In experimental evolution followed by whole genome resequencing, parallel evolution, defined as the increase in frequency of identical changes in independent populations adapting to the same environment, is often considered as the product of similar selection pressures and the parallel changes are interpreted as adaptive.
However, theory predicts that heterogeneity both in mutation rate and selection intensity across the genome can trigger patterns of parallel evolution. It is thus important to evaluate and quantify the contribution of both mutation and selection in determining parallel evolution to interpret more accurately experimental evolution genomic data and also potentially improve our capacity to predict the genes that will respond to selection.
In their manuscript, Bailey, Guo and Bataillon [1] derive a framework of statistical models to partition the role of mutation and selection in determining patterns of parallel evolution at the gene level. The rationale is to use the synonymous mutations dataset as a baseline to characterize the mutation rate heterogeneity, assuming a negligible impact of selection on synonymous mutations and then analyse the non-synonymous dataset to identify additional source(s) of heterogeneity, by examining the proportion of the variation explained by a number of genomic variables.
This framework is applied to a published data set of resequencing of 40 Saccharomyces cerevisiae populations adapting to a laboratory environment [2]. The model explaining at best the synonymous mutations dataset is one of homogeneous mutation rate along the genome with a significant positive effect of gene length, likely reflecting variation in the size of the mutational target. For the non-synonymous mutations dataset, introducing heterogeneity between sites for the probability of a change to increase in frequency is improving the model fit and this heterogeneity can be partially explained by differences in gene length, recombination rate and number of functional protein domains.
The application of the framework to an experimental data set illustrates its capacity to disentangle the role of mutation and selection and to identify genomic variables explaining heterogeneity in parallel evolution probability but also points to potential limits, cautiously discussed by the authors: first, the number of mutations in the dataset analysed needs to be sufficient, in particular to establish the baseline on the synonymous dataset. Here, despite a high replication (40 populations evolved in the exact same conditions), the total number of synonymous mutations that could be analysed was not very high and there was only one case of a gene with synonymous mutation in two independent populations. Second, although the models are able to identify factors affecting the mutation counts, the proportion of the variation explained is quite low. The consequence is that the models correctly predicts the mutation count distribution but the objective of predicting on which genes the response to selection will occur still seems quite far away.
The framework developed in this manuscript [1] clearly represents a very useful tool for the analysis of large “evolve and resequence” data sets and to gain a better understanding of the determinants of parallel evolution in general. The extension of its application to mutations others than SNPs would provide the possibility to get a more complete picture of the differences in contributions of mutation and selection intensity heterogeneities depending on the mutation types.

References

[1] Bailey SF, Guo Q and Bataillon T (2018) Identifying drivers of parallel evolution: A regression model approach. bioRxiv 118695, ver. 4 peer-reviewed by Peer Community In Evolutionary Biology. doi: 10.1101/118695

[2] Lang GI, Rice DP, Hickman, MJ, Sodergren E, Weinstock GM, Botstein D, and Desai MM (2013) Pervasive genetic hitchhiking and clonal interference in forty evolving yeast populations. Nature 500: 571–574. doi: 10.1038/nature12344

Starting with the seminar paper by Lande & Arnold (1983), several studies have addressed phenotypic selection in natural populations of a wide variety of organisms, with a recent renewed interest in forest trees (e.g., Oddou-Muratorio et al. 2018; Alexandre et al. 2020; Westergren et al. 2023). Because of their long generation times, long-lived organisms such as forest trees may suffer the most from maladaptation due to climate change, and whether they will be able to adapt to new environmental conditions in just one or a few generations is hotly debated.

In this study, Oddou-Muratorio and colleagues (2024) extend the current framework to add two additional selection components that may alter patterns of fecundity selection and the estimation of standard selection gradients, namely sexual selection (evaluated as differences in flowering phenology conducting to assortative mating) and growth (viability) selection. Notably, the study is conducted in two contrasted environments (low vs high altitude populations) providing information on how the environment may modulate selection patterns in spring phenology. Spring phenology is a key adaptive trait that has been shown to be already affected by climate change in forest trees (Alberto et al. 2013). While fecundity selection for early phenology has been extensively reported before (see Munguía-Rosas et al. 2011), the authors found that this kind of selection can be strongly modulated by sexual selection, depending on the environment. Moreover, they found a significant correlation between early phenology and seedling growth in a common garden, highlighting the importance of this trait for early survival in European beech.

As a conclusion, this original research puts in evidence the need for more integrative approaches for the study of natural selection in the field, as well as the importance of testing multiple environments and the relevance of common gardens to further evaluate phenotypic changes due to real-time selection.

