Another uninterpretable epigenetics study

Posted on 25 Aug, 2015 by Noah Snyder-Mackler

If you looked at your Twitter feed on Sunday you likely saw a lot of buzz about a new study that found that “Holocaust survivors trauma is passed on to children’s genes”. Many people have already taken time to blog about the issues with this study, but I wanted to ensure that the message was passed on to the molecular ecology community because I think that it is relevant to the community. So here is a quick (and by no means comprehensive) list of why you should be skeptical about the study (and, importantly, why the authors should not have been able to draw the conclusions that they drew):
Continue reading →

Posted in Uncategorized | 2 Comments

Where's your wine from?

Posted on 20 Aug, 2015 by Arun Sethuraman

Human-mediated selection of yeast cultures has played a huge role in the development of numerous unique strains of Sacchromyces cerevisiae, often attributed to production of a wide variety of wines the world over. Previous studies have indicated a single domesticated origin of S. cerevisiae, termed the “Wine-European” group, but detailed demographic history of the species has been thus far mired by insufficient sampling of wild cultures (especially from oak niches in the Mediterranean) that often coexist with domesticated strains. Almeida et al. (2015) analyze whole genome sequences of 145 strains of S. cerevisiae to understand (a) population genomic structure, and (b) ancestral demography of wild and domestic strains.

Network analysis showing geographic separation of S. cerevisiae strains. Image courtesy: Figure S1 of Almeida et al. (2015)

Network analysis showing geographic separation of S. cerevisiae strains. Image courtesy: Figure S1 of Almeida et al. (2015)


Phylogenetic network analyses indicated the positioning of Mediterranean oak strains, and commonly used wine strains in one horizontal extremity, and the other extremity occupied by North American, Asian African, and Caribbean strains, along with six Mediterranean strains. Largely geographical separations of strains were identified at K=10 ancestral subpopulations while using STRUCTURE under an admixture model. Haplotype structure indicated high degrees of shared structure between wine and Mediterranean oak strains, and complex admixture history among all sampled strains, also supported by maximum likelihood phylogenies. Diversity analyses indicate similar levels of polymorphism and diversity in the wine and Mediterranean oak strains, which are both lower than genomic diversity North American, or Asian strains. Demographic analyses of a subset of intronic regions under models of isolation, versus isolation with migration, versus population size growth post isolation using δaδi indicated better fit of the growth model. Estimates of demographic parameters under the growth model showed relatively low migration between strains, a strong bottleneck in the wine strains, with the isolation dated to around 5400-5000 BC, coincident with the first biochemical evidence of wine.

Perhaps you’d like to look at a genome list, Ms. Knope?

As in the case of crop and livestock domestication, linking wild and domesticated microbe genotypes is an essential step for understanding the roots and trajectories of man-driven artificial selection.

Reference:
Almeida, Pedro, et al. “A Population Genomics Insight into the Mediterranean Origins of Wine Yeast Domestication.” Molecular Ecology (2015). DOI: 10.1111/mec.13341

Posted in domestication, evolution, genomics, horizontal gene transfer, microbiology, Molecular Ecology, the journal, next generation sequencing, phylogenetics, phylogeography, population genetics, STRUCTURE, yeast | Tagged , , , | 1 Comment

Fossils and phylogenetics meet in the evolutionary middle

Posted on 19 Aug, 2015 by Rob Denton
Image by Matt Mechtley

Image by Matt Mechtley

…if evolutionary biologists are intent on documenting the
history of life, we need methods that can at least approximate patterns of evolution in deep time for clades without fossil information.

A scientists who wants to understand the evolutionary history of a group of organisms has some serious roadblocks. One of the most obvious of these issues: the majority of the evolutionary history is gone. Dead. Extinct! So unless you have a nice fossil record (most clades do not), you are left trying to understand millions of years of evolution by looking only on what is extant today. For the most part, this doesn’t provide the most accurate picture.
Jonathan Mitchell provides another cautionary tale about interpreting evolutionary history using only extant taxa in a new paper in Evolution. He uses birds as a study system, which have some of the most complete fossil records and phylogenies of all vertebrates, to test the two main hypotheses of when avian evolutionary radiations happened: one at the base of the tree during the Cretaceous and again when the Passeriformes first appeared.
Using morphological and phylogenetic data, Mitchell shows (as expected) that the model of diversification best supported by the data greatly depends on which taxa you include. When fossil data isn’t considered, the result tends to underestimate the diversity in the fossil record. However, the ratio of within-to-between clade differences does provide some evidence of early radiations. These combined results cause Mitchell to predict that modern radiations have washed out the signal of older radiations, providing another example of why interpretations of evolutionary history from only extant taxa can be problematic (feathered or otherwise).

