Epigenetic drift versus benefit: use of selection analysis to evaluate transgenerational phenotypic plasticity

Epigenetic drift versus benefit: use of selection analysis to evaluate transgenerational phenotypic plasticity | Authorea try { document.documentElement.classList.add('js'); } catch (e) { } var _gaq = _gaq || []; _gaq.push(['_setAccount', 'G-8VDV14Y67G']); _gaq.push(['_trackPageview']); (function() { var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true; ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js'; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s); })(); Skip to main content Preprints Collections Wiley Open Research IET Open Research Ecological Society of Japan All Collections About About Authorea FAQs Contact Us Quick Search anywhere Search for preprint articles, keywords, etc. Search Search ADVANCED SEARCH SCROLL This is a preprint and has not been peer reviewed. Data may be preliminary. 3 March 2025 V1 Latest version Share on Epigenetic drift versus benefit: use of selection analysis to evaluate transgenerational phenotypic plasticity Authors : Amanda Pettersen 0000-0001-6191-6563 [email protected] and Frank Seebacher 0000-0002-2281-9311 Authors Info & Affiliations https://doi.org/10.22541/au.174100927.71383129/v1 326 views 223 downloads Contents Abstract Information & Authors Metrics & Citations View Options References Figures Tables Media Share Abstract Transmission of environmentally induced variation from parents to offspring via inter- and trans-generational phenotypic plasticity is a major source of phenotypic variation across generations. In contrast to the slower process of genetic adaptation, epigenetic mechanisms such as altered DNA methylation patterns can change phenotypes between generations and may buffer offspring from environmental stress. Epigenetics thereby has the potential to promote future population persistence. However, epigenetic changes are not always beneficial. Similar to the adaptationist programme described in the classic paper by Gould and Lewontin (1979), a Panglossian approach has been adopted in recent years: transgenerational phenotypic plasticity is often assumed to be beneficial and to enhance offspring fitness. Yet, epigenetic transfer of information can also induce offspring phenotypes that are neutral or detrimental. Here, we challenge the implicit assumption that shifts in offspring phenotype in response to changed parent environments are necessarily adaptive. We instead advocate for the concept of “epigenetic drift” as the most parsimonious null-hypothesis. We propose a quantitative genetics framework to assess the fitness consequences of inter- and transgenerational phenotypic plasticity. We use worked examples to demonstrate how selection analysis can provide standardised estimates of selection to assess the fitness benefits of the transmission of epigenetic information across generations. Introduction Climate variability is one of the most ubiquitous forces shaping ecosystems. Natural environments change at different time scales, from variation at geological scales to the immediate impacts of weather events. These different rates and magnitudes of change impact the physiology that underlies reproduction, growth, and dispersal of individuals (Baduel et al. 2024; Lema 2020; Schulte et al. 2011), and thereby affect population dynamics and communities (Amarasekare 2024; Lema et al. 2024). For example, the composition of communities changes in response to current climate warming as more susceptible species are replaced by species more tolerant to warmer temperature (Khaliq et al. 2024). Similarly, drought and storm activity were associated with species turn-over in a subtropical forest (Wu et al. 2024). Environmental variation, and anthropogenic climate change in particular, is causing population declines in some species. For example, declines in arctic auks ( Alle alle ) are associated with climate change-induced increases in sea surface temperatures and decreases in sea ice (Jakubas et al. 2024). Population declines may be associated with loss of genetic diversity and thereby further reduce the potential for genetic adaptation to climate change (Aagaard et al. 2022; Zhou et al. 2024). However, the more or less ubiquitous climate variation over geological time also means that many species have evolved mechanisms to respond to environmental change and to reduce or alleviate its potentially negative impacts (McGaughran et al. 2021; Seebacher et al. 2015). Darwinian selection drives genetic adaptation to favour phenotypes that perform best under prevailing environmental conditions and thereby alters population allele frequencies. However, genetic adaptation is slow relative to the potential pace of environmental change. Even rapid evolution takes tens of generations before there are significant phenotypic effects that can render a population more resistant to environmental change (Lescak et al. 2015). The pace of adaptation is therefore too slow to mount effective responses to climate change, except for species with very short generation times. In addition to genetic adaptation, most if not all organisms have evolved the capacity for phenotypic plasticity, that is the expression of different phenotypes from a single genotype (Aliaga et al. 2019; DeWitt & Scheiner 2004). Plastic responses are much faster than genetic adaptation, and therefore can be more effective in buffering individuals and populations from rapid environmental variability. Plasticity is regulated by epigenetic mechanisms and manifests at different time scales. For example, DNA (de)methylation is an important mechanism regulating developmental plasticity, and altered methylation patterns may also be passed between generations to influence environmental responses of offspring generations (Heckwolf et al. 2020; Loughland et al. 2021; Sävilammi et al. 2021). Environmental conditions (e.g., temperature) experienced by parents can change the DNA methylation of their gametes and thereby alter offspring phenotypes (intergenerational plasticity) (Fellous et al. 2022; Heckwolf et al. 2020; Venney et al. 2022). In addition to DNA methylation, histone modifications (e.g. methylation, acetylation, phosphorylation) are wide-spread mechanisms by which phenotypes can be altered in an environmentally sensitive manner (Best et al. 2018) including across generations (Skvortsova et al. 2018). Inter- and trans-generational transmissions of epigenetic information are important mechanisms to match phenotypes to changing environmental conditions across single and multiple generations, respectively, and they can influence population dynamics and genetic adaptation to the environment (Brass et al. 2021; Chapelle & Silvestre 2022; Klughammer et al. 2023; Lynch & Kemp 2014; Stajic & Jansen 2021). However, despite their potential to improve resistance and resilience to environmental change, epigenetic modifications are not necessarily beneficial. Epigenetic modifications may be detrimental if epigenetically induced phenotypes do not match the environment experienced by offspring (Beaman et al. 2016; DeWitt & Scheiner 2004; Gibert et al. 2019; Schwanz et al. 2020). Epigenetic marks can also appear stochastically so that their effects are neutral. For example, DNA methylation marks can change stochastically as a result of spontaneous epimutations that arise during cell divisions (Johannes & Schmitz 2019). Epigenetic mutations increase the diversity of epigenetic marks and may be heritable (Bogan & Yi 2024), but do not necessarily alter phenotypes in response to specific environmental signals (Johannes & Schmitz 2019). Epigenetic drift is a related mechanism that is conceptually similar to genetic drift, and which denotes changes in epigenetic marks that accumulate randomly during the lifetime of an organism (Bertucci-Richter et al. 2024). Epigenetic drift is associated with ageing and does not induce phenotypic responses to environmental changes (Bertucci-Richter et al. 2024), and its phenotypic effects may be detrimental (Han 2024). In a now seminal paper, Gould and Lewontin (1979) argued against the ”adaptationist program” that accepted the ”near omnipotence of natural selection in forging organic design and fashioning the best among possible worlds” . In other words, any change in natural systems was viewed as resulting from natural selection to increase fitness. However, as Gould and Lewontin (1979) point out, there are a myriad of other constraints and stochastic changes that make it difficult to determine whether a particular change actually resulted from selection (Nielsen 2009). A similar argument can be made for epigenetic changes. The observations that epigenetic marks are altered in different environmental conditions, or that parental environments influence offspring characteristics by themselves reveal nothing about whether or not epigenetic changes are actually beneficial (”adaptive”). To link epigenetic processes to fitness benefits requires the explicit demonstration of the fitness benefit. Here we propose selection analysis as an approach to test whether epigenetic changes or inter- and trans-generational plasticity (hereon referred to collectively as ‘TGP’) are in fact beneficial, neutral, or detrimental to offspring fitness. These methods are grounded in quantitative genetics and can be applied using experimental measures of offspring trait variation and “fitness” (reproduction or survival) in response to parental exposure to environmental signals that induce changes in offspring phenotypes. The goals of incorporating this analysis are to produce formal, predictive, and comparable estimates of the form, direction, and strength of selection acting on offspring phenotypes, and hence the fitness benefits of TGP for offspring. Evidence for epigenetic drift versus benefit There is a growing focus on the potential benefits of “adaptive” TGP as a means for populations to persist in the face of rapid environmental change (Galloway & Etterson 2007; Harmon & Pfennig 2021; Via et al. 1995). Rather than relying on selection of heritable genetic variation with population-level effects observed after many generations, the parental environment can directly alter the phenotype of their offspring via epigenetic processes to cope with prevailing conditions (Burggren 2015; Skvortsova et al. 2018). This “adaptive” form of intergenerational plasticity can thus buy time for evolution via genetic adaptation to take place (Chevin et al. 2010). While the utility of TGP is appealing particularly in the face of anthropogenic effects, there is no consensus regarding the generality of its adaptive nature (Sánchez-Tójar et al. 2020). We suggest that a more neutral approach is needed to evaluate the fitness implications of TGP. It is important to consider that epigenetic responses are not always beneficial or “adaptive”. For example, moderate heat stress over a long period can confer resilience to heat stress on offspring, but more severe heat stroke can compromise basic cell function by inducing DNA methylation patterns that are ultimately detrimental (Murray et al. 2022). Hence, the nature of the environmental signal can determine whether epigenetic responses are beneficial or detrimental. Even when TGP has demonstrated benefits for offspring performance, these measures of offspring “fitness” are insufficient to conclude any adaptive responses. It is often logistically challenging to measure survival and reproduction in the laboratory or field – instead, the literature is dominated by measures of “performance”, including physiological and life history traits such as offspring size, running/swimming speed, aerobic capacity, as well as developmental and growth rates (Donelson et al. 2012; Le Roy et al. 2017; Pettersen et al. 2019). While shifts in offspring traits are interesting in and of themselves, performance may not reflect actual fitness and may even correlate negatively with actual fitness (lifetime reproductive output) ( Figure 1A ). For example, previous work has shown the presence of trade-offs between reproduction with aging or lifespan (Cohen et al. 2020), growth (Larue et al. 2021), resistance to parasitism (Ives et al. 2020), and survival (Cox et al. 2010). To infer whether a response is indeed beneficial, rigorous tests of the adaptive nature of TGP should include estimates of offspring fitness (lifetime reproductive output) or at least major components of fitness (survival, mating success [fertility], or reproductive output) (Burgess & Marshall 2014; Matesanz et al. 2022). Conversely, TGP can act as a conduit of environmental stress (Marshall & Uller 2007). Challenging or damaging environmental conditions experienced by parents can affect offspring phenotypes epigenetically. In this case, the parental effects on offspring are often viewed negatively. For example, chronic paternal stress in mice altered microRNA composition - an epigenetic mechanism - in sperm and was associated with altered HPA function in offspring (Rodgers et al. 2015). However, similar to the potential benefits of TGP, to provide evidence that TGP is maladaptive, it is important to demonstrate that it reduces offspring fitness. Here, we propose that an a priori assumption of neutral epigenetic effects (i.e., random phenotypic drift) provides a more agnostic and thus unbiased approach to TGP studies. From this premise the potential benefits or detrimental effects of TGP should be tested experimentally, and below we outline an experimental approach and analysis, for those that may be unfamiliar with these techniques, that can be used to achieve this. Figure 1. Measures of parental performance, such as offspring size or aerobic scope may not be correlated, or may even trade-off, with actual fitness (that has not been measured). In this hypothetical example, there appears to be adaptive inter-generational plasticity in offspring size: parents reared in Environment 2 (‘ENV 2’, orange shading) produce larger offspring than parents exposed to Environment 1 (‘ENV 1’, green shading); A) . However, if there is no relationship between offspring fitness and size ( B ) then there will be no fitness benefit of the inter-generational plasticity ( C ). It is therefore essential to link measures of offspring fitness (survival and reproductive output) to plastic traits (size in this case) to avoid misinterpreting the adaptive nature of TGP. Applying a quantitative genetics framework to epigenetics To demonstrate that TGP is either beneficial or detrimental, it is necessary to quantify whether selection (measured as e.g., increased survival or reproductive output) occurs in the direction of, or opposite to the shift in the plastic offspring phenotype (e.g., metabolic rate) across generations, such that TGP increases or decreases offspring fitness, respectively ((Pettersen et al. 2024); Figure 2 ). Natural selection is the differential fitness of phenotypically different individuals (Lande & Arnold 1983). Selection therefore manifests as non-random differential survival or reproductive output across phenotypes (Falconer & Mackay 1996; Fisher 1930; Haldane 1954). Selection acts on individual phenotypes, but its outcomes can alter genetic and thus phenotypic distributions that may be observed at the population level. The potential for selection to shape trait variation of populations relies on heritable genetic variation, and a lack of genetic constraints on those traits (Arnold & Wade 1984). Quantitative measures of selection therefore provide an important first step to provide evidence for the adaptive potential of TGP. Phenotypic plasticity is often thought to reduce the efficacy of selection by increasing phenotypic but not genetic (heritable) variation and thereby reducing the accuracy and effectiveness of selection, and therefore reducing adaptive evolution (Chevin et al. 2010; Gienapp et al. 2007; Merilä & Hendry 2014; Oostra et al. 2018). Yet, selection acts on phenotypes, irrespective of whether they have a genetic or epigenetic basis (Lande & Arnold 1983). Hence, selection analysis is a useful tool that can be utilised to determine whether phenotypic plasticity is producing phenotypic variance