United by conflict: Convergent signatures of parental conflict in angiosperms and placental mammals

Abstract

Endosperm in angiosperms and placenta in eutherians are convergent innovations for efficient embryonic nutrient transfer. Despite advantages, this reproductive strategy incurs metabolic costs that maternal parents disproportionately shoulder, leading to potential inter-parental conflict over optimal offspring investment. Genomic imprinting—parent-of-origin-biased gene expression—is fundamental for endosperm and placenta development and has convergently evolved in angiosperms and mammals, in part, to resolve parental conflict. Here, we review the mechanisms of genomic imprinting in these taxa. Despite differences in the timing and spatial extent of imprinting, these taxa exhibit remarkable convergence in the molecular machinery and genes governing imprinting. We then assess the role of parental conflict in shaping evolution within angiosperms and eutherians using four criteria: 1) Do differences in the extent of sibling relatedness cause differences in the inferred strength of parental conflict? 2) Do reciprocal crosses between taxa with different inferred histories of parental conflict exhibit parent-of-origin growth effects? 3) Are these parent-of-origin growth effects caused by dosage-sensitive mechanisms and do these loci exhibit signals of positive selection? 4) Can normal development be restored by genomic perturbations that restore stoichiometric balance in the endosperm/placenta? Although we find evidence for all criteria in angiosperms and eutherians, suggesting that parental conflict may help shape their evolution, many questions remain. Additionally, myriad differences between the two taxa suggest that their respective biologies may shape how/when/where/to what extent parental conflict manifests. Lastly, we discuss outstanding questions, highlighting the power of comparative work in quantifying the role of parental conflict in evolution.

Introduction

Despite belonging to vastly different biological kingdoms, eutherian mammals and flowering plants have independently evolved key innovations to support the nourishment of their developing embryos and enhance their chances of successful reproduction; the placenta and endosperm, respectively. The placenta is a chimeric organ, composed of maternal and fetal tissues, that facilitates nutrient and gas exchange between the mother and the developing fetus (Cross et al. 1994). Similarly, the endosperm is a nutritive tissue formed through a second fertilization event alongside the embryo. Endosperm is often triploid (2 maternal: 1 paternal), although some early diverging angiosperms have a diploid endosperm (1 maternal: 1 paternal) (Williams and Friedman 2002, 2004). The endosperm is vital for embryo growth by providing nutrition and facilitating its development (Brink and Cooper 1947). These convergently evolved structures are hypothesized to contribute to rapid diversifications (Fitzpatrick 2004; Garratt et al. 2013; Roberts et al. 2016; Springer et al. 2019; Coughlan 2023c); while there are ~300 species of marsupials, there are over 6,000 species of eutherians (Nilsson et al. 2004; Burgin et al. 2018; Koepfli and Gooley 2020). Similarly, gymnosperms consist of ~1,000 species, whereas angiosperms encompass over 300,000 species (Wang and Ran 2014; Christenhusz and Byng 2016; Yang et al. 2022).

Yet, the metabolic burden of this investment strategy is costly and unequally distributed between parents, setting the stage for potential conflict of evolutionary interests known as parent-offspring conflict (Hamilton 1964a, 1964b; Trivers 1974; Charnov 1979; Queller 1984). This occurs when the maternal parent assumes the sole responsibility of nourishing the developing embryo, whereas the paternal parent merely contributes genetic material. This discrepancy in investment creates opposing selective pressures on each parent; in polyandrous taxa where multiple males sire different embryos in a single brood, selection can favor paternal alleles that promote increased maternal investment in their offspring (Hamilton 1964a, 1964b; Trivers 1974; Queller 1984). Selection should then favor maternally derived compensatory mutations that minimize and equalize investment among developing embryos since all offspring are equally related to the maternal parent. Genomic imprinting—the phenomenon by which genes are expressed in a parent-of-origin manner—is proposed to have evolved to mitigate locus-specific conflict among parents over resource allocation to developing offspring (i.e. kinship or parental conflict theory; Haig and Westoby 1989; Moore and Haig 1991; Haig 1997). Genomic imprinting has independently evolved in angiosperms and mammals (Feil and Berger 2007; Pires and Grossniklaus 2014).