The variance and range in morphology observed in the fossil assemblages from the “halfway point” of avian evolution is ~70% of the modern, which is substantially higher than models based solely on extant taxa would predict for ~50Ma. This observation, of crown Aves having achieved such ecological disparity by the Eocene, stands in stark contrast to expectations from modern data alone. None of the models based on extant taxa only consistently predicted this high level of early disparity, and the fossil-informed method was unable to predict both the high level of ancient disparity and the relatively low modern disparity simultaneously. All of these models are known as extreme simplifications, but they are commonly used to at least predict the broad contours of morphological evolution.

 
Mitchell, J. S. (2015). Extant‐only comparative methods fail to recover the disparity preserved in the bird fossil record. Evolution. DOI: 10.1111/evo.12738
 

Posted in phylogenetics | Tagged , , | 2 Comments

What's the most replicated finding in population genetics?

Posted on 18 Aug, 2015 by Jeremy Yoder

Cloning Experiments: Jess Payne

The more the merrier. (Flickr: Dan Foy)


DrugMonkey tells a tale of a specific finding in addiction research — that rats provided with an intravenous drip of cocaine solution will push a lever to self-administer the drug — which has been replicated countless times over the decades. Past the point of usefulness, you’d think. But it turns out that in all this replication, folks have turned up a lot of factors that make the replication, um, not replicate. Everything from the cocaine dose available in each infusion to whether or not the rat-handler wears a clean lab coat.
And this, as he concludes, has taught addiction researchers a lot about the mechanisms underlying a seemingly unassailable “classic” result.

I can’t speak to how many “failure to replicate” studies were discussed at conferences and less formal interactions. But given what I do know about science, I am confident that there was a little bit of everything. Probably some accusations of faking data popped up now and again. Some investigators no doubt were considered generally incompetent and others were revered (sometimes unjustifiably). No doubt. Some failures to replicate were based on ignorance or incompetence…and some were valid findings which altered the way the field looked upon prior results.
Ultimately the result was a good one. The rat IVSA model of cocaine use has proved useful to understand the neurobiology of addiction.

Science! Where you learn things even when you screw up. (Maybe especially when you screw up.)
It left me wondering, though, what the equivalent experiment or result would be for my own field, population genetics. My first thought is isolation by distance, the finding that populations distributed across a landscape will show greater genetic differentiation as the geographic distance between them grows, even if there is no meaningful difference in the environments they encounter. Testing for IBD is a terribly basic thing to do with your shiny new population genetic dataset, and it’s no surprise when it turns up — but if you don’t find it, you know something odd is going on.
Or maybe there’s a better alternative that I haven’t thought of? Submit your nominations in the comments.

Posted in population genetics | Tagged , | 4 Comments

Survival of the fittest: a marine snail toughs it out through a salty time

Posted on 17 Aug, 2015 by Melissa DeBiasse

Dendropoma and its its associated calcareous algae Neogoniolithon brassica-florida. Photo from the Mediterranean Sea climate and environmental change blog

The vermetid gastropod Dendropoma and its its associated calcareous algae Neogoniolithon brassica-florida. Photo from the Mediterranean Sea climate and environmental change blog.


For marine organisms, salinity plays an important role in determining how populations and species are distributed across time and space, particularly in the Mediterranean Sea. During the Mesozoic, about 252 to 66 million years ago, the Tethys Ocean, a body of water that would become the Mediterranean Sea, connected the Atlantic and Pacific oceans. According to Lejeusne et al., “the Mediterranean is a peculiar sea, a product of a tormented geological history, where continents collide and water masses come and go, a crossroads of biogeographical influences between cold temperate biota and subtropical species.” Continue reading →

Posted in Uncategorized | Leave a comment

Models matter when linking genetic diversity to niche model predictions

Posted on 13 Aug, 2015 by Rob Denton

Ecological niche models and the methods to create them continue to evolve. These techniques provide a tidy way to relate the distributions of taxa to environmental variables from the present, past, or future. Oh, and they are pretty too:

Image from Bram Breure


Those pretty maps of niche models assign some measure of suitability to each pixel, which indicates how close the conditions at that point are to a species’ most-optimum environmental conditions. One particular juicy prediction that results from niche modelling is that there may be a positive relationship between the “probability of occurrence” and the abundance of a species. Having this sort of forecasting power would provide some helpful inference for population demography and associated genetic diversity, right?
However, combining genetic data with ecological niche models can be a tricky business. Some of the concerns and caveats are summarized nicely in a review by Alvarado-Serrano and Knowles last year in Molecular Ecology Resources. One empirical example for these considerations appears in the same journal from Diniz-Filho and colleagues. They used 14 general methods for niche modelling with four climate data models and tested how these variations affect inferences of genetic diversity (heterozygosity measured by microsatellites) of Dipteryx alata, a widely distributed tree species in Brazil.