PS: The recommender and the first author of the preprint have shared authorship in a recent paper in a similar topic (Westergren et al. 2023). Nevertheless, the recommender has not contributed in any way or was aware of the content of the current preprint before acting as recommender, and steps have been taken for a fair and unpartial evaluation.

References

Alberto, F. J., Aitken, S. N., Alía, R., González‐Martínez, S. C., Hänninen, H., Kremer, A., Lefèvre, F., Lenormand, T., Yeaman, S., Whetten, R., & Savolainen, O. (2013). Potential for evolutionary responses to climate change - evidence from tree populations. Global Change Biology, 19(6), 1645‑1661.
https://doi.org/10.1111/gcb.12181
 
Alexandre, H., Truffaut, L., Klein, E., Ducousso, A., Chancerel, E., Lesur, I., Dencausse, B., Louvet, J., Nepveu, G., Torres‐Ruiz, J. M., Lagane, F., Musch, B., Delzon, S., & Kremer, A. (2020). How does contemporary selection shape oak phenotypes? Evolutionary Applications, 13(10), 2772‑2790.
https://doi.org/10.1111/eva.13082
 
Lande, R., & Arnold, S. J. (1983). The measurement of selection on correlated characters. Evolution, 37(6), 1210-1226.
https://doi.org/10.2307/2408842
 
Munguía-Rosas, M. A., Ollerton, J., Parra-Tabla, V., & De-Nova, J. A. (2011). Meta-analysis of phenotypic selection on flowering phenology suggests that early flowering plants are favoured. Ecology Letters, 14(5), 511-521
https://doi.org/10.1111/j.1461-0248.2011.01601.x

Oddou-Muratorio S, Bontemps A, Gauzere J, Klein E (2024) Interplay between fecundity, sexual and growth selection on the spring phenology of European beech (Fagus sylvatica L.). bioRxiv, 2023.04.27.538521, ver. 2 peer-reviewed and recommended by Peer Community In Evolutionary Biology https://doi.org/10.1101/2023.04.27.538521 

Oddou-Muratorio, S., Gauzere, J., Bontemps, A., Rey, J.-F., & Klein, E. K. (2018). Tree, sex and size: Ecological determinants of male vs. female fecundity in three Fagus sylvatica stands. Molecular Ecology, 27(15), 3131‑3145.
https://doi.org/10.1111/mec.14770
 
Westergren, M., Archambeau, J., Bajc, M., Damjanić, R., Theraroz, A., Kraigher, H., Oddou‐Muratorio, S., & González‐Martínez, S.C. (2023). Low but significant evolutionary potential for growth, phenology and reproduction traits in European beech. Molecular Ecology, Early View 
https://doi.org/10.1111/mec.17196

The human genome not only encodes for biological functions and for what makes us human, it also encodes the population history of our ancestors. Changes in past population sizes, for example, affect the distribution of times to the most recent common ancestor (tMRCA) of genomic segments, which in turn can be inferred by sophisticated modelling along the genome.
A key framework for such modelling of local tMRCA tracts along genomes is the Sequentially Markovian Coalescent (SMC) (McVean and Cardin 2005, Marjoram and Wall 2006) . The problem that the SMC solves is that the mosaic of local tMRCAs along the genome is unknown, both in their actual ages and in their positions along the genome. The SMC allows to effectively sum across all possibilities and handle the uncertainty probabilistically. Several important tools for inferring the demographic history of a population have been developed built on top of the SMC, including PSMC (Li and Durbin 2011), diCal (Sheehan et al 2013), MSMC (Schiffels and Durbin 2014), SMC++ (Terhorst et al 2017), eSMC (Sellinger et al. 2020) and others.
In this paper, Sellinger, Abu Awad and Tellier (2020) review these SMC-based methods and provide a coherent simulation design to comparatively assess their strengths and weaknesses in a variety of demographic scenarios (Sellinger, Abu Awad and Tellier 2020). In addition, they used these simulations to test how breaking various key assumptions in SMC methods affects estimates, such as constant recombination rates, or absence of false positive SNP calls.
As a result of this assessment, the authors not only provide practical guidance for researchers who want to use these methods, but also insights into how these methods work. For example, the paper carefully separates sources of error in these methods by observing what they call “Best-case convergence” of each method if the data behaves perfectly and separating that from how the method applies with actual data. This approach provides a deeper insight into the methods than what we could learn from application to genomic data alone.
In the age of genomics, computational tools and their development are key for researchers in this field. All the more important is it to provide the community with overviews, reviews and independent assessments of such tools. This is particularly important as sometimes the development of new methods lacks primary visibility due to relevant testing material being pushed to Supplementary Sections in papers due to space constraints. As SMC-based methods have become so widely used tools in genomics, I think the detailed assessment by Sellinger et al. (2020) is timely and relevant.
In conclusion, I recommend this paper because it bridges from a mere review of the different methods to an in-depth assessment of performance, thereby addressing both beginners in the field who just seek an initial overview, as well as experienced researchers who are interested in theoretical boundaries and assumptions of the different methods.