Eutherians and angiosperms are thus united by the convergent evolution of these nutritive tissues, and by the conflict that can evolve as a consequence. Here we take advantage of this broad-scale convergence in phenotype to assess similarities and differences between angiosperms and eutherians in genomic imprinting. We then evaluate whether parental conflict shapes more contemporary evolution within these two taxa using four criteria: 1) Do differences in the extent of sibling relatedness cause differences in the inferred strength of parental conflict? 2) Do reciprocal crosses between taxa with different inferred histories of parental conflict exhibit parent-of-origin growth effects? 3) Are these parent-of-origin growth effects caused by dosage-sensitive mechanisms and do these loci exhibit a history of strong, positive selection? 4) Can normal development be restored by genomic perturbations that restore stoichiometric balance in the endosperm/placenta? Finally, we highlight unresolved questions that could provide further insight into the role of parental conflict in angiosperm and eutherian evolution, and the ways in which synthesis between taxa can provide novel insights into how conflict can shape evolution.

Genomic imprinting in angiosperms and placental mammals

Both the endosperm and placenta are hotspots for genomic imprinting, and imprinted expression is a canonical feature of their development (McGrath and Solter 1984; Surani et al. 1984; Kinoshita et al. 1999; Vielle-Calzada et al. 1999; Köhler et al. 2003a, 2003b, 2005; Inoue et al. 2018; Mei et al. 2021; Zhao et al. 2022). Despite remarkable similarities in the convergent evolution of genomic imprinting, key details of imprinting differ between mammals and angiosperms (Figs. 1 and 2). In mammals, while some imprinted genes are found singly or in pairs, ~65%–80% of imprinted genes are arranged in highly conserved clusters, where gene expression is controlled by imprinting control regions (ICRs; Wan and Bartolomei 2008; Bonaldi et al. 2017). ICRs, rich in CpG dinucleotide repeats, are often upstream of genes where their methylation status controls gene expression via the differential binding to chromatin remodeling complexes or transcription factors (Spahn and Barlow 2003; Edwards and Ferguson-Smith 2007; Barlow and Bartolomei 2014). The parent-of-origin specific methylation pattern of ICRs is established de novo in mammalian gametes via DNA methyltransferase (DNMT3a), somatically maintained in the embryo via DNMT1, then erased in primordial germ cells for subsequent re-establishment during gametogenesis (Fig. 1a; Li et al. 1992; Okano et al. 1998; Chedin et al. 2002; Hata et al. 2002; Jaenisch and Bird 2003). These comprise the canonical imprints (Andergassen et al. 2021), a prime example of which includes the ICR located between Insulin-like growth factor II (Igf2) and H19 (Bartolomei et al. 1991; Murrell et al. 2004; Nordin et al. 2014). The CCCTC-binding factor (CTCF) binds to the unmethylated ICR on the maternal allele, insulating Igf2 from its downstream enhancers and limiting its expression while enabling the expression of H19 (Bell and Felsenfeld 2000; Hark et al. 2000; Kanduri et al. 2000; Szabó et al. 2000). Conversely, the methylated ICR on the paternal allele prevents CTCF binding, allowing for the expression of Igf2 while silencing H19 (Bell and Felsenfeld 2000; Hark et al. 2000; Kanduri et al. 2000; Szabó et al. 2000; Howell et al. 2001). Together, this creates maternal-specific expression of H19 (a maternally expressed gene; MEG) and paternal-specific expression of Igf2 (a paternally expressed gene; PEG). Other less common forms of imprinting—referred to as non-canonical imprints—also exist in eutherian mammals, but have been exclusively reported in rodents (Andergassen et al. 2021; Richard Albert et al. 2023). One form of non-canonical imprints is histone H3 lysine-27 tri-methylation (H3K27me3). These marks are established via Polycomb Repressive Complex 2 (PRC2) which is composed of multiple subunits such as Eed (Embryonic Ectoderm Development) and Ezh1/2 (Enhancer of Zeste 1 and 2), some of which act in concert with the deposition of a second histone modifying mark—H2AK119ub1—via PRC1 (Inoue et al. 2017; Du et al. 2020; Andergassen et al. 2021; Mei et al. 2021; Inoue 2023; Rodriguez-Caro et al. 2023). These imprints arise during oogenesis, mediating the expression of ~30 PEGs (Fig. 2c and d; Andergassen et al. 2021; Richard Albert et al. 2023; Rodriguez-Caro et al. 2023). Postfertilization, PRCs maintain epigenetic modifications on the maternal allele of PEGs via positive feedback (Holoch and Margueron 2017). Conversely, deposition of H3K27me3 is repressed on paternal alleles, promoting paternal expression, potentially via the deposition of another type of histone modification—H3K36me3 (Xu et al. 2019). After embryo implantation, non-canonical imprints are maintained in the placenta through a second round of differential DNA methylation via DNMT3b and G9a/Glp (an H3K9 methyltransferase), but lost in the embryo (Chen et al. 2019; Andergassen et al. 2021; Zeng et al. 2021). Non-canonical imprints are therefore re-established each generation.

