Genomic analyses of ancestry of Caribbean populations

Blog author Rajiv McCoy is a graduate student in the lab of Dmitri Petrov.

Blog author Rajiv McCoy is a graduate student in the lab of Dmitri Petrov.

In the Author Summary of their paper, “Reconstructing the Population Genetic History of the Caribbean”, Andrés Moreno-Estrada and colleagues point out that Latinos are often falsely depicted as a homogeneous ethnic or cultural group.  In reality, however, Latinos, including inhabitants of the Caribbean basin, represent a diverse mixture of previously separate human populations, such as indigenous groups, European colonists, and West Africans brought over during the Atlantic slave trade.  This mixing process, which geneticists call “admixture”, left a distinct footprint on genetic variation within and between Caribbean populations.  By surveying genotypes of 330 Caribbean individuals and comparing to a database of variation from more than 3000 individuals from European, African, and Native American populations, Moreno et al., explore the genomic outcomes of this complex admixture process and reveal intriguing demographic patterns that could not be obtained from the historical record alone. The paper, featured in the latest edition of PLOS Genetics, represents a collaborative project with co-senior authorship by Stanford CEHG professor Carlos Bustamante and Professor Eden Martin from the University of Miami Miller School of Medicine.

Reconstructing the demographic history of admixed populations

Because parental DNA is only moderately shuffled before being incorporated into gametes (the process of meiotic recombination), admixture results in discrete genomic segments that can be traced to a particular ancestral population.  In early generations after the onset of admixture, these segments are large.  However, after many generations, segments will be quite small.  By investigating the distribution of sizes of these ancestry “tracts”, Moreno and colleagues inferred the timing of various waves of migration and admixture.  For Caribbean Island populations, they infer that European gene flow first occurred ~16-17 generations ago, which matches very closely to the historical record of ~500 years, assuming ~30 years per generation.  In contrast, for neighboring mainland populations from Colombia and Honduras, they find that European gene flow occurred in waves, starting more recently (~14 generations ago).

Identifying sub-continental ancestry of admixed individuals

Those familiar with human population genetics will recognize principal component analysis (PCA), which transforms a matrix of correlated observed genotypes into a set of uncorrelated variables where the first component explains the most possible variance, the second variable explains the second most variance, and so on.  Individuals’ transformed genotypes can be plotted on the first two principle components, and when performed on a worldwide scale, distinct clusters appear which represent populations of ancestry.  On conventional PCA plots, admixed individuals fall between their different ancestral populations, as they possess sets of genotypes diagnostic of multiple ancestral groups.  As virtually all Caribbean individuals are admixed to some degree, this pattern is apparent for Caribbean populations (see Figure 1B from the paper, reproduced below).

Fig1B

While interesting, this means that the sub-continental ancestry of these admixed individuals is difficult to ascertain.  An individual may want to know which Native American, West African, and European populations contribute to his or her ancestry, and this analysis does not have sufficient resolution to answer these questions.

Moreno and colleagues therefore devised a new version of PCA called ancestry-specific PCA (ASPCA), which extracts genomic segments assigned to Native American, West African, and European ancestry, then analyzes these segments separately, dealing with the large proportions of missing data that result.  In the case of Native American ASPCA, they observe two overlapping clusters.  The first represents mostly Colombians and Hondurans, who cluster most closely with indigenous groups from Western Colombia and Central America and have a greater overall proportion of Native American ancestry.  The second cluster represents mostly Cubans, Dominicans, and Puerto Ricans, who cluster most closely with Eastern Colombian and Amazonian indigenous groups.  This makes sense in light of the fact that Amazonian populations from the Lower Orinoco Valley settled on rivers and streams, which could have facilitated their migration.  Because indigenous ancestry proportions were relatively consistent and closely clustered across different Caribbean Islands, the authors posit that there was a single pulse of expansion of Amazonian natives across the Caribbean prior to European arrival, along with gene flow among the islands.

In the case of European ASPCA, Moreno et al. found that Caribbean samples clustered closest to, but clearly distinct from, present day individuals from the Iberian Peninsula in Southern Europe.  In fact, the differentiation between this “Latino-specific component” and Southern Europe is at least as great as the differentiation between Northern and Southern Europe.  The authors hypothesize that this is due to very small population sizes among European colonists, which would have introduced noise into patterns of genomic variation through the process of random genetic drift.