References

[1] Li, H., and Durbin, R. (2011). Inference of human population history from individual whole-genome sequences. Nature, 475(7357), 493-496. doi: https://doi.org/10.1038/nature10231
[2] Marjoram, P., and Wall, J. D. (2006). Fast"" coalescent"" simulation. BMC genetics, 7(1), 16. doi: https://doi.org/10.1186/1471-2156-7-16
[3] McVean, G. A., and Cardin, N. J. (2005). Approximating the coalescent with recombination. Philosophical Transactions of the Royal Society B: Biological Sciences, 360(1459), 1387-1393. doi: https://doi.org/10.1098/rstb.2005.1673
[4] Schiffels, S., and Durbin, R. (2014). Inferring human population size and separation history from multiple genome sequences. Nature genetics, 46(8), 919-925. doi: https://doi.org/10.1038/ng.3015
[5] Sellinger, T. P. P., Awad, D. A., Moest, M., and Tellier, A. (2020). Inference of past demography, dormancy and self-fertilization rates from whole genome sequence data. PLoS Genetics, 16(4), e1008698. doi: https://doi.org/10.1371/journal.pgen.1008698
[6] Sellinger, T. P. P., Awad, D. A. and Tellier, A. (2020) Limits and Convergence properties of the Sequentially Markovian Coalescent. bioRxiv, 2020.07.23.217091, ver. 3 peer-reviewed and recommended by PCI Evolutionary Biology. doi: https://doi.org/10.1101/2020.07.23.217091
[7] Sheehan, S., Harris, K., and Song, Y. S. (2013). Estimating variable effective population sizes from multiple genomes: a sequentially Markov conditional sampling distribution approach. Genetics, 194(3), 647-662. doi: https://doi.org/10.1534/genetics.112.149096
[8] Terhorst, J., Kamm, J. A., and Song, Y. S. (2017). Robust and scalable inference of population history from hundreds of unphased whole genomes. Nature genetics, 49(2), 303-309. doi: https://doi.org/10.1038/ng.3748

Genetic markers are used for in modern population genetics/genomics to uncover the past neutral and selective history of population and species. Besides Single Nucleotide Polymorphisms (SNPs) obtained from whole genome data, microsatellites (or Short Tandem Repeats, SSR) have been common markers of choice in numerous population genetics studies of non-model species with large sample sizes [1]. Microsatellites can be used to uncover and draw inference of the past population demography (e.g. expansion, decline, bottlenecks…), population split, population structure and gene flow, but also life history traits and modes of reproduction (e.g. [2,3]). These markers are widely used in conservation genetics [4] or to study parasites or disease vectors [5]. Microsatellites do show higher mutation rate than SNPs increasing, on the one hand, the statistical power to infer recent events (for example crop domestication, [2,3]), while, on the other hand, decreasing their statistical power over longer time scales due to homoplasy [6].
To perform such analyses, however, an excellent and reliable quality of data is required. As emphasized in the article by De Meeûs et al. [7] three main issues do bias the observed heterozygosity at microsatellites: null alleles, short allele dominance (SAD) and stuttering. These originates from poor PCR amplification. As a result, an excess of homozygosity is observed at the microsatellite loci leading to overestimation of the variation statistics FIS and FST as well as increased linage disequilibrium (LD). For null alleles, several methods and software do help to reduce the bias, and in the present study, De Meeûs et al. [7] propose a way to tackle issues with SAD and stuttering.
The authors study a dataset consisting of 387 samples from 61 subsamples genotyped at nine loci of the species Ixodes scapularis, i.e. ticks transmitting the Lyme disease. Based on correlation methods and FST, FIS they can uncover null alleles and SAD. Stuttering is detected by evaluating the heterozygote deficit between alleles displaying a single repeat difference. Without correction, six loci are affected by one of these amplification problems generating a large deficit of heterozygotes (measured by significant FIS and FST) remaining so after correction for the false discovery rate (FDR). These results would be classically interpreted as a strong Wahlund effect and/or selection at several loci.
After correcting for null alleles, the authors apply two novel corrections: 1) a re-examination of the chromatograms reveals previously disregarded larger alleles thus decreasing SAD, and 2) pooling alleles close in size decreasing stuttering. The corrected dataset shows then a significant excess of heterozygotes as could be expected in a dioecious species with strong population structure. The FDR correction removes then the significant excess of homozygotes and LD between pairs of loci. FST on the cured dataset is used to demonstrate the strong population structure and small effective subpopulation sizes. This is confirmed by a clustering analysis using discriminant analysis of principal components (DAPC).
While based on a specific dataset of ticks from different populations sampled across the USA, the generality of the authors’ approach is presented in Figure 6 in which they provide a step by step flowchart to cure microsatellite datasets from null alleles, SAD and stuttering. Several criteria based on FIS, FST and LD between loci are used as decision keys in the flowchart. An excel file is also provided as help for the curation steps. This study and the proposed methodology are thus extremely useful for all population geneticists working on non-model species with large number of samples genotyped at microsatellite markers. The method not only allows more accurate estimates of heterozygosity but also prevents the thinning of datasets due to the removal of problematic loci. As a follow-up and extension of this work, an exhaustive simulation study could investigate the influence of these data quality issues on past demographic and population structure inference under a wide range of scenarios. This would allow to quantify the current biases in the literature and the robustness of the methodology devised by De Meeûs et al. [7].