The correlation between He and environmental suitability would then reflect the effects of variable population size in geographical space that, under distinct environmental conditions, leads to a well-known pattern in which larger populations are able to maintain more genetic diversity.

Modelling method was the most influential factor in predicting genetic diversity, with our old friend Maxent having the highest mean correlation across the different climate models (Pearson correlation = 0.438, P = 0.037). However, the overall message is that the type of modelling methodology can have drastic effects on the correlations between predicted occurrence and genetic diversity.

Figure 2 from Diniz-Filho et al. (2015)

Figure 2 from Diniz-Filho et al. (2015)


 
As an alternative to stacking an ensemble of methods for modelling, Diniz-Filho and colleagues provide a R-script that creates distributions of correlations with random combinations of modelling method and climate data, providing the ability to simply visualize if variation in modelling method affects the ability to detect a pattern of interest (genetic diversity here, but could be something else!).
 
Cited
Alvarado‐Serrano, D. F., & Knowles, L. L. (2014). Ecological niche models in phylogeographic studies: applications, advances and precautions. Molecular Ecology Resources, 14(2), 233-248.
Diniz‐Filho, J. A. F., Rodrigues, H., Telles, M. P. D. C., Oliveira, G. D., Terribile, L. C., Soares, T. N., & Nabout, J. C. (2015). Correlation between genetic diversity and environmental suitability: taking uncertainty from ecological niche models into account. Molecular Ecology Resources, 15(5), 1059-1066.
 

Posted in methods, R | Tagged , , | Leave a comment

Inferring kinship from low coverage sequencing data

Posted on 11 Aug, 2015 by Noah Snyder-Mackler

The kinship structure of the Noble and Most Ancient House of BLACK (from the “Harry Potter” series)


Knowing the relatedness structure of your population is essential for pretty much any study. Until recently, the only way to determine the kinship structure was to have a detailed pedigree or to estimate relatedness (poorly) using microsatellites. The spurt of genomic data coming out of many non-model systems and wild populations has drastically increased the number of markers that can be used to infer kinship, which should allow for more accurate estimates of kinship. Indeed, many programs take advantage the large number of variants that can be genotyped using high throughput sequencing technology (e.g., PLINK and KING). However these programs all assume that every genotype is known with high confidence. In other words, it doesn’t take into account the uncertainty of each genotype call that inevitably comes with low coverage sequencing. A new biorxiv preprint from Mikhail Lipatov and colleagues details a method that incorporates this uncertainty to allow for more accurate inference of kinship in low coverage sequencing data.
Continue reading →

Posted in Uncategorized | 2 Comments

Notes from Edmonton and #Botany2015

Posted on 11 Aug, 2015 by Jeremy Yoder

2015.07.29 - Shaw Convention Centre, Edmonton

The Shaw Convention Centre looks sort of like a Mayan temple inside a greenhouse, which is sort of appropriate for botanists, I guess? (Flickr: jby)


As noted previously, I broke with my usual habit and skipped the Evolution meetings this year. Instead, I attended Botany 2015, a joint meeting of multiple U.S. and Canadian plant-focused scholarly societies held in Edmonton, Alberta — I’d never been to Botany, but I do think of myself as a plant-focused evolutionary biologist, and Jim Leekens-Mack was kind enough to invite me to co-manage a colloquium on intimate mutualisms featuring folks who study specialized plants-pollinators interactions, legume-rhizobium symbiosis, and interactions between plants and endophytes featuring every scientific approach from field experiments to phylogenetics, population genetics to experimental evolution, gas chromatography to computer simulations. That colloquium was a great success, if I do say so myself, and the conference as a whole was fun and informative.
Continue reading →

Posted in bioinformatics, conferences, genomics, natural history, phylogenetics, plants, population genetics | Tagged | 1 Comment

Selection scans, and the genomics of adaptive/maladaptive introgression

Posted on 6 Aug, 2015 by Arun Sethuraman

Natural selection, and the adaptive evolution of hybrid reproductive incompatibilities post divergence are known to be major drivers of speciation. At the phenotype level, these manifest as fitness differences between introgressing populations. At the genomic level, speciation “genes” or “islands” are often identified via quantifying these incompatibilities as barriers to migration, and hybridization (see my previous post for an account of “islands” and secondary contact). Three recent publications discuss these incompatibilities in different contexts (1) genome wide scans of selection (Haasl and Peyseur 2015), (2) when incompatible gene flow post divergence does not increase selection (Rolshausen et al. 2015), and generally, (3) the genomic architecture of incompatibilities under secondary contact (Lindtke and Buerkle 2015).
1) Fifteen years of genome-wide scans for selection; trends, lessons, and unaddressed genetic sources of complication – Haasl and Peyseur, 2015, Molecular Ecology
In this detailed meta-analysis and review of > 100 studies across a host of species, Haasl and Peyseur (a) discuss evolutionary processes – mutational rate differences among marker types, recombination rate variation across the genome, confounding effects of reduction in diversity due to linked/background selection and positive selection, selection on polygenic traits that often complicate, confound, or complement the identification of targets of selection, (b) report observed trends in GWSS across taxa –predominantly human, use Fst outlier detection, and strongly biased towards the detection of positive, and directional selection, and (c) offer recommendations for best practices for researchers working on GWSS – the use of recombination and mutation rates, and the use of simulations eg. under an ABC framework, to test hypotheses of demographic histories.