Unlike mammals, imprinted genes in angiosperms are semi-scattered throughout the genome (Wolff et al. 2011; Waters et al. 2013; Hatorangan et al. 2016; Zhang et al. 2016b). Imprinting establishment, maintenance, and erasure are distinctly different in angiosperms than mammals, with both methylation- and non-methylation-based imprinting being common (Figs. 1 and 2; Batista and Köhler 2020). In angiosperms, differential imprinting is established during late gametogenesis (Figs. 1 and 2). In both the central cell (in ovules) and the vegetative cell (in pollen), DEMETER (DME) removes DNA methylation, leading to an increased transposable element (TE) activity (Choi et al. 2002; Gehring et al. 2009; Hsieh et al. 2009). This uptick in TE activity generates small interfering RNAs (siRNAs), which can then migrate from the central or vegetative cell to the egg or sperm cells triggering DNA methylation at TE-flanked imprinted genes. In the sperm cell, siRNAs mediate methylation via the RNA-directed DNA methylation (RdDM) that silence the paternal copies of MEGs (Fig. 2a; Calarco et al. 2012; Batista and Köhler 2020). Methylation status can subsequently be reinforced by DNA METHYLTRANSFERASE 1 (MET1) and/or CHROMOMETHYLASE3 (CMT3; Calarco et al. 2012). Like mammals, this creates differential methylation profiles between the parental alleles. However, unlike mammals, differential methylation is largely associated with MEGs in angiosperms—the maternal copy of MEGs is hypomethylated and expressed, while the paternal copy is hypermethylated and silenced. Meanwhile, the maternal copies of many PEGs are silenced via H3K27me3 histone modifications deposited by the FERTILIZATION-INDEPENDENT SEED PRC2 (FIS-PRC2) (Fig. 2b; Wolff et al. 2011; Moreno-Romero et al. 2016, 2019), and in some cases further reinforced by H3K9me2 and CHG methylation depositions (Moreno-Romero et al. 2019). This is largely due to the exclusive expression of PRC2 proteins in the female gametophyte (Batista and Köhler 2020). Although these two models of imprinting—paternal silencing of MEGs via hypermethylation and maternal silencing of PEGs via H3K27me3 deposition—are common mechanisms of imprinting, other molecular mechanisms are likely, as 65%–68% of MEGs and up to 30% of PEGs lack either differential DNA methylation or H3K27me3 activity (Batista and Köhler 2020). Since imprinting predominantly occurs in the endosperm, rather than the embryo (Gehring et al. 2011; Hsieh et al. 2011; Waters et al. 2013; Pignatta et al. 2014), imprints do not persist in plant tissues post-germination and must be re-established each generation during gametogenesis (Fig. 1b; Rodrigues and Zilberman 2015).

Despite spatial and temporal differences in imprinting between angiosperms and mammals, these groups share remarkable similarities in the underlying molecular mechanisms. In both taxa, DNA methylation and histone modifications are the main regulatory mechanisms governing imprinting and both recruit homologous DNA methylase/demethylase enzymes. For instance, MET1 and DNMT1 are homologs, and both play a key role in maintaining CpG methylation (Finnegan and Dennis 1993; Finnegan and Kovac 2000; Xiao et al. 2006; Feil and Berger 2007). Furthermore, both taxa recruit homologous PRC2 Complexes—FIS-PRC2 in angiosperms and PRC2 in mammals—to deposit H3K27me3 marks, predominantly on maternal copies of PEGs (Reyes and Grossniklaus 2003; Baroux et al. 2006; Feil and Berger 2007; Batista and Köhler 2020; Inoue 2023). Moreover, some subunits comprising each PRC2 are homologous; for example, both MEDEA (MEA; in angiosperms) and Ezh2 (in mammals) have histone methyltransferase domains (Hennig and Derkacheva 2009; Mozgova et al. 2015). The molecular machinery governing imprinting is ancient, dating back to early eukaryotic evolution, and may have origins in epigenetic regulation, maintenance of cell identity, and multicellularity (Shaver et al. 2010; Mozgova et al. 2015; Fischer et al. 2022; Sharaf et al. 2022; Vijayanathan et al. 2022). Nonetheless, it is striking that this molecular machinery has been independently recruited to govern imprinting in both angiosperms and placental mammals.