Finally, the authors demonstrate that Caribbean populations have a higher proportion of African ancestry compared to mainland American populations, a result of admixture during and after the Atlantic slave trade.  Surprisingly, the authors found that all samples tightly clustered with present day Yoruba samples from Nigeria rather than being dispersed throughout West Africa.  However, because other analyses suggested that there might have been two major waves of migration from West Africa, the authors decided to analyze “old” and “young” blocks of African ancestry separately.  This analysis revealed that “older” segments are primarily derived from groups from the Senegambia region of Northwest Africa, while “younger” segments likely trace to groups from the Gulf of Guinea and Equatorial West Africa (including the Yoruba).

Conclusions and perspectives

This groundbreaking study has immediate implications for the field of personalized medicine, especially due to the discovery of a distinct Latino-specific component of European ancestry.  The hypothesis that European colonists underwent a demographic bottleneck (a process termed the “founder effect”) has expected consequences for the frequency of damaging mutations contributing to genetic disease. The observation of extensive genetic differences among Caribbean populations also argues for more such studies characterizing genetic variation on a smaller geographic scale. The newly developed ASPCA method will surely be valuable for other admixed populations.  In addition to medical implications, studies such as this help dispel simplistic notions of race and ethnicity and inform cultural identities based on unique and complex demographic history.

Citation: Moreno-Estrada A, Gravel S, Zakharia F, McCauley JL, Byrnes JK, et al. (2013) Reconstructing the Population Genetic History of the Caribbean. PLoS Genet 9(11): e1003925. doi:10.1371/journal.pgen.1003925

Paper author Andres Moreno-Estrada is a research associate in the lab of Carlos Bustamante.

Paper author Andrés Moreno-Estrada is a research associate in the lab of Carlos Bustamante.

How recombination and changing environments affect new mutations

Blog author: David Lawrie was a graduate student in Dmitri Petrov’s lab. He is now a postdoc at USC.

I recently sat down with Oana Carja, a graduate student with Marc Feldman, to discuss her paper published in the journal of Theoretical Population Biology entitled “Evolution with stochastic fitnesses: A role for recombination”. In it, the authors Oana Carja, Uri Liberman, and Marcus Feldman explore when a new mutation can invade an infinite, randomly mating population that experiences temporal fluctuations in selection.

The one locus case

This work builds off of previous research in the field on how the fluctuations in fitness over time (i.e., increased variance of fitness) affect the invasion dynamics of a mutation at a single locus. For a single locus, it has been shown that the geometric mean of the fitness of the allele over time determines the ability of an allele to invade a population. This effect is known as the geometric mean principle. Fluctuations in fitness increase the variance and therefore decrease the geometric mean fitness. The variance of the fitness of the allele over time thus greatly impacts the ability of that allele to invade a population.

What if there are two loci?

In investigating a two locus model, the researchers split the loci by their effect on the temporally-varying fitness: one locus only affects the mean, while the other controls the variance. The authors demonstrate through theory and simulation that:

1)    allowing for recombination between the two loci increases the threshold for the combined fitness of the two mutant alleles to invade the population beyond the geometric mean (see figure).

2)    periodic oscillations in the fitness of the alleles over time lead to higher fitness thresholds for invasion over completely random fluctuations (see figure).

3)    edge case scenarios allow for the maintenance of polymorphisms in the population despite clear selective advantages of a subset of allelic combinations.

Temporally changing environments and recombination thus make it overall more difficult for new alleles to invade a population.

Invasibility thresholds as a function of recombination rate. Recombination makes it more difficult for new alleles to invade a population.

Invasibility thresholds as a function of recombination rate. If there is no recombination (the left-most edge of the figure), the geometric mean of the pair of new alleles needs to be higher than 0.5 to allow for invasion, because the resident alleles’ geometric mean fitness is set to 0.5. However, as recombination between the two loci increases, the geometric mean needed for invasion increases rapidly. If there is free recombination (r = 0.5) then  invasion can only happen if the new alleles’ geometric mean fitness is twice the resident alleles’ geometric mean fitness (light grey area). If the environment is changing periodically, it is even harder for new alleles to invade a population (dark grey area).

The evolution of models of evolution

This work is important for addressing the evolutionary dynamics of loci controlling phenotypic variance – in this case, controlling the ability of a phenotype to maintain its fitness even if the environment is variable. Most environments undergo significant temporal shifts from the simple changing of the seasons to larger scale weather changes such as El Niño and climate change, in which species must survive and thrive. For organisms in the wild, many alleles that confer a benefit in one environment will be deleterious when the environment and selective pressures change. There may be modifier-loci which buffer the fitness of those loci in the face of changing environments. Such modifier-loci have been recently found in GWAS studies and may be important for overall phenotypic variance. Thus modeling the patterns of evolution for multiple loci in temporarily varying environments is a key component to advancing our understanding of the patterns found in nature.