References

[1] Jarne, P., and Lagoda, P. J. (1996). Microsatellites, from molecules to populations and back. Trends in ecology & evolution, 11(10), 424-429. doi: 10.1016/0169-5347(96)10049-5
[2] Cornille, A., Giraud, T., Bellard, C., Tellier, A., Le Cam, B., Smulders, M. J. M., Kleinschmit, J., Roldan-Ruiz, I. and Gladieux, P. (2013). Postglacial recolonization history of the E uropean crabapple (Malus sylvestris M ill.), a wild contributor to the domesticated apple. Molecular Ecology, 22(8), 2249-2263. doi: 10.1111/mec.12231
[3] Parat, F., Schwertfirm, G., Rudolph, U., Miedaner, T., Korzun, V., Bauer, E., Schön C.-C. and Tellier, A. (2016). Geography and end use drive the diversification of worldwide winter rye populations. Molecular ecology, 25(2), 500-514. doi: 10.1111/mec.13495
[4] Broquet, T., Ménard, N., & Petit, E. (2007). Noninvasive population genetics: a review of sample source, diet, fragment length and microsatellite motif effects on amplification success and genotyping error rates. Conservation Genetics, 8(1), 249-260. doi: 10.1007/s10592-006-9146-5
[5] Koffi, M., De Meeûs, T., Séré, M., Bucheton, B., Simo, G., Njiokou, F., Salim, B., Kaboré, J., MacLeod, A., Camara, M., Solano, P., Belem, A. M. G. and Jamonneau, V. (2015). Population genetics and reproductive strategies of African trypanosomes: revisiting available published data. PLoS neglected tropical diseases, 9(10), e0003985. doi: 10.1371/journal.pntd.0003985
[6] Estoup, A., Jarne, P., & Cornuet, J. M. (2002). Homoplasy and mutation model at microsatellite loci and their consequences for population genetics analysis. Molecular ecology, 11(9), 1591-1604. doi: 10.1046/j.1365-294X.2002.01576.x
[7] De Meeûs, T., Chan, C. T., Ludwig, J. M., Tsao, J. I., Patel, J., Bhagatwala, J., and Beati, L. (2019). Deceptive combined effects of short allele dominance and stuttering: an example with Ixodes scapularis, the main vector of Lyme disease in the USA. bioRxiv, 622373, ver. 4 peer-reviewed and recommended by Peer Community In Evolutionary Biology. doi: 10.1101/622373

Despite advances in genomic research, many views of genome evolution are still based on what we know from a handful of species, such as humans. This also applies to our knowledge of sex chromosomes. We've apparently been too much used to the situation in which a highly degenerate Y chromosome coexists with an almost normal X chromosome to be able to fully grasp all the questions implied by this situation. Lately, many more sex chromosomes have been studied in other organisms, such as in plants, and the view is changing radically: there is a large diversity of situations, ranging from young highly divergent sex chromosomes to old ones that are so similar that they're hard to detect. Undoubtedly inspired by these recent findings, a few theoretical studies have been published around 2 years ago that put an entirely new light on the evolution of sex chromosomes. The differences between these models have however remained somewhat difficult to appreciate by non-specialists. 