There are many hypotheses to explain the repeated evolution of genomic imprinting (Spencer and Clark 2014). While it is likely that imprinting may have evolved in response to several selective pressures, we focus on the evidence for parental conflict. First, the presence of imprinting has a distinct phylogenetic distribution. Imprinting is a fundamental feature of angiosperm and eutherian development, although it is also present in some eusocial insects, marsupials, and at least one liverwort (Feil and Berger 2007; Pires and Grossniklaus 2014; Kocher et al. 2015; Montgomery and Berger 2024). Notably, in oviparous taxa (e.g. birds, amphibians, and most reptiles)—where maternal investment is predetermined and cannot be altered postfertilization irrespective of the male’s genetic contribution—genomic imprinting is absent (Clutton-Brock and Scott 1991; Guillette 1993; Frésard et al. 2014; Pires and Grossniklaus 2014). Similarly in plants, imprinting has not been reported in gymnosperms, wherein embryos are surrounded by a predetermined unfertilized nutritive tissue derived from the female gametophyte (Montgomery and Berger 2021). Even within mammals, the extent of imprinting differs between eutherians and their close relatives. Over 200 imprinted genes have been reported in placental mammals (Gregg et al. 2010; Wang et al. 2011; Santini et al. 2021; Ishihara et al. 2024), 9 in marsupials (Edwards 2019; Ishihara et al. 2022) and none in monotremes (Renfree et al. 2009; Edwards 2019). Thus, the origin and diversification of imprinting appear to be linked to the origin of tissues that facilitate dynamic nutritive exchange during offspring development.

Second, the expression of genomic imprinting is highly tissue-specific, and imprinted genes often function to actively facilitate nutrient provisioning from the maternal donor to developing offspring (Table 1). In angiosperms, most imprinted genes are found in the endosperm where they regulate its proper development (Jahnke and Scholten 2009; Pignatta et al. 2014; Lafon-Placette and Köhler 2016; Jiang et al. 2017; Flores-Vergara et al. 2020; Köhler et al. 2021; Butel et al. 2023). In eutherians, imprinted genes are predominantly found in the placenta, fetus, and brain and are crucial for regulating the placenta and embryonic prenatal, neonatal, and early postnatal growth (Renfree et al. 2009; Barlow and Bartolomei 2014; Tucci et al. 2019; Arévalo and Campbell 2020). Even more specifically; tissues within the placenta and endosperm that are more directly responsible for maternal-filial nutrient transfer appear to be enriched for particular imprinted genes. In eutherians, many PEGs that have central roles in nutrient acquisition and placental growth are specifically expressed in the labyrinth zone—a layer of tissue composed of trophoblast cells, fetal capillaries, and maternal blood that is specialized in nutrient transfer (Rodriguez-Caro et al. 2023). For example, the PEG Igf2 described above is positively associated with placental growth (DeChiara et al. 1991; Constância et al. 2002). Similarly, PEGs appear to be enriched in endosperm structures directly associated with nutrient transfer, such as the chalazal endosperm in Arabidopsis thaliana (Picard et al. 2021) and the Basal Endosperm Transfer Layer in maize (Zhan et al. 2015). This relationship is less pronounced with MEGs. Although certainly some eutherian MEGs have been shown to restrict growth, such as insulin-like growth factor 2 Receptor (Igf2r; Lau et al. 1994; Ludwig et al. 1996), several other MEGs are involved in maternal immune function and maternal–fetal molecular signaling (Rodriguez-Caro et al. 2023). In angiosperms, recent work in A. thaliana suggests that a family of MEGs that are responsible for Auxin signaling—centromeric auxin response factors (cARFs)—directly cause cellularization; a crucial developmental step after which nutrient transfer between the female and offspring is greatly reduced in species with nuclear-type endosperm (Butel et al. 2023). Moreover, overexpression of cARFs results in maternal-excess phenotypes; premature cellularization, reduced endosperm, and overall smaller seeds while diminishing expression results in paternal-excess phenotypes; delayed cellularization, excessive endosperm size, and overall larger seeds (Butel et al. 2023). Similarly, in maize and tomato, MEGs are over-represented in the Auxin signaling pathway, and involved in nutrient uptake and allocation (Xin et al. 2013; Roth et al. 2018b). Although many MEGs that function in nutrient allocation play a restrictive role in endosperm growth in line with parental conflict, one opposing example is MEG1 in maize, which positively regulates nutrient transfer through the mother:seed interface (Costa et al. 2012). Lastly, in angiosperms MEGs and PEGs have been found to co-localize for quantitative trait loci involved in grain size, and disruption of MEGs and PEGs has been shown to increase or reduce starch deposition, respectively (Costa et al. 2012; Yuan et al. 2017). Thus, in both eutherians and angiosperms PEGs appear to be largely associated with nutrient acquisition and are expressed in tissues that mediate such function. Conversely, MEGs appear to function in facilitating nutrient transfer and key developmental stages in endosperm, but exhibit a broader range of functions in eutherians, including maternal–fetal communications and stress response. These latter functions may highlight other selective pressures governing the evolution of imprinted loci, such as the maternal–fetal coadaptation theory, which posits that maternally biased expression should evolve at interacting loci that coordinate the development of maternal–fetal tissue to avoid possible negative epistatic interactions between paternal and maternal genomes (Wolf and Hager 2006).

