be presented in closed form.…”

Section: Methodsmentioning

confidence: 99%

Paleolithic Contingent in Modern Japanese: Estimation and Inference using Genome-wide Data

Wang

et al. 2012

Sci Rep

View full text Add to dashboard Cite

The genetic origins of Japanese populations have been controversial. Upper Paleolithic Japanese, i.e. Jomon, developed independently in Japanese islands for more than 10,000 years until the isolation was ended with the influxes of continental immigrants about 2,000 years ago. However, the knowledge of origin of Jomon and its contribution to the genetic pool of contemporary Japanese is still limited, albeit the extensive studies using mtDNA and Y chromosomes. In this report, we aimed to infer the origin of Jomon and to estimate its contribution to Japanese by fitting an admixture model with missing data from Jomon to a genome-wide data from 94 worldwide populations. Our results showed that the genetic contributions of Jomon, the Paleolithic contingent in Japanese, are 54.3∼62.3% in Ryukyuans and 23.1∼39.5% in mainland Japanese, respectively. Utilizing inferred allele frequencies of the Jomon population, we further showed the Paleolithic contingent in Japanese had a Northeast Asia origin.

show abstract

“…They evaluate the performance of the m Y i using simulations and conclude that despite the increased number of parameters the level of precision in the estimates remains approximately the same as in the original study with two parental populations. A general improvement to the method of Bertorelle and Excoffier (1998) for sequence data, assuming an infinite sites model, has been described by Wang (2006). He derives formulae assuming an infinite sites model for the expected number of segregating sites in each population, in each pair of populations pooled together, and in all 3 populations pooled together.…”

Section: Admixturementioning

confidence: 99%

Conservation Genetics

Beaumont¹

2007

Handbook of Statistical Genetics

View full text Add to dashboard Cite

This chapter reviews some of the statistical methodologies used in the genetic analysis of endangered and managed populations. The topics covered include the estimation of effective population size, N e , the detection of past changes in population size, the estimation of admixture proportions, and the analysis of local population structure, kinship, and relatedness through genotypic methods. The reasons why it may be useful to measure N e are discussed. Using genetic information it is possible to estimate N e from changes in gene frequencies between samples taken at different times, or from genotypic disequilibria in one sample, or from differences in gene frequencies between two recently descended populations. The different statistical methodologies that have been applied are discussed in some detail. The ability to detect past changes in population size may also be useful in conservation management, and some of the more recent approaches are described. Another area of concern in conservation biology is introgressive hybridisation. At the population level it is possible to infer admixture proportions by comparing the gene frequencies in the admixed and non-admixed populations, providing they are available, and the different statistical approaches to analysing this problem are reviewed. Statistical methods have also been developed to detect immigrant and hybrid individuals from their multi-locus genotype, and the main methods developed in this area are also discussed.

show abstract

A Coalescent-Based Estimator of Admixture From DNA Sequences

Cited by 17 publications

References 34 publications

Approximate Bayesian Computation Without Summary Statistics: The Case of Admixture

Approximate Bayesian Computation Without Summary Statistics: The Case of Admixture

Paleolithic Contingent in Modern Japanese: Estimation and Inference using Genome-wide Data

Conservation Genetics

Contact Info

Product

Resources

About