that enhances fitness. Further, since these analyses produce selection coefficients that are bound by 0 and 1, they provide comparable estimates across studies, allowing for direct comparisons regarding the direction and strength of selection across treatments and taxa. Selection analysis serves as a simple, yet powerful statistical tool for estimating the direction, form, and strength of selection acting on a trait. Although it cannot distinguish phenotypic selection from an evolutionary (genetic) response to selection (Haldane 1954), applying the same logic and framework developed for genetics (selection analysis) to epigenetics, allows to determine whether epigenetic changes in single or multiple traits are neutral, beneficial, detrimental, via standardised measures of the change in offspring fitness among parental treatments ( Figure 2 ). Figure 2. Hypothetical examples of different forms of selection on single (panels A – E) or multiple (panels F – G) offspring traits. Here, the trait of interest is the shift in offspring phenotype between parental environment treatments. In cases where a shift in an offspring phenotype is correlated with increased offspring fitness, there is positive linear (i.e., directional) selection on TGP that is beneficial (panel A). Conversely, where a shift in offspring phenotype (TGP) is detrimental for offspring fitness, there is negative directional selection (panel B). When selection is not aligned with TGP, there is no significant correlation between TGP and a change in offspring fitness (panel C). Selection gradients can also be nonlinear, where selection favours parents producing intermediate phenotypes via TGP (stabilising or concave selection; panel D) or extreme phenotypes (disruptive or convex selection; panel E). Nonlinear selection can also be detected for two traits, referred to as correlational selection, whereby a shift in offspring fitness is correlated with a shift in two offspring phenotypes in the same (positive correlational selection; panel F) or opposite (negative directional selection; panel G) direction. Selection analysis methods The basic data requirement for a selection analysis are phenotypic measurements from individuals in a population before and after a potential selection event (Lande & Arnold 1983), in parallel with measures of fitness. Fitness measures of survival, ability to reproduce, and reproductive output, can be used to quantify viability, fertility, and fecundity selection, respectively. The type of data collected for inter- and trans-generational studies lend themselves well to implementing selection analysis – typically, parents are exposed to differing treatments (usually representing an environmental change) and the offspring and grand-offspring are tracked in a longitudinal study to measure potential variation in traits and performance. There are several key limitations and methodological considerations for isolating the effects of parental and offspring environments that are discussed below. Once the data have been collected for each combination of parent and offspring environments, the selection landscape of each can be quantified using the following steps: Steps for selection analysis: Check for multicollinearity among traits. Before including multiple traits in a selection analysis, it is important to assess whether these traits are too strongly correlated with one another. To determine whether multiple traits can be included in the same analysis, variance inflation factor (VIF) values should be calculated. VIF values provide a measure of how much the variance of one trait is influenced by its relationship with other traits. A VIF value less than 5 generally indicates low multicollinearity, meaning that the traits are not overly correlated and can be analysed together. A VIF value greater than 5 suggests high multicollinearity, meaning the traits of interest are strongly correlated so that it will be difficult to determine statistically how selection acts on each trait independently. In cases of high multicollinearity, you may need to i) combine or transform correlated traits into composite variables, ii) drop one of the correlated traits to reduce redundancy, or iii) use statistical methods specifically designed to handle multicollinearity, such as ridge regression or principal component analysis. Check for significance of random effects and their interactions. Many experimental studies include random factors such as “family”, “tank”, “run” or “block” to achieve the required sample sizes and replication per treatment or environment. Before proceeding with the analysis, it is important to evaluate whether these random factors significantly influence the traits of interest. This will help to determine whether the traits need to be standardised within or across levels of the random effects. This evaluation can be conducted using linear mixed-effects models that account for both fixed effects (such as treatments or environment) and random effects (such as family or tank) and can be tailored to the nature of your data. For example, use a Gaussian family model for continuous fitness traits (e.g., reproductive output), and a Binomial or Logistic family model for binary traits (e.g., survival or fertility). Assessing the significance of random effects helps to ensure that your model appropriately accounts for variability introduced by experimental design elements. If certain effects are not significant, they can be excluded to simplify the model and improve interpretability. Consider whether selection on a trait is under “hard” or “soft” selection. Determining whether trait values should be standardised and fitness should be mean centred among or within treatments or environments (see step 5) depends on whether the selection pressures are the same or differ across contexts. This concept is based on Wallace (1975): Hard selection: the fitness consequences (i.e., direction and strength of selection) of a trait are consistent across all treatments or environments. In this case, fitness should be mean centred among all groups. Soft selection: the fitness consequences of a trait differ across treatments or environments. Here, fitness should be mean centred within each group to account for these differences. The distinction between hard and soft selection often depends on the specific trait, species, and environmental context (De Lisle & Svensson 2017). Consideration of this distinction can ensure that your analysis appropriately reflects the scale of ecological and evolutionary processes influencing selection. 1. Test for differences in selection across treatments or environments. It is important to evaluate whether selection differs significantly across the treatments or environments tested. This can be achieved by running hypothesis tests to determine if there are significant interactions between selection (linear, quadratic, and correlational) and treatment or environmental variables using logistic regression (Morrissey 2014). To perform this test, i) define your fitness trait (e.g., survival, reproductive success) and the traits under selection, ii) incorporate interaction terms between the traits and the treatment or environmental variable into the model, and iii) use statistical tests to determine whether these interaction terms are significant. If interactions are found, this indicates that selection differs across treatments or environments, and it may be necessary to estimate separate selection coefficients for each group. This approach ensures that using separate selection coefficients is justified, such that your results accurately capture how selection operates under different conditions. 