The reliability of published scientific papers has been the topic of much recent discussion, notably in the biomedical sciences [1]. Although small sample size is regularly pointed as one of the culprits, big data can also be a concern. The advent of high-throughput sequencing, and the processing of sequence data by opaque bioinformatics workflows, mean that sequences with often high error rates are produced, and that exact but slow analyses are not feasible.
The troubles with bioinformatics arise from the increased complexity of the tools used by scientists, and from the lack of incentives and/or skills from authors (but also reviewers and editors) to make sure of the quality of those tools. As a much discussed example, a bug in the widely used PLINK software [2] has been pointed as the explanation [3] for incorrect inference of selection for increased height in European Human populations [4].
High-throughput sequencing often generates high rates of genotyping errors, so that the development of bioinformatics tools to assess the quality of data and correct them is a major issue. The work of Gillingham et al. [5] contributes to the latter goal. In this work, the authors propose a new bioinformatics workflow (ACACIA) for performing genotyping analysis of multigene complexes, such as self-incompatibility genes in plants, major histocompatibility genes (MHC) in vertebrates, and homeobox genes in animals, which are particularly challenging to genotype in non-model organisms. PCR and sequencing of multigene families generate artefacts, hence spurious alleles. A key to Gillingham et al.‘ s method is to call candidate genes based on Oligotyping, a software pipeline originally conceived for identifying variants from microbiome 16S rRNA amplicons [6]. This allows to reduce the number of false positives and the number of dropout alleles, compared to previous workflows.
This method is not based on an explicit probability model, and thus it is not conceived to provide a control of the rate of errors as, say, a valid confidence interval should (a confidence interval with coverage c for a parameter should contain the parameter with probability c, so the error rate 1- c is known and controlled by the user who selects the value of c). However, the authors suggest a method to adapt the settings of ACACIA to each application.
To compare and validate the new workflow, the authors have constructed new sets of genotypes representing different extents copy number variation, using already known genotypes from chicken MHC. In such conditions, it was possible to assess how many alleles are not detected and what is the rate of false positives. Gillingham et al. additionally investigated the effect of using non-optimal primers. They found better performance of ACACIA compared to a preexisting pipeline, AmpliSAS [7], for optimal settings of both methods. However, they do not claim that ACACIA will always be better than AmpliSAS. Rather, they warn against the common practice of using the default settings of the latter pipeline. Altogether, this work and the ACACIA workflow should allow for better ascertainment of genotypes from multigene families.

References

[1] Ioannidis, J. P. A, Greenland, S., Hlatky, M. A., Khoury, M. J., Macleod, M. R., Moher, D., Schulz, K. F. and Tibshirani, R. (2014) Increasing value and reducing waste in research design, conduct, and analysis. The Lancet, 383, 166-175. doi: 10.1016/S0140-6736(13)62227-8
[2] Chang, C. C., Chow, C. C., Tellier, L. C. A. M., Vattikuti, S., Purcell, S. M. and Lee, J. J. (2015) Second-generation PLINK: rising to the challenge of larger and richer datasets. GigaScience, 4, 7, s13742-015-0047-8. doi: 10.1186/s13742-015-0047-8
[3] Robinson, M. R. and Visscher, P. (2018) Corrected sibling GWAS data release from Robinson et al. http://cnsgenomics.com/data.html
[4] Field, Y., Boyle, E. A., Telis, N., Gao, Z., Gaulton, K. J., Golan, D., Yengo, L., Rocheleau, G., Froguel, P., McCarthy, M.I . and Pritchard J. K. (2016) Detection of human adaptation during the past 2000 years. Science, 354(6313), 760-764. doi: 10.1126/science.aag0776
[5] Gillingham, M. A. F., Montero, B. K., Wihelm, K., Grudzus, K., Sommer, S. and Santos P. S. C. (2020) A novel workflow to improve multi-locus genotyping of wildlife species: an experimental set-up with a known model system. bioRxiv 638288, ver. 3 peer-reviewed and recommended by Peer Community In Evolutionary Biology. doi: 10.1101/638288
[6] Eren, A. M., Maignien, L., Sul, W. J., Murphy, L. G., Grim, S. L., Morrison, H. G., and Sogin, M.L. (2013) Oligotyping: differentiating between closely related microbial taxa using 16S rRNA gene data. Methods in Ecology and Evolution 4(12), 1111-1119. doi: 10.1111/2041-210X.12114
[7] Sebastian, A., Herdegen, M., Migalska, M. and Radwan, J. (2016) AMPLISAS: a web server for multilocus genotyping using next‐generation amplicon sequencing data. Mol Ecol Resour, 16, 498-510. doi: 10.1111/1755-0998.12453