The phylogenetic distribution, tissue specificity, and function of many imprinted genes support a role of parental conflict in the convergent evolution of imprinting in mammals and angiosperms, although it also highlights that other selective pressures may govern imprinted gene expression. Since the origin of genomic imprinting, imprinted genes have diversified and expanded within each of these taxonomic groups. This diversification has been linked with the evolution of reproductive isolation between closely related lineages (Gray 1972; Haig and Westoby 1991; Zeh and Zeh 2000; Brekke and Good 2014; Coughlan 2023c). As with the origin of imprinting itself, divergence in imprinting within each taxa is likely caused by many selective pressures as well as genetic drift. For example, ecological divergence or differences in maternal–fetal coadaptation may cause divergence in placenta/endosperm development (Leishman et al. 2000; Wolf and Hager 2006, 2009; Jong and Scott 2007; Patten et al. 2014; Yuan et al. 2017; Farnitano and Sweigart 2023; Wilsterman et al. 2023), potentially via divergence in imprinted expression. Given these myriad potential evolutionary pressures, we propose four criteria for evaluating whether parental conflict has shaped the divergence in genomic imprinting between closely related lineages within eutherians and angiosperms.

Criteria for parental conflict

Criterion 1: Genetic relatedness mediates the strength of parental conflict

Parental conflict arises because maternal resources are limited and the ability to solicit greater resources during early development should confer greater fitness later in life (Haig and Westoby 1989). As maternity is guaranteed and selection should therefore favor equivalent resource distribution among offspring via maternally inherited alleles, the ability of offspring to solicit extra resources must stem from paternally inherited alleles (Coughlan 2023a). It is this observation—that sibling relatedness via sires should mediate the strength of selection by parental conflict—that has sparked several hypotheses about what factors influence the severity of parental conflict.

Before describing these hypotheses, it is important to recognize that a parallel framework was developed by botanists—that of endosperm balance number (EBN; i.e. effective ploidy). Both placenta and endosperm are dosage-sensitive tissues (McGrath and Solter 1984; Surani et al. 1984; Kinoshita et al. 1999; Vielle-Calzada et al. 1999; Köhler et al. 2003a, 2003b, 2005; Inoue et al. 2018; Mei et al. 2021; Zhao et al. 2022). Botanists had long documented that interploidy crosses often failed in predictable ways (reviewed in Brink and Cooper 1947). Crosses involving a higher ploidy dam largely resulted in precocious endosperm development and smaller seeds, while the reciprocal cross yielded delayed endosperm development and larger seeds; phenotypes that are referred to as maternal- and paternal-excess phenotypes, respectively (Johnston et al. 1980; Johnston and Hanneman 1982; Lin 1984; Scott et al. 1998; Zhang et al. 2016a). Yet, many intra-ploidy crosses also exhibit the same asymmetric growth patterns, despite maintaining the proper 2m:1p endosperm ratio. To reconcile these parallel crossing phenotypes, Johnston et al. (1980) created the EBN theory, wherein individuals are described as having a higher EBN if their offspring exhibited maternal-excess phenotypes when they are dams and paternal-excess phenotypes when sires. Although EBN is a description of a population’s ability to cross to others, differences in EBN likely reflect differences in the strength of parental conflict between populations (Städler et al. 2021; Coughlan 2023c). These differences are also sometimes referred to in terms of “genome strength” (Brandvain and Haig 2018). Here, we describe populations as “stronger” or “weaker,” denoting higher or lower EBN lineages with stronger or weaker inferred history of parental conflict, respectively.