2. Standardise phenotypic traits values and mean-centre fitness. To produce standardised selection coefficients, it is essential to i) standardise each trait value (predictor variable) to have a mean of zero and standard deviation of one, to ensure all traits are on the same scale and thus comparable, and ii) mean-centre fitness by subtracting the mean fitness value from each observation, to centre fitness data around zero, and simplify interpretation of selection gradients. If your analysis includes random effects or multiple treatments or environments, standardisation and mean-centring should also be performed within each random effect or treatment/environment level, as determined in steps 2 and 3. This ensures that traits and fitness are appropriately scaled and centred within the relevant context, avoiding biases introduced by variation among groups. 3. Test for significant selection coefficients. Use multiple regression to evaluate the presence of significant linear, quadratic, and correlational selection coefficients for the traits of interest. Multiple regression models allow for estimating selection gradients and therefore determine how traits influence fitness. • Linear selection: represents directional selection, where higher or lower trait values are correlated with higher and lower fitness, respectively. • Quadratic selection: detects stabilising (selection for intermediate values, reflected by negative coefficient values) or disruptive (selection for extreme values, reflected by positive coefficient values) selection. • Correlational selection: identifies selection on trait combinations, indicating how interactions between traits influence fitness. Choose the appropriate regression family based on the fitness measure being used. For example, Gaussian regression for continuous fitness traits (e.g., reproductive output), and Binomial or Logistic regression for binary fitness traits (e.g., survival or fertility). Add linear, quadratic, then correlational terms in a stepwise fashion to the model and assess the significance of each using statistical tests. Calculate and interpret selection coefficients. Once your models are fitted, extract and interpret the selection coefficients for each trait. Depending on the type of selection and fitness measure, additional transformations may be needed. • Linear selection coefficients: these can often be directly interpreted, showing how a one-unit increase in the trait affects fitness (i.e., the slope of the relationship in Fig. 2A, for example). • Quadratic selection coefficients: these measure curvature in the fitness landscape. Quadratic coefficients must be doubled for accurate interpretation (Stinchcombe et al. 2008), because the raw coefficient from the regression represents only half the curvature. • Fecundity selection: for traits influencing reproductive output (Gaussian-distributed fitness), the regression coefficients can be directly interpreted. • Viability and fertility selection: for binary fitness traits (e.g., survival, fertility), logistic regression slopes need to be transformed to obtain average selection gradients. This can be done using the approach outlined by Janzen and Stern (1998), which accounts for the logistic scale. Create selection surface plots. Selection surface plots provide a visual representation of how fitness varies with combinations of traits. These plots use predictions from your regression model (including linear, quadratic, and correlational terms) to generate surfaces that depict the fitness landscape. These visualisations can be valuable for interpreting selection dynamics, particularly for combinations of traits, and can help to reveal complex patterns, such as for nonlinear selection. To create selection surface plots: • Generate predictions: use your fitted regression model (including linear, quadratic, and correlational selection terms) to predict fitness values across a grid of trait combinations. • Choose visualisation methods: for example, contour plots display lines of equal fitness on a 2D grid, showing how fitness changes across trait combinations, while heatmaps use colour gradients to represent fitness levels, offering a clear and intuitive visualisation of the fitness landscape. • Select tools for plotting: R packages, including “ggplot2” and “modelR” can be used to generate high-quality visualisations based on predictions from your model (Wickham et al. 2023). Case study 1: Pettersen et al., 2024, Phil Trans B Methods This study measured the effects of parent rearing temperature and feeding frequency, and offspring temperature, on embryo mass, embryo metabolic rate, and survival to two weeks post hatching, in zebrafish ( Danio rerio ). There were four parent treatments (temperature: 24 °C and 30 °C and feeding frequency: low and high) and two offspring treatments (temperature: 24 °C and 30 °C) that were fully crossed to produce eight offspring combinations in total. The high temperature (30 °C) and low feeding frequency were assumed to be stressful treatments. Survival to two weeks post hatching was considered to be an important measure of fitness (Pettersen et al. 2024). TGP Parental exposure to 30 °C and low feeding frequency reduced overall offspring survival. A significant TGP response was detected, whereby parents from 30 °C and low feeding frequency produced offspring with lower metabolic rates ( Figure 3 ). Parents from the low food environment also produced heavier embryos than parents from the high feeding frequency treatment ( Figure 3 ). Figure 3. Intergenerational phenotypic plasticity of A) Embryo mass and B) Embryo metabolic rate under different combinations of parent feeding frequency (green = high, yellow = low) and temperature (blue = 24 °C, red = 30 °C) treatments, and offspring temperature (x-axis; 24 C° or 30 °) treatments. Error bars are ± standard error (SE). Selection analysis Significant negative directional selection (linear selection; β ) for metabolic rate was found for offspring reared at 30 °C ( Table 1, Figure 4, Panels B, D, H ) or in offspring from the low food parent treatment ( Figure 4, Panels E, G, H ). There was also positive directional selection for embryo mass in offspring from high food parent treatment ( Figure 4, Panels B, C ), but negative directional selection for embryo mass in offspring from parents exposed to low food or 30 °C ( Figure 4, Panels F, G, H ). Significant stabilising selection (nonlinear quadratic; γ ) on metabolic rates was found for offspring from the P30LO24 parent-offspring treatment ( Figure 4, Panel G ). Table 1. Selection coefficients (linear; β and non-linear; γ ) for offspring embryo mass and metabolic rate (mass-independent) across combinations of parent (‘P’) and offspring (‘O’) environments (temperatures: 24 °C (‘24’) or 30 °C (‘30’) for both parents and offspring and feeding frequency: high (‘H’) or low (‘L’) for parents). Statistically significant coefficients shown in bold. Parent environment Offspring environment β Embryo mass Embryo metabolic rate P24H O24 Embryo mass 0.134 (0.041) -0.053 (0.114) 0.010 (0.055) Embryo metabolic rate -0.061 (0.041) -0.054 (0.092) O30 Embryo mass 0.039 (0.036) -0.117 (0.105) 0.052 (0.057) Embryo metabolic rate -0.157 (0.040) 0.051 (0.113) P30H O24 Embryo mass 0.180 (0.056) 0.371 (0.224) 0.040 (0.125) Embryo metabolic rate 0.052 (0.033) 0.002 (0.068) O30 Embryo mass -0.109 (0.089) 0.210 (0.215) -0.165 (0.181) Embryo metabolic rate -0.167 (0.074) -0.091 (0.074) P24L O24 Embryo mass -0.092 (0.070) -0.106 (0.166) -0.083 (0.124) Embryo metabolic rate -0.122 (0.047) 0.346 (0.320) O30 Embryo mass -0.070 (0.035) -0.343 (0.270) -0.329 (0.222) Embryo metabolic rate -0.050 (0.037) 0.329 (0.320) P30L O24 Embryo mass -0.122 (0.056) -0.039 (0.113) 0.038 (0.095) Embryo metabolic rate -0.161 (0.054) -0.353 (0.138) O30 Embryo mass -0.166 (0.056) 0.218 (0.233) 0.171 (0.111) Embryo metabolic rate -0.120 (0.042) -0.168 (0.104) Figure 4. Viability selection surfaces for parent/offspring treatments (parental (‘P’) and offspring (‘O’) environments (temperatures: 24 °C (‘24’) or 30 °C (‘30’) for both parents and offspring and feeding frequency: high (‘H’) or low (‘L’) for parents). Coloured bars show relative fitness within each temperature (high = yellow, blue = low). Dot points show distribution of data for individual offspring phenotypes within each treatment combination. Conclusions There is evidence for beneficial TGP in metabolic rates, but detrimental TGP in offspring mass. Parents produce offspring with