Phylogenetic trees can be used to extract information about the process of diversification that has generated them. The most common approach to conduct this inference is to rely on a likelihood, defined here as the probability of generating a dated tree T given a diversification model (e.g. a birth-death model), and then use standard maximum likelihood. This idea has been explored extensively in the context of the so-called diversification studies, with many variants for the models and for the questions being asked (diversification rates shifting at certain time points or in the ancestors of particular subclades, trait-dependent diversification rates, etc).
However, all this assumes that the dated tree T is known without error. In practice, trees (that is, both the tree topology and the divergence times) are inferred based on DNA sequences, possibly combined with fossil information for calibrating and informing the divergence times. Molecular dating is a delicate exercise, however, and much more so in fact than reconstructing the tree topology. In particular, a mis-specificied model for the relaxed molecular clock, or a mis-specifiied prior, can have a substantial impact on the estimation of divergence dates - which in turn could severely mislead the inference about the underlying diversification process. This thus raises the following question: would that be possible to conduct inference and testing of diversification models without having to go through the dangerous step of molecular dating?
In his article ""Probabilities of tree topologies with temporal constraints and diversification shifts"" [1], Gilles Didier introduces a recursive method for computing the probability of a tree topology under some diversification model of interest, without knowledge of the exact dates, but only interval constraints on the dates of some of the nodes of the tree. Such interval constraints, which are derived from fossil knowledge, are typically used for molecular dating: they provide the calibrations for the relaxed clock analysis. Thus, what is essentially proposed by Gilles Didier is to use them in combination with the tree topology only, thus bypassing the need to estimates divergence times first, before fitting a diversification model to a phylogenetic tree.
This article, which is primarily a mathematical and algorithmic contribution, is then complemented with several applications: testing for a diversification shift in a given subclade of the phylogeny, just based on the (undated) tree topology, with interval constraints on some of its internal nodes; but also, computing the age distribution of each node and sampling on the joint distribution on node ages, conditional on the interval constraints. The test for the presence of a diversification shift is particularly interesting: an application to simulated data (and without any interval constraint in that case) suggests that the method based on the undated tree performs about as well as the classical method based on a dated tree, and this, even granting the classical approach a perfect knowledge of the dates - given that, in practice, one in fact relies on potentially biased estimates. Finally, an application to a well-known example (rate shifts in cetacean phylogeny) is presented.
This article thus represents a particularly meaningful contribution to the methodology for diversification studies; but also, for molecular dating itself: it is a well known problem in molecular dating that computing and sampling from the conditional distributions on node ages, given fossil constraints, and more generally understanding and visualizing how interval constraints on some nodes of the tree impact the distribution at other nodes, is a particularly difficult exercise. For that reason, the algorithmic routines presented in the present article will be useful in this context as well.