The genome provides an organic record of evolution that is frequently likened to a palimpsest – a writing medium that is recycled, continuously written over, and reoriented so as to partially or wholly obscure older text. By this metaphor, chromosomes are the parchment, and DNA sequence the text.

2) When maladaptive gene flow does not increase selection – Rolshausen et al. (2015), Evolution

Morphological differences in river (inlet) and lake (outlet) threespine stickleback. Image courtesy: http://www.registrelep-sararegistry.gc.ca/default.asp?lang=En&n=708618BB-1


Rolshausen et al. report results of a long-term mark-recapture observational study on inlet (river/stream) and outlet (lake) populations of the threespine stickleback (Gasterosteus aculeatus) in British Columbia which have large levels of previously observed unidirectional (inlet-outlet) gene flow, and phenotypic differences in life histories. They monitor individual survival of ~4000 fish over long periods of time, with temporal replicates, and use logistic regression models to estimate selection coefficients. Key findings of the study include (a) increased winter mortality in outlet (lake) fish, (b) smaller average body lengths in inlet fish (river) compared to outlet fish (lake), (c) positive directional selection for deeper bodies in inlet fish, and no selection for body size in outlet fish, with seasonally varying levels of intensity – contrary to theory that suggests stronger selection in outlet fish (receiving maladaptive gene flow from the inlet).

Of particular interest is the novel idea that high gene flow can causally reduce selection by broadening the fitness function – a result we demonstrated by means of a general population genetic model.

Classic DMI’s, indicating allelic incompatibility of hybrids. Image courtesy: http://www.nature.com/scitable/content/dobzhansky-muller-model-of-hybrid-incompatibility-7883


3) The genetic architecture of hybrid incompatibilities and their effect on barriers to introgression in secondary contact – Lindtke and Buerkle, 2015, Evolution
Dobzhansky-Muller Incompatibilities, or DMI’s are often broken down in populations with few viable hybrids under secondary contact, but yet suffer reduced fitness. Lindtke and Buerkle use simulations of whole genomes in contact zones, to explore the classic DMI model, and an alternative model of coadaptation in genomes due to a ‘pathway’ model (wherein incompatibilities arise not from alleles, but from disruptions of pathway interactions). Quantifying Fst differences between diverged populations, they report (a) strong selection, and low migration showed the classic signatures of DMIs, and vice versa – weaker selection, and high migration indicated breakdown of DMI’s due to enabling interspecific recombination, and (b) genome-wide reduction in Fst under the pathway model, and declined linearly with time, with more introgression, stabilizing after a few generations, and (c) strong influence of demography – particularly migration rates on the genomic outcomes of hybridization.

…This highlights the potential contribution of intra-genomic interactions to speciation with gene flow and suggests the value of a broader set of epistatic models in speciation research.

References
Haasl, Ryan J., and Bret A. Payseur. “Fifteen years of genome‐wide scans for selection: trends, lessons, and unaddressed genetic sources of complication.”Molecular Ecology (2015). DOI:10.1111/mec.13339
Rolshausen, Gregor, et al. “When maladaptive gene flow does not increase selection.” Evolution (2015).DOI: 10.1111/evo.12739
Lindtke, Dorothea, and C. Alex Buerkle. “The genetic architecture of hybrid incompatibilities and their effect on barriers to introgression in secondary contact.” Evolution (2015). DOI: 10.1111/evo.12725

Posted in adaptation, Coevolution, evolution, genomics, Molecular Ecology, the journal, mutation, natural history, population genetics, selection, speciation, theory | Tagged , , , , | 3 Comments

Should we use Mantel tests in molecular ecology?

Posted on 5 Aug, 2015 by Rob Denton


No. Stop.
At least that is the message from a new publication in Methods in Ecology and Evolution by Pierre Legendre and colleagues (pay-walled, but I found a pdf here).

Mantel tests should simply not be used to test hypotheses that concern the raw data from which dissimilarity matrices can be computed or to control for spatial structures in tests of relationships between two autocorrelated data sets.

Continue reading →

Posted in Uncategorized | Tagged , , | 2 Comments