A prominent hypothesis to explain differences in genome strength is that of the mating system. In predominantly self-fertilizing or monogamous organisms, siblings are equally related via their dams and sires, and should therefore experience less parental conflict than a highly outcrossing or polyandrous organism where siblings are equally related via dams but not necessarily equally related via sires. This concept was formalized by the Weak-Inbreeding/Strong-Outbreeder (WISO) hypothesis, which states that largely inbreeding populations should comprise relatively weaker dams and sires than outcrossing populations, and that crosses between the two should result in signals of paternal-excess when the outcrossing population serves as the sire, and maternal-excess when serving as the dam (Haig 1997; Brandvain and Haig 2005). In plants, much evidence supports WISO in both inter- and intraspecific crosses. In both cases, when the more outcrossing population is the dam, seeds exhibit maternal-repression phenotypes and paternal-excess in the reciprocal cross (Brandvain and Haig 2005; Willi 2013; Lafon-Placette et al. 2018; Raunsgard et al. 2018; İltaş et al. 2021; Gustafsson et al. 2022; Petrén et al. 2023). In animals, this hypothesis is much less tested, in part due to the challenges of quantifying mating systems. Nonetheless, evidence of maternal- and paternal-excess has been demonstrated in both Mus and Peromyscus hybrids, wherein reciprocal crosses between a monogamous and polyandrous species exhibit paternal-excess phenotypes when the polyandrous species is the sire and maternal-excess phenotypes in the reciprocal cross (Rogers and Dawson 1970; Zechner et al. 1996; Vrana et al. 1998, 2000; Cassaing and Isaac 2007; Duselis and Vrana 2007). In crosses between lion (Panthera leo) and tiger (Panthera tigris) hybrids exhibit paternal-excess phenotypes when sired by lions and maternal-excess phenotypes in the reciprocal cross (Gray 1972; Vrana 2007). Lion prides often contain more than one male and frequently these males are not close genetic relatives (Packer and Pusey 1982; Packer et al. 1991), while tigers maintain single-male territories (Smith and Mcdougal 1991). It is conceivable that these differences may result in increased multiple paternity in lions relative to tigers. However, both tigers and lions exhibit significant extra-pair paternity (Liu et al. 2013; Lyke et al. 2013), suggesting that they do not conform to the strict monogamy/polyandrous dichotomy of Mus and Peromyscus.

Factors besides mating system can also influence the severity of parental conflict that a population/species has experienced through evolutionary time. Sex-biased survival, fecundity, or dispersal can result in unequal relatedness among siblings via their paternally inherited alleles, thereby influencing the strength of parental conflict (Fig. 3; Van Cleve et al. 2010; Brandvain et al. 2011). Moreover, demographic factors can also directly influence mating system (Fig. 3), influencing the strength of selection via parental conflict. For example, a severe bottleneck may decrease the number of effective sires, thus increasing rates of biparental inbreeding. The potential impacts of demographic and life history factors in shaping patterns of parental conflict are exemplified by patterns of maternal- and paternal-excess in Mimulus (Garner et al. 2016; Coughlan et al. 2020; Sandstedt et al. 2020) and tomatoes (Roth et al. 2018a), in which maternal- and paternal-excess phenotypes cannot be explained by differences in mating systems. In mammals, these hypotheses have been much less explored. However, several interspecific mammalian crosses exhibit signals of maternal- and paternal-excess in line with differences in parental conflict despite no obvious differences in mating system (Reviewed in: Gray 1972; Vrana 2007; Brekke and Good 2014). Whether these parent-of-origin growth effects stem from yet undescribed differences in mating system or other demographic factors can elucidate the importance of different life history factors in shaping the severity of parental conflict in nature.

Criterion 2: Phenotypic asymmetry in the hybrid offspring

Conflict between parents arises via the evolution of resource-soliciting paternally derived alleles and the compensatory evolution of resource-repressive maternally derived alleles. Therefore crosses between individuals from populations with different conflict histories should reveal these differences in paternal solicitation and maternal repression, manifesting as asymmetric phenotypes in development and size in reciprocal hybrids (Haig and Westoby 1989; Brandvain and Haig 2005, 2018; Vrana 2007). When sired by individuals with a stronger history of conflict, hybrids should exhibit signs of overgrowth in the embryo and/or supporting structures (i.e. paternal-excess). Conversely, when individuals with stronger conflict histories are the dams, growth should be restricted (i.e. maternal-excess; Vrana 2007; Brekke and Good 2014; Rebernig et al. 2015; Lafon-Placette et al. 2018; Arévalo and Campbell 2020; Coughlan et al. 2020; Arévalo et al. 2021; İltaş et al. 2021; Sandstedt and Sweigart 2022; Petrén et al. 2023).