References

[1] Didier, G. (2020) Probabilities of tree topologies with temporal constraints and diversification shifts. bioRxiv, 376756, ver. 4 peer-reviewed and recommended by PCI Evolutionary Biology. doi: 10.1101/376756

Ecological specialisation, especially among parasites infecting a set of host species, is ubiquitous in nature. Host specialisation can be understood as resulting from trade-offs in parasite infectivity, virulence and growth. However, it is not well understood how variation in these trade-offs shapes the overall fitness trade-off a parasite faces when adapting to multiple hosts. For instance, it is not clear whether a strong trade-off in one fitness component may sufficiently constrain the evolution of a generalist parasite despite weak trade-offs in other components. A second mechanism explaining variation in specialisation among species is habitat availability and quality. Rare habitats or habitats that act as ecological sinks will not allow a species to persist and adapt, preventing a generalist phenotype to evolve. Understanding the prevalence of those mechanisms in natural systems is crucial to understand the emergence and maintenance of host specialisation, and biodiversity in general.
In their study "Trait-specific trade-offs prevent niche expansion in two parasites", Lievens et al. [1] report the results of an evolution experiment involving two parasitic microsporidians, Anostracospora rigaudi and Enterocytospora artemiae, infecting two sympatric species of brine shrimp, Artemia franciscana and Artemia parthenogenetica. The two parasites were originally specialised on their primary host: A. rigaudi on A. parthenogenetica and E. artemiae on A. franciscana, although they encounter both species in the wild but at different rates. After passaging each parasite on each single host and on both hosts alternatively, Lievens et al. asked how host specialisation evolved. They found no change in specialisation at the fitness level in A. rigaudi in either treatment, while E. artemiae became more of a generalist after having been exposed to its secondary host, A. parthenogenetica. The most interesting part of the study is the decomposition of the fitness trade-off into its underlying trade-offs in spore production, infectivity and virulence. Both species remained specialised for spore production on their primary host, interpreted as caused by a strong trade-off between hosts preventing improvements on the secondary host. A. rigaudi evolved reduced virulence on its primary host without changes in the overall fitness trad-off, while E. artemiae evolved higher infectivity on its secondary host making it a more generalist parasite and revealing a weak trade-off for this trait and for fitness. Nevertheless, both parasites retained higher fitness on their primary host because of the lack of an evolutionary response in spore production.
This study made two important points. First, it showed that despite apparent strong trade-off in spore production, a weak trade-off in infectivity allowed E. artemiae to become less specialised. In contrast, A. rigaudi remained specialised, presumably because the strong trade-off in spore production was the overriding factor. The fitness trade-off that results from the superposition of multiple underlying trade-offs is thus difficult to predict, yet crucial to understand potential evolutionary outcomes. A second insight is related to the ecological context of the evolution of specialisation. The results showed that E. artemiae should be less specialised than observed, which points to a role played by source-sink dynamics on A. parthenogenetica in the wild. The experimental approach of Lievens et al. thus allowed them to nicely disentangle the various sources of constraints on the evolution of host adaptation in the Artemia system.

References

[1] Lievens, E.J.P., Michalakis, Y. and Lenormand, T. (2019). Trait-specific trade-offs prevent niche expansion in two parasites. bioRxiv, 621581, ver. 4 peer-reviewed and recommended by PCI Evolutionary Biology. doi: 10.1101/621581