lower metabolic rates when exposed to stressful conditions (30 °C and low food) and this is in line with selection on metabolic rates, where offspring from stressed parents, or offspring exposed to 30 °C, have higher survival when they have relatively low metabolic rates. Despite significant selection for heavier and lighter embryo mass when parents were reared under high and low food, respectively, low food parents produced heavier offspring. This study also emphasised the importance of parent-offspring environment matching. For example, offspring reared at 30 °C from parents reared at 24 °C had the highest metabolic rates, despite the detrimental effects for survival of a high metabolism at 30 °C. Case study 2: Matesanz et al., 2022, Proceedings of the Royal Society B Methods To separate within- versus inter-generational responses to drought stress, this study reared full-siblings in the parental generation of a wild crop, Lupinus angustifolius, to two different watering environments (high-moisture or drought) and reciprocally assigned their offspring to the same treatments (high-moisture or drought), to produce four treatment combinations in total (Matesanz et al. 2022). Among other offspring traits, they measured seed mass and relative growth rate and assessed fitness as lifetime reproductive output (seed number). Results TGP There was no significant effect of parental watering treatment on offspring reproductive output. A significant TGP response was found, whereby seeds from drought-stressed parents were lighter but grew bigger, compared with offspring from high-moisture parents ( Figure 5 ). Figure 5. Intergenerational phenotypic plasticity of A) Individual seed mass and B) Relative growth rate under different combinations of parent and offspring watering treatment (dark red = drought, dark blue = high moisture). Error bars are ±SE. Selection analysis Negative directional selection (linear selection; β ) for relative growth rate was found for offspring reared under high moisture conditions but was only significant for offspring also from parents exposed to the drought treatment ( Table 2, Figure 6, Panel C ). There was also significant stabilising selection (nonlinear quadratic; γ ) on seed mass for offspring both reared under drought and from high-moisture parents ( Figure 6, Panel B ). Table 2. Selection coefficients (linear; β and non-linear; γ ) for offspring seed mass and relative growth rate across combinations of parent (‘P’) and offspring (‘O’) environments (high moisture: ‘H’ versus drought: ‘D’). Statistically significant coefficients shown in bold. Parent environment Offspring environment β Seed mass Relative growth rate PH OH Seed mass -0.027 (0.109) -0.179 (0.154) 0.031 (0.121) Relative growth rate -0.058 (0.104) -0.176 (0.154) OD Seed mass 0.088 (0.124) -0.436 (0.190) 0.125 (0.114) Relative growth rate 0.009 (0.124) -0.203 (0.176) PD OH Seed mass -0.020 (0.108) -0.066 (0.201) 0.027 (0.116) Relative growth rate -0.263 (0.109) -0.104 (0.157) OD Seed mass -0.002 (0.128) -0.078 (0.233) 0.113 (0.405) Relative growth rate -0.188 (0.130) 0.059 (0.165) Figure 6 . Fecundity selection surfaces for parent/offspring treatments (parental (‘P’) and offspring (‘O’) environments (drought (‘D’) or high moisture (‘H’)). Coloured bars show relative fitness within each temperature (high = yellow, blue = low). Dot points show distribution of data for individual offspring phenotypes within each treatment combination. Conclusions: There is little support for TGP aligning with selection in this study. While there is evidence that when exposed to drought conditions, offspring of intermediate mass have highest reproductive output, parents from drought conditions produce significantly lighter seeds. Similarly, while there is selection for low growth rate in offspring from the parental drought - offspring high moisture (PD-OH) treatment, this is not reflected in TGP, where offspring environment is substantially more important for growth rate than parental environment. Hence, a shift in offspring phenotype from drought-stressed parents to produce lighter, faster growing offspring, compared with offspring from high-moisture parents, may be detrimental for offspring. Case study 3: Sun et al., 2023, Ecology and Evolution Methods This study manipulated temperature and parasite exposure in parents and offspring of Daphnia dentifera using a fully factorial design (Sun et al. 2023). There were two parasite (control versus infection) and two temperature (control; 20 °C versus warming; 24 °C) treatments used across both parents and offspring to produce 16 treatment combinations in total (note: we analyse only eight treatments combinations here). Offspring immune response and body size were the phenotypic traits of interest, and lifetime fecundity was the measure of offspring fitness. Results TGP Both parental and offspring treatments, and their interactions, affected offspring lifetime fecundity. There was evidence for TGP when parents experienced only one stressors (infection or warming), but not both. Infected parents produced smaller offspring relative to control parents, while parents in the warming treatment produced larger offspring than those in the control treatment. There were no differences in the immune response of offspring from parents reared in the control versus warming treatment ( Figure 7 ). Figure 7. Intergenerational phenotypic plasticity of offspring body size (panels A and B) and immune response (haemocytes per spore; panel C) under different combinations of A) parent and offspring parasite treatment (green = control, orange = infection), and B and C) parent and offspring temperature treatment (light blue = control, pink = warming). Note, offspring immune response under combinations of parent and offspring parasite treatment not shown due to a lack of immune response in offspring control treatments. Error bars are ±SE. Results Selection analysis No significant selection coefficients detected in any parent x offspring treatment combination tested. Conclusions While there is evidence for TGP that enhances lifetime fecundity in offspring when parents are exposed to either infection or warming, there is no support that this is mediated by offspring body size or immune response. Discussion Selection on TGP is often context dependent In natural populations, environments can be stressful, resources can be limited, or predation may be present, such that variation in phenotypes can pose important fitness consequences (Reznick et al. 2002). In the case studies presented here, we found evidence for selection acting on embryo mass and metabolism ( Case study 1 ) and seed mass ( Case study 2 ). Given that environments also fluctuate across time and space, laboratory experiments often aim to test the implications of environmental variation on individual fitness or performance. In benign environments, such as ‘control’ treatments used in the case studies explored here, individuals exposed to high food availability ( Case study 1 ), high moisture ( Case study 2 ), and no infection ( Case study 3 ), showed evidence for relaxed or no selection. In these examples, variation in fitness of offspring possessing different phenotypes (e.g., small versus large size) were indistinguishable. Larger body and seed size, which are often thought to confer starvation and predation resistance, were not correlated with survival and reproduction in the control environment. Similarly, exposure of parents and offspring to extremely stressful environmental conditions, such as the combined high temperature/low food ( Case study 1 ) and drought ( Case study 2 ) treatments also did not filter out different phenotypes ( Case study 1 : embryo mass and metabolism, Case study 2 : seed size and growth rate) discriminately. It may be a common trend that both benign and extremely stressful environments result in reduced strength of selection, however further selection analyses are needed to elucidate any potentially general patterns. Key limitations and methodological considerations Selection analysis requires trait and fitness