In angiosperms, many species exhibit asymmetries in reciprocal F1 seed size, viability, and endosperm development (Rebernig et al. 2015; Garner et al. 2016; Lafon-Placette et al. 2018; Raunsgard et al. 2018; Coughlan et al. 2020; Sandstedt et al. 2020; İltaş et al. 2021; Sandstedt and Sweigart 2022; Petrén et al. 2023). These parent-of-origin effects on seed size are largely caused by parent-of-origin specific development of the endosperm, rather than the embryo. In angiosperms, endosperm can either grow via nuclear division until a certain developmental stage after which these nuclei cellularize (i.e. nuclear-type endosperm), or each nuclear division can be accompanied by a cellular division (i.e. cellular-type endosperm). In paternal-excess crosses, endosperm often displays delayed development: in organisms with nuclear endosperm development, endosperm fails to cellularize, while in organisms with cellular endosperm development cells continuously grow, but fail to divide further. In both cases, paternal-excess seeds often exhibit an overly large endosperm (Rebernig et al. 2015; Lafon-Placette and Köhler 2016; Oneal et al. 2016; Lafon-Placette et al. 2017, 2018; Roth et al. 2018a; Coughlan et al. 2020; İltaş et al. 2021; Sandstedt and Sweigart 2022). In maternal-excess crosses, endosperm cells exhibit precocious growth—endosperm cellularizes prematurely in nuclear-type species or cell division occurs rapidly in species with cellular-type endosperm, in both cases resulting in less endosperm than intraspecific seeds (Rebernig et al. 2015; Lafon-Placette and Köhler 2016; Oneal et al. 2016; Lafon-Placette et al. 2017, 2018; Roth et al. 2018a; Coughlan et al. 2020; İltaş et al. 2021; Sandstedt and Sweigart 2022). Notably, a recent study by Sandstedt and Sweigart (2022) revealed that paternal- and maternal-excess phenotypes manifested in the chalazal haustorium, an endosperm tissue that serves as a maternal–fetal interface and controls nutrient transport. In paternal-excess crosses, the chalazal haustorium was enlarged and remained intact later than intraspecific crosses, while in maternal-excess crosses, this tissue was diminished and degraded prematurely. These parent-of-origin growth effects also have asymmetric effects on viability; in hybrid crosses with asymmetric viability, paternal-excess phenotypes are generally more lethal (Rebernig et al. 2015; Roth et al. 2018a; Coughlan et al. 2020; Sandstedt et al. 2020; İltaş et al. 2021). Inappropriate development of the endosperm, rather than incompatibilities in the embryo itself, likely causes seed failure, as dissected embryos from normally inviable seeds are often viable when grown on nutrient-rich media (Rebernig et al. 2015; Lafon-Placette et al. 2017, 2018; Sandstedt and Sweigart 2022).

Eutherians also often exhibit significant parent-of-origin effects on growth and development. Phenotypic asymmetry is common in reciprocal mammalian crosses (Gray 1972). For example, the F1 hybrids between closely related Mus and Peromyscus species discussed above exhibit striking asymmetries in embryo and/or placenta size (Rogers and Dawson 1970; Foltz 1981; Zechner et al. 1996; Vrana et al. 1998, 2000; Duselis and Vrana 2007; Gutierrez-Marcos et al. 2012; Brekke and Good 2014; Brekke et al. 2016, 2021; Arévalo and Campbell 2020; Arévalo et al. 2021). Similar patterns are well documented in reciprocal crosses between the two dwarf hamster species, Phodopus sungorus and Phodopus campbelli; hybrids sired by P. campbelli exhibit excessive embryo and placental growth which is ultimately lethal to the dam and offspring, while the reciprocal cross yields smaller embryos and placenta (Brekke and Good 2014; Brekke et al. 2016, 2021). In the hybrids between lions and tigers described above, paternal-excess hybrids can be 150% the size of either parent, with some hybrids weighing as much as both parents combined (Gray 1972), while the reciprocal cross is significantly smaller than either parent (Vrana 2007). Within humans, standing genetic variation in imprinting is associated with several diseases, many of which involve parent-of-origin growth phenotypes that persist into adulthood (Eggermann et al. 2008; Buiting 2010; Haig 2010). Thus, unlike in angiosperms, many cases of maternal- and paternal-excess arise in both the embryo and placenta and continue through adulthood. Intriguingly, in poeciliid fishes, reciprocal crosses between populations that differ in the extent of viviparity do not show reciprocal F1 size differences, highlighting that the evolution of hemochorial placentation (i.e. placentation in which direct and dynamic nutrient exchange is mediated by fetal access to the maternal bloodstream) and its connection with imprinting underlies these reciprocal F1 phenotypes (Schrader and Travis 2008). While the observed growth asymmetries in mammals fit the parental conflict model, it also is consistent with alternative hypotheses, such as the maternal–fetal coadaptation theory. Disentangling the role of conflict versus maternal–fetal coadaptation will require a better understanding of the genetic basis of these parent-of-origin growth phenotypes.