It is well known that the rearing environment has strong effects on life history and fitness traits of organisms. Microbes are part of every environment and as such likely contribute to such environmental effects. Gut bacteria are a special type of microbe that most animals harbor, and as such they are part of most animals’ environment. Such microbial symbionts therefore likely contribute to local adaptation [1]. The main question underlying the laboratory study by Guilhot et al. [2] was: How much do particular gut bacteria affect the organismal phenotype, in terms of life history and larval foraging traits, of the fruit fly Drosophila melanogaster, a common laboratory model species in biology?
To investigate the above question, the authors isolated 4 taxa of bacteria from the gut of a (randomly picked) Drosophila melanogaster lab strain, and subsequently let Drosophila melanogaster eggs and larvae (stemming from their own, different lab strain) develop both in the typical artificial laboratory medium as well as in grapes, a natural “new” habitat for Drosophila larvae, inoculated with theses bacteria, singly and in combination, also including a bacteria-free control. By investigating various relevant developmental and size traits, the authors found that adding particularly Enterobacteria had some visible effects on several traits, both upward (indicting improvement) and downward (being detrimental) (with three other types of bacteria showing only minor or even no effects). In general, the grape medium reduced performance relative to the standard lab medium. Strongest interactive effects occurred for development time and body size, together making up growth plasticity [3], with lesser such effects on some related behavioral (feeding) traits (Figs. 2,3).
The study premise is interesting, its general objectives are clearly laid out, and the practical work was conducted correctly as far as I can evaluate. The study remains largely descriptive in that no particular a priori hypotheses or predictions in relation to the specific bacteria isolated were formulated, not least because the bacteria were necessarily somewhat arbitrarily chosen and there were apparently no prior studies from which to derive concrete predictions. Overall, the results of this study should be of interest to the community of evolutionary ecologists, especially those working on nutritional and microbiome effects on animal life histories. I consider this work to be primarily ecological, with limited evolutionary content (e.g. no genetics) though some evolutionary implications, as mentioned in the paper’s Conclusions. So this paper would best fit in a microbial or physiological ecology outlet/journal.
The inclusion of a natural medium (grapes) must be commended because this permits inferences and conclusions for at least one natural environment, whereas inferences drawn from laboratory studies in the artificial medium that most Drosophila researchers seem to use are typically limited. Unsurprisingly perhaps, the study showed that Drosophila melanogaster fared generally better in the artificial than the chosen natural medium (grape). Crucially, however, the bacterial symbionts modified both media differentially. Although common bacterial taxa were chosen, the particular bacteria isolated and used remain arbitrary, as there are many. I note that the main and strongest interactive effects between medium and bacterial type are apparent for the Enterobacteria, and they probably also strongly, if not exclusively, mediate the overall effect of the bacterial mixture.
While these specific data are novel, they are not very surprising. If we grow animals in different environments we can expect some detectable effects of these environments, including the bacterial (microbiome) environment, on the hosts life history. The standard and predicted [4] life history response of Drosophila melanogaster (but not all insects [3]) facing stressful nutritional environments, as apparently created by the Enterobacteria, is to extend development but come out smaller in the end. This is what happened here for the laboratory medium ([2]: Fig. 5). The biological interpretation is that individuals have more trouble ingesting and/or digesting the nutrients available (thus prolonging their foraging period and development), yet cannot convert the nutrients effectively into body size increments (hence emerging smaller). This is what the authors here refer to as developmental plasticity, which is ultimately nutritionally mediated. However, interestingly, a signal in the opposite direction was indicated for the bacterial mixture in the grape medium (flies emerging larger after accelerated development: Fig. 5), suggesting some positive effects on growth rate of the natural medium, perhaps related to grapes being a limited resource that needs to be escaped quickly [3]? The reversal of sexual size dimorphism across bacterial treatments in the grape environment detectable in Fig. 4 is interesting, too, though I don’t understand why this happens, and this is not discussed.
In general, more encompassing and increased questions in this context to be researched in the future could be: 1) are these effects predictable (not (yet) at this point, or so it seems); and 2) how strong are these environmental bacterial effects relative to other, more standard effects (e.g. relative to genetic variation, population variation, etc., or relative to other types of environmental effects like, say, temperature)? (3) It could further be asked why not natural but laboratory populations of Drosophila were used for this experiment, if the aim was to draw inferences for the wild situation. (4) Although Genotype x Environment effects are invoked in the Discussion, they were not tested here, lacking genetically different Drosophila families or populations. From an evolutionary standpoint, I consider this the greatest weakness of the study. I was also not too thrilled by the particular statistical analyses employed, though this ultimately does not negate the results. Nevertheless, this work is a good start in this huge field investigating the microbiome. In conclusion, I can recommend this paper after review by PCI Evol Biol.