data (with sufficient variation) for individuals within each treatment. It is often logistically difficult to measure relative fitness, hence many studies obtain proxies for fitness, such as body size, growth rate, or locomotor performance (Stinchcombe et al. 2017), which can also be used in selection analysis in place of measures of reproduction and survival. For example, it is possible to analyse TGP using morphometric parameters to assess the benefits or costs to locomotor performance and hence broaden the applicability of standardised measures of selection. However, caution should be exercised when analysing TGP using fitness proxies to infer its adaptive potential (Acasuso-Rivero et al. 2019). Franklin and Morrissey (2017) found that using performance measures in place of fitness often underestimates the actual strength of selection. Given the requirement of selection analyses for datasets containing measures of many individuals often tracked longitudinally, it might be assumed that selection analyses are constrained to model organisms measured in the laboratory. Yet, selection analyses have been performed on natural populations of plants (Munguía-Rosas et al. 2011), birds (Charmantier et al. 2004), lizards (Losos et al. 2004), fish (Brooks & Endler 2001), and mammals (Boratyński & Koteja 2010; Kruuk et al. 2002) for several decades. Further selection estimates, particularly on TGP, are crucially needed to build an understanding of the fitness consequences of trait variation under rapid environmental change (Arnold et al. 2019; Rivkin et al. 2019; Svensson 2023). Here we outline some potential limitations and methodological considerations when utilising selection analysis for inferring the adaptive nature of TGP: 1. Parent and offspring treatments need to be applied independently wherever possible. Studying species with short brooding times, such as egg laying taxa, often allows for selection on offspring phenotypes to be analysed without the issue of parental environment as a confounding factor. For species with high parental care, however, it may be difficult to disentangle selection on offspring phenotype independently of parent environment or condition (Kielland et al. 2017). 2. Detecting selection requires that there is sufficient variation in both the traits of interest and fitness. One potential issue for using this type of analysis is that TGP might result in depletion of phenotypic variation in a given environment, such that even if TPG in response to a change in environment increases offspring fitness, it cannot be detected. For example, parents exposed to predation might increase offspring size, leading to relatively higher offspring survival in large versus small offspring, but selection (correlation between fitness and trait) cannot be detected statistically (Engqvist & Reinhold 2016). In line with this, studies on populations already exposed to strong past selection, may not possess the available phenotypic variation required for selection to act on in the present generation. 3. Caution should be exercised when identifying the target of selection. For example, rather than acting upon the measured trait(s), selection may be targeting other unmeasured trait(s) that are correlated with trait(s) of interest, leading to a spurious correlation between fitness and the measured trait(s). It is important therefore to carefully select trait measurements, and to measure and test for selection on multiple traits. 4. It is important to apply a treatment independently from other confounding environmental effects (Rausher 1992; Stinchcombe et al. 2002). For example, the spatial variation of treatments can be randomised to ensure that environmental conditions other than the treatment do not systematically alter phenotypic or fitness variation. 5. Measures of lifetime reproductive output in offspring and grand offspring from TGP studies are exceedingly rare. Hence, it is difficult to determine how long the benefits or costs of epigenetic effects can persist. Future studies measuring whether fitness consequences persist throughout the life history and across generations are needed to determine the temporal stability of selection acting on TGP (Siepielski et al. 2009). These studies could also shed light on the mechanisms underlying beneficial, detrimental, or neutral TGP. 6. It is important to consider that fitness benefits for individual offspring are not always beneficial for parents (Einum & Fleming 2000; Stearns 1992). Selection acting on offspring traits may not lead to adaptive evolution if there is antagonistic selection acting across generations (Wilson et al. 2005). To determine whether TGP is likely to evolve in response to selection, future studies that span generations and quantify phenotypic and fitness covariances between parents and offspring, are an important next step (Wolf & Wade 2001). Acknowledgments We wish to thank all authors of the publications used as case studies used in this paper (Matesanz et al. 2022; Sun et al. 2023) for making their data and code publicly accessible. References Aagaard, A., Liu, S., Tregenza, T., Braad Lund, M., Schramm, A., Verhoeven, K.J.F., et al. (2022). Adapting to climate with limited genetic diversity: Nucleotide, DNA methylation and microbiome variation among populations of the social spider Stegodyphus dumicola. Molecular Ecology , 31, 5765–5783.Acasuso-Rivero, C., Murren, C.J., Schlichting, C.D. & Steiner, U.K. (2019). Adaptive phenotypic plasticity for life-history and less fitness-related traits. Proceedings of the Royal Society B: Biological Sciences , 286, 20190653.Aliaga, B., Bulla, I., Mouahid, G., Duval, D. & Grunau, C. (2019). Universality of the DNA methylation codes in Eucaryotes. Sci Rep , 9, 173.Amarasekare, P. (2024). Temperature-dependent dispersal and ectotherm species’ distributions in a warming world. Journal of Animal Ecology , 93, 428–446.Arnold, P.A., Nicotra, A.B. & Kruuk, L.E.B. (2019). Sparse evidence for selection on phenotypic plasticity in response to temperature. Philosophical Transactions of the Royal Society B: Biological Sciences , 374, 20180185.Arnold, S.J. & Wade, M.J. (1984). On the Measurement of Natural and Sexual Selection: Theory. Evolution , 38, 709–719.Baduel, P., Sammarco, I., Barrett, R., Coronado-Zamora, M., Crespel, A., Díez-Rodríguez, B., et al. (2024). The evolutionary consequences of interactions between the epigenome, the genome and the environment. Evolutionary Applications , 17, e13730.Beaman, J.E., White, C.R. & Seebacher, F. (2016). Evolution of Plasticity: Mechanistic Link between Development and Reversible Acclimation. Trends in Ecology & Evolution , 31, 237–249.Bertucci-Richter, E.M., Shealy, E.P. & Parrott, B.B. (2024). Epigenetic drift underlies epigenetic clock signals, but displays distinct responses to lifespan interventions, development, and cellular dedifferentiation. Aging , 16, 1002–1020.Best, C., Ikert, H., Kostyniuk, D.J., Craig, P.M., Navarro-Martin, L., Marandel, L., et al. (2018). Epigenetics in teleost fish: From molecular mechanisms to physiological phenotypes. Comparative Biochemistry and Physiology Part B: Biochemistry and Molecular Biology , 50 Years of Comparative Biochemistry: The Legacy of Peter Hochachka, 224, 210–244.Bogan, S.N. & Yi, S.V. (2024). Potential Role of DNA Methylation as a Driver of Plastic Responses to the Environment Across Cells, Organisms, and Populations. Genome Biology and Evolution , 16, evae022.Boratyński, Z. & Koteja, P. (2010). Sexual and natural selection on body mass and metabolic rates in free-living bank voles. Functional Ecology , 24, 1252–1261.Brass, D.P., Cobbold, C.A., Ewing, D.A., Purse, B.V., Callaghan, A. & White, S.M. (2021). Phenotypic plasticity as a cause and consequence of population dynamics. Ecology Letters , 24, 2406–2417.Brooks, R. & Endler, J.A. (2001). Direct and indirect sexual selection and quantitative genetics of male traits in guppies (Poecilia reticulata). Evolution , 55, 1002–1015.Burgess, S.C. & Marshall, D.J. (2014). Adaptive parental effects: the importance of estimating environmental predictability and offspring fitness appropriately. Oikos , 123, 769–776.Burggren, W.W. (2015). Dynamics of epigenetic phenomena: intergenerational and intragenerational phenotype ‘washout.’ Journal of Experimental Biology , 218, 80–87.Chapelle, V. & Silvestre, F. (2022). Population Epigenetics: The Extent of DNA Methylation Variation in Wild Animal Populations. Epigenomes , 6, 31.Charmantier, A., Kruuk, L.E.B., Blondel, J. & Lambrechts, M.M. (2004). Testing for microevolution in body size in three blue tit populations. Journal of Evolutionary Biology , 17, 732–743.Chevin, L.-M., Lande, R. & Mace, G.M. (2010). Adaptation, Plasticity, and Extinction in a Changing Environment: Towards a Predictive Theory. PLOS Biology , 8, e1000357.Cohen, A.A., Coste, C.F.D., Li, X.-Y., Bourg, S. & Pavard, S. (2020). Are trade-offs really the key drivers of ageing and life span? Functional Ecology , 34, 153–166.Cox, R.M., Parker, E.U., Cheney, D.M., Liebl, A.L., Martin, L.B. & Calsbeek, R. (2010). Experimental evidence for physiological costs underlying the trade-off between reproduction and survival. Functional Ecology , 24, 1262–1269.De Lisle, S.P. & Svensson, E.I. (2017). On the standardization of fitness and traits in comparative studies of phenotypic selection. Evolution , 71, 2313–2326.DeWitt, T. & Scheiner, S. (2004). Phenotypic Plasticity: Functional and Conceptual Approaches . Oxford University Press.Donelson, J.M., Munday, P.L., McCormick, M.I. & Pitcher, C.R. (2012). Rapid transgenerational acclimation of a tropical reef fish to climate change. Nature Clim Change , 2, 30–32.Einum, S. & Fleming, I.A. (2000). Highly fecund mothers sacrifice offspring survival to maximize fitness. Nature , 405, 3.Engqvist, L. & Reinhold, K. (2016). Adaptive trans-generational phenotypic plasticity and the lack of an experimental control in reciprocal match/mismatch experiments. Methods in Ecology and Evolution , 7, 1482–1488.Falconer, D.S. & Mackay, T.F.C. (1996). Introduction to Quantitative Genetics . Harlow: Pearson.Fellous, A., Wegner, K.M., John, U., Mark, F.C. & Shama, L.N.S. (2022). Windows of opportunity: Ocean warming shapes temperature-sensitive epigenetic reprogramming and gene expression across gametogenesis and embryogenesis in marine stickleback. Global Change Biology , 28, 54–71.Fisher, R.A. (1930). The genetical theory of natural selection . Clarendon Press, Oxford, UK.Franklin, O.D. & Morrissey, M.B. (2017). Inference of selection gradients using performance measures as fitness proxies. Methods in Ecology and Evolution , 8, 663–677.Galloway, L.F. & Etterson, J.R. (2007). Transgenerational Plasticity Is Adaptive in the Wild. Science , 318, 1134–1136.Gibert, P., Debat, V. & Ghalambor, C.K. (2019). Phenotypic plasticity, global change, and the speed of adaptive evolution. Current Opinion in Insect Science , Global change biology • Molecular Physiology, 35, 34–40.Gienapp, P., Leimu, R. & Merilä, J. (2007). Responses to climate change in avian migration timemicroevolution versus phenotypic plasticity. Clim. Res. , 35, 25–35.Gould, S.J. & Lewontin, R.C. (1979). The spandrels of San Marco and the Panglossian paradigm: a critique of the adaptationist programme. Proceedings of the Royal Society of London. Series B. Biological Sciences , 205, 581–598.Haldane, J.B.S. (1954). The measurement of natural selection. Proceedings of the 9th International Congress of Genetics , 480–487.Han, J.-D.J. (2024). The ticking of aging clocks. Trends in Endocrinology & Metabolism , 35, 11–22.Harmon, E.A. & Pfennig, D.W. (2021). Evolutionary rescue via transgenerational plasticity: Evidence and implications for conservation. Evolution & Development , 23, 292–307.Heckwolf, M.J., Meyer, B.S., Häsler, R., Höppner, M.P., Eizaguirre, C. & Reusch, T.B.H. (2020). Two different epigenetic information channels in wild three-spined sticklebacks are involved in salinity adaptation. Science Advances , 6, eaaz1138.Ives, A.R., Barton, B.T., Penczykowski, R.M., Harmon, J.P., Kim, K.L., Oliver, K., et al. (2020). Self-perpetuating ecological–evolutionary dynamics in an agricultural host–parasite system. Nat Ecol Evol , 4, 702–711.Jakubas, D., Wojczulanis-Jakubas, K., Petersen, A., Evans, P.G.H. & Boertmann, D. (2024). Decline in the West Greenland population of a zooplanktivorous seabird, the little auk Alle alle. Sci Rep , 14, 20686.Janzen, F.J. & Stern, H.S. (1998). Logistic Regression for Empirical Studies of Multivariate Selection. Evolution , 52, 1564–1571.Johannes, F. & Schmitz, R.J. (2019). Spontaneous epimutations in plants. New Phytologist , 221, 1253–1259.Khaliq, I., Rixen, C., Zellweger, F., Graham, C.H., Gossner, M.M., McFadden, I.R., et al. (2024). Warming underpins community turnover in temperate freshwater and terrestrial communities. Nat Commun , 15, 1921.Kielland, Ø.N., Bech, C. & Einum, S. (2017). No evidence for thermal transgenerational plasticity in metabolism when minimizing the potential for confounding effects. Proceedings of the Royal Society B: Biological Sciences , 284, 20162494.Klughammer, J., Romanovskaia, D., Nemc, A., Posautz, A., Seid, C.A., Schuster, L.C., et al. (2023). Comparative analysis of genome-scale, base-resolution DNA methylation profiles across 580 animal species. Nat Commun , 14, 232.Kruuk, L.E.B., Slate, J., Pemberton, J.M., Brotherstone, S., Guinness, F. & Clutton‐Brock, T. (2002). Antler size in red deer: heritability and selection but no evolution. Evolution , 56, 1683–1695.Lande, R. & Arnold, S.J. (1983). The Measurement of Selection on Correlated Characters. Evolution , 37, 1210–1226.Larue, B., Pelletier, F., Côté, S.D., Hamel, S. & Festa-Bianchet, M. (2021). Growth and reproduction trade-offs can estimate previous reproductive history in alpine ungulates. Journal of Applied Ecology , 58, 869–878.Le Roy, A., Loughland, I. & Seebacher, F. (2017). Differential effects of developmental thermal plasticity across three generations of guppies (Poecilia reticulata): canalization and anticipatory matching. Sci Rep , 7, 4313.Lema, S.C. (2020). Hormones, developmental plasticity, and adaptive evolution: Endocrine flexibility as a catalyst for ‘plasticity-first’ phenotypic divergence. Molecular and Cellular Endocrinology , 502, 110678.Lema, S.C., Luckenbach, J.A., Yamamoto, Y. & Housh, M.J. (2024). Fish reproduction in a warming world: vulnerable points in hormone regulation from sex determination to spawning. Philosophical Transactions of the Royal Society B: Biological Sciences , 379, 20220516.Lescak, E.A., Bassham, S.L., Catchen, J., Gelmond, O., Sherbick, M.L., von Hippel, F.A., et al. (2015). Evolution of stickleback in 50 years on earthquake-uplifted islands. Proceedings of the National Academy of Sciences , 112, E7204–E7212.Losos, J.B., Schoener, T.W. & Spiller, D.A. (2004). Predator-induced behaviour shifts and natural selection in field-experimental lizard populations. Nature , 432, 505–508.Loughland, I., Little, A. & Seebacher, F. (2021). DNA methyltransferase 3a mediates developmental thermal plasticity. BMC Biology , 19, 11.Lynch, K.E. & Kemp, D.J. (2014). Nature-via-nurture and unravelling causality in evolutionary genetics. Trends in Ecology & Evolution , 29, 2–4.Marshall, Dustin.J. & Uller, T. (2007). When is a maternal effect adaptive? Oikos , 116, 1957–1963.Matesanz, S., Ramos-Muñoz, M., Rubio Teso, M.L. & Iriondo, J.M. (2022). Effects of parental drought on offspring fitness vary among populations of a crop wild relative. Proceedings of the Royal Society B: Biological Sciences , 289, 20220065.McGaughran, A., Laver, R. & Fraser, C. (2021). Evolutionary Responses to Warming. Trends in Ecology & Evolution , 36, 591–600.Merilä, J. & Hendry, A.P. (2014). Climate change, adaptation, and phenotypic plasticity: the problem and the evidence. Evolutionary Applications , 7, 1–14.Morrissey, M.B. (2014). In search of the best methods for multivariate selection analysis. Methods in Ecology and Evolution , 5, 1095–1109.