Future work

Epigenetic modifiers are a hot area of research and one potential biological mechanism to control phenotypic variance. The evolution of such epigenetic regulation is a particular research interest of Oana. Future work will continue to explore the evolutionary dynamics of epigenetic regulation and focus on applying the above results to finite populations.

Paper author: Oana Carja is a graduate student with Marc Feldman

Reference

Oana Carja, Uri Liberman, Marcus W. Feldman, Evolution with stochastic fitnesses: A role for recombination, Theoretical Population Biology, Volume 86, June 2013, Pages 29-42, ISSN 0040-5809.

Missing the forest for the trees: How frequent adaptation can confound its own inference

FabianStaubach

Blog author: Fabian Staubach was a postdoc in Dmitri Petrov’s lab and is now an assistant professor in Freiburg, Germany.

This post was written by Fabian Staubach. 

The neutral theory of molecular evolution assumes that adaptation is rare and that the effect of adaptation on linked variation, the so-called hitchhiking effect, typically has only little influence on the dynamics of molecular genetic variation. Because of this assumption, it is widely assumed that in most natural populations, hitchhiking can be neglected, or at least reasonably well approximated by the introduction of effective parameters, such as an effective population size. But if molecular adaptation is in fact common, then the assumption may be violated, and we should worry whether population genetic methods and estimates of evolutionary parameters obtained from them are robust to frequent hitchhiking.

In their paper “Frequent adaptation and the McDonald-Kreitman test” (PNAS, 2013), Philipp Messer and Dmitri Petrov investigate this question for one of the key population genetic methods — the McDonald-Kreitman (MK) test. This test forms the basis of most commonly used approaches to measure the rate of adaptation from population genomic data and has been used to argue that in some organisms, such as Drosophila, the rate of adaptation is surprisingly high.

The MK test can substantially underestimate the true rate of adaptation

Messer and Petrov employ their powerful forward simulation software, SLiM (see here), to simulate the evolution of entire chromosomes under a range of parameter values relevant to humans and other organisms, and apply various forms of the MK test to the population genomic data resulting from their simulations. They then study how accurately these methods re-infer the true evolutionary parameters in the simulations. Strikingly, they find that the MK test can substantially underestimate the true rate of adaptation. For instance, they present scenarios where 40% of the amino acid changing substitutions were in fact strongly adaptive in the simulations, while other population parameters resembled those commonly inferred for human evolution, yet the standard MK estimates yield that none of these substitutions were actually adaptive. Fortunately, Messer and Petrov propose a way to avoid these problems by using a simple, asymptotic extension of the MK test.

Figure: Illustration of the asymptotic MK estimation of the rate of adaptive substitutions : The standard MK approach assumes that all polymorphisms (non-synonymous and synonymous) are neutral. This assumption is likely violated for low frequency polymorphisms, as some of these are likely to be deleterious. The assumption should hold for very high frequency polymorphisms, because they are very unlikely to be deleterious. The asymptotic MK approach uses this fact by looking at the estimated rate from different frequency classes of alleles, and extrapolating to x=1, where the rate is expected to have asymptoted.  

The bigger claim of this straightforward and easy-to-read paper is that the effects of linked selection cannot be simply swept under the rug by introducing effective parameters, such as effective population size or effective strength of selection, and then using these effective parameters in formulae derived from the diffusion approximation under the assumption of free recombination.

Quantifying known biases

Surely, this paper will ruffle some feathers. Some people will argue that these problems have been know for a while in theory. Yet despite this, the vast majority of studies that continue to appear in the literature still pay only cursory lip service, if anything, to these issues. Presumably, this is because it is not well understood analytically to what extent linkage effects affect population genetic estimates, and Messer and Petrov therefore do an important job in quantifying these biases. Hopefully this will help focus the community’s attention to spend some time figuring out how to modify commonly used approaches to place them on a more solid foundation.

Citation: Messer, P. W., & Petrov, D. A. (2013). Frequent adaptation and the McDonald-Kreitman test. Proceedings of the National Academy of Sciences of the United States of America, 110(21), 8615–20. doi:10.1073/pnas.1220835110

PhilippMesser

Paper author: Philipp Messer is a research associate in Dmitri Petrov’s lab at Stanford, where he studies the population genetics of adaptation using theoretical and computational approaches in concert with the analysis of large-scale population genomic data.