References

[1] Kawecki, T. J. and Ebert, D. (2004) Conceptual issues in local adaptation. Ecology Letters 7: 1225-1241. doi: 10.1111/j.1461-0248.2004.00684.x
[2] Guilhot, R., Rombaut, A., Xuéreb, A., Howell, K. and Fellous, S. (2019). Environmental specificity in Drosophila-bacteria symbiosis affects host developmental plasticity. BioRxiv, 717702, v3 peer-reviewed and recommended by PCI Evolutionary Biology. doi: 10.1101/717702
[3] Blanckenhorn, W.U. (1999) Different growth responses to temperature and resource limitation in three fly species with similar life histories. Evolutionary Ecology 13: 395-409. doi: 10.1023/A:1006741222586
[4] Stearns, S. C. and Koella, J. (1986) The evolution of phenotypic plasticity in life history traits: predictions of reaction norms for age and size at maturity. Evolution 40: 893-914. doi: 10.1111/j.1558-5646.1986.tb00560.x

In partially clonal organisms, genetic markers are often used to characterize the genotypic diversity of populations and infer thereof the relative importance of clonal versus sexual reproduction. Most studies report a measure of genotypic diversity based on a ratio, R, of the number of distinct multilocus genotypes over the sample size, and qualitatively interpret high / low R as indicating the prevalence of sexual / clonal reproduction. However, a theoretical framework allowing to quantify the relative rates of clonal versus sexual reproduction from genotypic diversity is still lacking, except using temporal sampling. Moreover, R is intrinsically highly dependent on sample size and sample design, while alternative measures of genotypic diversity are more robust to sample size, like D*, which is equivalent to the Gini-Simpson diversity index applied to multilocus genotypes. Another potential indicator of reproductive strategies is the inbreeding coefficient, Fis, because population genetics theory predicts that clonal reproduction should lead to negative Fis, at least when the sexual reproduction component occurs through random mating. Taking advantage of this prediction, Arnaud-Haond et al. [1] reanalysed genetic data from 165 populations of four partially clonal seagrass species sampled in a standardized way. They found positive correlations between Fis and both R and D* within each species, reflecting variation in the relative rates of sexual versus clonal reproduction among populations. Moreover, the differences of mean genotypic diversity and Fis values among species were also consistent with their known differences in reproductive strategies. Arnaud-Haond et al. [1] also conclude that previous works based on the interpretation of R generally lead to underestimate the prevalence of clonality in seagrasses. Arnaud-Haond et al. [1] confirm experimentally that Fis merits to be interpreted more properly than usually done when inferring rates of clonal reproduction from population genetics data of species reproducing both sexually and clonally. An advantage of Fis is that it is much less affected by sample size than R, and thus should be more reliable when comparing studies differing in sample design. Hence, when the rate of clonal reproduction becomes significant, we expect Fis < 0 and D* < 1. I expect these two indicators of clonality to be complementary because they rely on different consequences of clonality on pattern of genetic variation. Nevertheless, both measures can be affected by other factors. For example, null alleles, selfing or biparental inbreeding can pull Fis upwards, potentially eliminating the signature of clonal reproduction. Similarly, D* (and other measures of genotypic diversity) can be low because the polymorphism of the genetic markers used is too limited or because sexual reproduction often occurs through selfing, eventually resulting in highly similar homozygous genotypes.
The work of Arnaud-Haond et al. [1] shows that the populations genetics of partially clonal organisms should be better studied, an endeavour encompassed in a companion paper using numerical simulations [2]. A further step that remains to be accomplished is to build a mathematical framework for developing estimators of rates of clonal versus sexual reproduction based on genotypic diversity.

References

[1] Arnaud-Haond, S., Stoeckel, S., and Bailleul, D. (2019). New insights into the population genetics of partially clonal organisms: when seagrass data meet theoretical expectations. ArXiv:1902.10240 [q-Bio], v6 peer-reviewed and recommended by Peer Community in Evolutionary Biology. Retrieved from http://arxiv.org/abs/1902.10240
[2] Stoeckel, S., Porro, B., and Arnaud-Haond, S. (2019). The discernible and hidden effects of clonality on the genotypic and genetic states of populations: improving our estimation of clonal rates. ArXiv:1902.09365 [q-Bio], v4 peer-reviewed and recommended by Peer Community in Evolutionary Biology. Retrieved from http://arxiv.org/abs/1902.09365