In particular, the models by Lenormand & Roze (2022) and by Jay et al. (2022) seemed quite similar. Indeed, both rely on the same mechanism for initial recombination suppression: a ``lucky'' inversion, i.e. one with less deleterious mutations than the population average, encompassing the sex-determination locus, is initially selected. However, as it doesn't recombine, it will quickly accumulate deleterious mutations lowering its fitness. And it's at this point the models diverge: according to Lenormand & Roze (2022), nascent dosage compensation not only limits the deleterious effects on fitness by the ongoing degeneration, but it actually opposes recombination restoration as this would lead gene expression away from the optimum that has been reached. On the other hand, in the model by Jay et al. (2022), no additional ingredient is required: they argue that once an inversion had been fixed, reversions that restore recombination are extremely unlikely.

This is what Lenormand & Roze (2024) now call a ``constraint'': in Jay et al.'s model, recombination restoration is impossible for mechanistic reasons. Lenormand & Roze (2024) argue such constraints cannot explain long-term recombination suppression. Instead, a mechanism should evolve to limit the negative fitness effects of recombination arrest, otherwise recombination is either restored, or the population goes extinct due to a dramatic drop in the fitness of the heterogametic sex. These two arguments work together: given the huge fitness cost of the lack of ongoing degeneration of the non-recombining Y, in the absence of compensatory mechanisms, there is a very strong selection for the restoration of recombination, so that even when restoration a priori is orders of magnitude less likely than inversion (leading to recombination suppression), it will eventually happen. 

One way the negative fitness effects of recombination suppression can be limited, is the way the authors propose in their own model: dosage compensation evolves through regulatory evolution right at the start of recombination suppression. This changes our classical, simplistic view that dosage compensation evolves in response to degeneration: rather, Lenormand & Roze (2024) argue, that degeneration can only happen when dosage compensation is effective.

The reasoning is convincing and exposes the difference between the models to readers without a firm background in mathematical modelling. Although Lenormand & Roze (2024) target the "constraint theory", it seems likely that other theories for the maintenance of recombination suppression that don't imply the compensation of early degeneration are subject to the same criticism. Indeed, they mention the widely-cited "sexual antagonism" theory, in which mutations with a positive effect in males but a negative in females will select for recombination suppression that will link them to the sex-determining gene on the Y. However, once degeneration starts, the sexually-antagonistic benefits should be huge to overcome the negative effects of degeneration, and it's unlikely they'll be large enough.

A convincing argument by Lenormand & Roze (2024) is that there are many ways recombination could be restored, allowing to circumvent the possible constraints that might be associated with reverting an inversion. First, reversions don't have to be exact to restore recombination. Second, the sex-determining locus can be transposed to another chromosome pair, or an entirely new sex-determining locus might evolve, leading to sex-chromosome turnover which has effectively been observed in several groups.

These modelling studies raise important questions that need to be addressed with both theoretical and empirical work. First, is the regulatory hypothesis proposed by Lenormand & Roze (2022) the only plausible mechanism for the maintenance of long-term recombination suppression? The female- and male-specific trans regulators of gene expression that are required for this model, are they readily available or do they need to evolve first? Both theoretical work and empirical studies of nascent sex chromosomes will help to answer these questions. However, nascent sex chromosomes are difficult to detect and dosage compensation is difficult to reveal.

Second, how many species today actually have "stable" recombination suppression? Maybe many species are in a transient phase, with different populations having different inversions that are either on their way to being fixed or starting to get counterselected. The models have now shown us some possibilities qualitatively but can they actually be quantified to be able to fit the data and to predict whether an observed case of recombination suppression is transient or stable? 

The debate will continue, and we need the active contribution of theoretical biologists to help clarify the underlying hypotheses of the proposed mechanisms. 

Conflict of interest statement: I did co-author a manuscript with D. Roze in 2023, but do not consider this a conflict of interest. The manuscript is the product of discussions that have taken place in a large consortium mainly in 2019. It furthermore deals with an entirely different topic of evolutionary biology.

References

Jay P, Tezenas E, Véber A, and Giraud T. (2022) Sheltering of deleterious mutations explains the stepwise extension of recombination suppression on sex chromosomes and other supergenes. PLoS Biol.;20:e3001698. https://doi.org/10.1371/journal.pbio.3001698
 
Lenormand T and Roze D. (2022) Y recombination arrest and degeneration in the absence of sexual dimorphism. Science;375:663-6. https://doi.org/10.1126/science.abj1813
 
Lenormand T and Roze D. (2024) Can mechanistic constraints on recombination reestablishment explain the long-term maintenance of degenerate sex chromosomes? bioRxiv, ver. 5 peer-reviewed and recommended by Peer Community in Evolutionary Biology. https://doi.org/10.1101/2023.02.17.528909