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, 511–521.Murray, K.O., Clanton, T.L. & Horowitz, M. (2022). Epigenetic responses to heat: From adaptation to maladaptation. Experimental Physiology , 107, 1144–1158.Nielsen, R. (2009). Adaptionism—30 Years After Gould and Lewontin. Evolution , 63, 2487–2490.Oostra, V., Saastamoinen, M., Zwaan, B.J. & Wheat, C.W. (2018). Strong phenotypic plasticity limits potential for evolutionary responses to climate change. Nat Commun , 9, 1005.Pettersen, A.K., Metcalfe, N.B. & Seebacher, F. (2024). Intergenerational plasticity aligns with temperature-dependent selection on offspring metabolic rates. Philosophical Transactions of the Royal Society B: Biological Sciences , 379, 20220496.Pettersen, A.K., White, C.R., Bryson-Richardson, R.J. & Marshall, D.J. (2019). Linking life-history theory and metabolic theory explains the offspring size-temperature relationship. Ecology Letters , 22, 518–526.Rausher, M.D. (1992). The measurement of selection on quantitative traits: biases due to environmental covariances between traits and fitness. Evolution , 46, 616–626.Reznick, D., Bryant, M.J. & Bashey, F. (2002). r- and K-Selection Revisited: The Role of Population Regulation in Life-History Evolution.Rivkin, L.R., Santangelo, J.S., Alberti, M., Aronson, M.F.J., de Keyzer, C.W., Diamond, S.E., et al. (2019). A roadmap for urban evolutionary ecology. Evolutionary Applications , 12, 384–398.Sánchez-Tójar, A., Lagisz, M., Moran, N.P., Nakagawa, S., Noble, D.W.A. & Reinhold, K. (2020). The jury is still out regarding the generality of adaptive ‘transgenerational’ effects. Ecology Letters , 23, 1715–1718.Sävilammi, T., Papakostas, S., Leder, E.H., Vøllestad, L.A., Debes, P.V. & Primmer, C.R. (2021). Cytosine methylation patterns suggest a role of methylation in plastic and adaptive responses to temperature in European grayling (Thymallus thymallus) populations. Epigenetics , 16, 271–288.Schulte, P.M., Healy, T.M. & Fangue, N.A. (2011). Thermal Performance Curves, Phenotypic Plasticity, and the Time Scales of Temperature Exposure. Integr Comp Biol , 51, 691–702.Schwanz, L.E., Crawford-Ash, J. & Gale, T. (2020). Context dependence of transgenerational plasticity: the influence of parental temperature depends on offspring environment and sex. Oecologia , 194, 391–401.Seebacher, F., White, C.R. & Franklin, C.E. (2015). Physiological plasticity increases resilience of ectothermic animals to climate change. Nature Climate Change , 5, 61–66.Siepielski, A.M., DiBattista, J.D. & Carlson, S.M. (2009). It’s about time: the temporal dynamics of phenotypic selection in the wild. Ecology letters , 12, 1261–1276.Skvortsova, K., Iovino, N. & Bogdanović, O. (2018). Functions and mechanisms of epigenetic inheritance in animals. Nat Rev Mol Cell Biol , 19, 774–790.Stajic, D. & Jansen, L.E.T. (2021). Empirical evidence for epigenetic inheritance driving evolutionary adaptation. Philosophical Transactions of the Royal Society B: Biological Sciences , 376, 20200121.Stearns, S.C. (1992). The Evolution of Life Histories . OUP Oxford.Stinchcombe, J.R., Agrawal, A.F., Hohenlohe, P.A., Arnold, S.J. & Blows, M.W. (2008). Estimating nonlinear selection gradients using quadratic regression coefficients: double or nothing? Evolution , 62, 2435–2440.Stinchcombe, J.R., Kelley, J.L. & Conner, J.K. (2017). How to measure natural selection. Methods in Ecology and Evolution , 8, 660–662.Stinchcombe, J.R., Rutter, M.T., Burdick, D.S., Tiffin, P., Rausher, M.D. & Mauricio, R. (2002). Testing for Environmentally Induced Bias in Phenotypic Estimates of Natural Selection: Theory and Practice. The American Naturalist , 160, 511–523.Sun, S.-J., Dziuba, M.K., Jaye, R.N. & Duffy, M.A. (2023). Transgenerational plasticity in a zooplankton in response to elevated temperature and parasitism. Ecology and Evolution , 13, e9767.Svensson, E.I. (2023). Phenotypic selection in natural populations: what have we learned in 40 years? Evolution , 77, 1493–1504.Venney, C.J., Wellband, K.W., Normandeau, E., Houle, C., Garant, D., Audet, C., et al. (2022). Thermal regime during parental sexual maturation, but not during offspring rearing, modulates DNA methylation in brook charr (Salvelinus fontinalis). Proceedings of the Royal Society B: Biological Sciences , 289, 20220670.Via, S., Gomulkiewicz, R., De Jong, G., Scheiner, S.M., Schlichting, C.D. & Van Tienderen, P.H. (1995). Adaptive phenotypic plasticity: consensus and controversy. Trends in Ecology & Evolution , 10, 212–217.Wallace, B. (1975). Hard and Soft Selection Revisited. Evolution , 29, 465–473.Wickham, H., Çetinkaya-Rundel, M. & Grolemund, G. (2023). R for Data Science, 2nd Edition[Book] .Wilson, A.J., Pilkington, J.G., Pemberton, J.M., Coltman, D.W., Overall, A.D.J., Byrne, K.A., et al. (2005). Selection on Mothers and Offspring: Whose Phenotype Is It and Does It Matter? Evolution , 59, 451–463.Wolf, J.B. & Wade, M.J. (2001). On the assignment of fitness to parents and offspring: whose fitness is it and when does it matter? Journal of Evolutionary Biology , 14, 347–356.Wu, A., Xiong, X., Zhou, G., Barmon, M., Li, A., Tang, X., et al. (2024). Climate change-related biodiversity fluctuations and composition changes in an old-growth subtropical forest: A 26-yr study. Science of The Total Environment , 914, 169899.Zhou, Y., Song, R., Nevo, E., Fu, X., Wang, X., Wang, Y., et al. (2024). Genomic evidence for climate-linked diversity loss and increased vulnerability of wild barley spanning 28 years of climate warming. Science of The Total Environment , 913, 169679. Information & Authors Information Version history V1 Version 1 03 March 2025 Copyright This work is licensed under a Non Exclusive No Reuse License. Keywords adaptive plasticity dna methylation epigenetics fitness intergenerational phenotypic plasticity natural selection offspring parental effects reproduction survival Authors Affiliations Amanda Pettersen 0000-0001-6191-6563 [email protected] The University of Sydney View all articles by this author Frank Seebacher 0000-0002-2281-9311 University of Sydney View all articles by this author Metrics & Citations Metrics Article Usage 326 views 223 downloads .FvxKWukQNSOunydq8rnd { width: 100px; } Citations Download citation Amanda Pettersen, Frank Seebacher. Epigenetic drift versus benefit: use of selection analysis to evaluate transgenerational phenotypic plasticity. Authorea . 03 March 2025. DOI: https://doi.org/10.22541/au.174100927.71383129/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu . Format Please select one from the list RIS (ProCite, Reference Manager) EndNote BibTex Medlars RefWorks Direct import Tips for downloading citations document.getElementById('citMgrHelpLink').addEventListener('click', function() { popupHelp(this.href); return false; }); $(".js__slcInclude").on("change", function(e){ if ($(this).val() == 'refworks') $('#direct').prop("checked", false); $('#direct').prop("disabled", ($(this).val() == 'refworks')); }); View Options View options PDF View PDF Figures Tables Media Share Share Share article link Copy Link Copied! Copying failed. Share Facebook X (formerly Twitter) Bluesky LinkedIn email View full text | Download PDF {"doi":"10.22541/au.174100927.71383129/v1","type":"Article"} Now Reading: Share Figures Tables Close figure viewer Back to article Figure title goes here Change zoom level Go to figure location within the article Download figure Toggle share panel Toggle share panel Share Toggle information panel Toggle information panel Go to previous graphic Go to next graphic Go to previous table Go to next table All figures All tables View all material View all material xrefBack.goTo xrefBack.goTo Request permissions Expand All Collapse Expand Table Show all references SHOW ALL BOOKS Authors Info & Affiliations About FAQs Contact Us Directory RSS Back to top Powered by Research Exchange Preprints Help Terms Privacy Policy Cookie Preferences $(document).ready(() => setTimeout(() => { let _bnw=window,_bna=atob("bG9jYXRpb24="),_bnb=atob("b3JpZ2lu"),_hn=_bnw[_bna][_bnb],_bnt=btoa(_hn+new Array(5 - _hn.length % 4).join(" ")); $.get("/resource/lodash?t="+_bnt); },4000)); (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'a00090d87af409d6',t:'MTc3OTUwMjY2Mg=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();