Criterion 3: Loci controlling dosage-sensitive processes should evolve rapidly and underlie transgressive growth phenotypes

Although it is well established that imprinted genes function within placental mammals and angiosperms to mediate parent-of-origin specific growth in the embryo and/or the placenta/endosperm (see “Genomic imprinting in mammals and angiosperms”; Table 1), antagonistic evolution precipitated by parental conflict should manifest as the rapid and continual turnover of genes involved in nutrient allocation, consequently resulting in divergence in imprinted expression between close relatives (Moore and Haig 1991; Zeh and Zeh 2000; Brandvain and Haig 2018). Under this model, parental conflict predicts that the parent-of-origin growth effects described above should be caused by divergence in the expression of parent-of-origin effect alleles among closely related species, wherein species with stronger histories of parental conflict should possess PEGs and MEGs that are better able to acquire or restrict resources than species with a weaker history of parental conflict. This persistent turnover of alleles caused by successive rounds of antagonistic evolution should leave signals of positive selection at the molecular level, with stronger signals of selection in populations that are inferred to have a stronger history of parental conflict (Moore and Haig 1991; Zeh and Zeh 2000; Johnson 2010; Crespi and Nosil 2013; Coughlan 2023c). Below we examine the evidence that parental conflict has shaped divergence between closely related species by evaluating the evidence for positive selection on imprinted genes and assessing whether divergence between close relatives in imprinting influences parent-of-origin effects on growth and development.

Angiosperms and eutherians may exhibit different levels of constraint in imprinted expression (Waters et al. 2013; Hatorangan et al. 2016; Lafon-Placette et al. 2018; Roth et al. 2018b; Picard and Gehring 2020; Richard Albert et al. 2023; Rodriguez-Caro et al. 2023). In angiosperms, imprinting is lowly conserved (Waters et al. 2013; Florez-Rueda et al. 2016; Hatorangan et al. 2016; Lafon-Placette et al. 2018; Roth et al. 2018b, 2019; Picard and Gehring 2020; but also see Klosinska et al. 2016). For example, in tomatoes only ~10% of MEGs and ~21% of PEGs are conserved across 3 species that are ~550,000 years diverged (Städler et al. 2008; Roth et al. 2018b). In Capsella only 2 PEGs and 1 MEG are conserved across another trio of species that diverged < 1,000,000 years ago (out of 152 and 56 PEGs and MEGs respectively; Hurka et al. 2012; Lafon-Placette et al. 2018). Strikingly, up to 10% of imprinted genes are variably imprinted in A. thaliana and maize (Waters et al. 2013; Pignatta et al. 2014, 2018). In contrast, in placental mammals the imprinted status of several canonical imprints are highly conserved across mammals (although there is variation in expression level; McVean and Hurst 1997; Wang et al. 2013; Babak et al. 2015; Kobayashi 2021; Rodriguez-Caro et al. 2023), particularly those expressed in the embryo (Richard Albert et al. 2023). Conversely, imprinted genes expressed in the placenta are less conserved in imprinted status, particularly non-canonical imprints (Richard Albert et al. 2023). Thus, despite their fundamental role in development, many imprinted loci do not appear to be evolutionarily constrained in angiosperms, rather, abundant standing genetic variation allows for rapid divergence in imprinting. In contrast, although some imprinted genes appear somewhat evolutionarily flexible in mammals (particularly in the extent of their expression), others are deeply conserved, suggesting significant constraint.

Under parental conflict, antagonistic evolution should also cause rapid allelic turnover, resulting in signals of positive selection. Some evidence for positive selection on imprinted genes has been found in angiosperms (Waters et al. 2013; Hatorangan et al. 2016; Tuteja et al. 2019). In Capsella and Arabidopsis, PEGs tend to exhibit stronger signals of positive selection (Hatorangan et al. 2016; Tuteja et al. 2019). Although not exclusively imprinted genes, some of the highest rates of molecular evolution in Arabidopsis are seed-associated genes, particularly genes expressed in the chalazal endosperm (Geist et al. 2019). In mammals, fewer studies have sought to quantify patterns of molecular evolution at imprinted genes. Earlier work in rodents identified signals of strong positive selection in placenta-associated genes (again, these are not necessarily imprinted genes; Chuong et al. 2010). More recently, work in three closely related Mus species/subspecies shows divergence in a small number of imprinted genes in both the extent of parental bias and total expression levels (Rodriguez-Caro et al. 2023). Moreover, work in angiosperms provides direct comparisons of the number or extent of imprinting as a function of genome strength; outcrossers tend to exhibit more imprinted genes and elevated rates of adaptive evolution at such genes (Spillane et al. 2007; Miyake et al. 2009; Wolff et al. 2011; Lafon-Placette et al. 2018). In particular, the number of PEGs and their expression levels scales with inferred genome strength in Capsella and tomato (Lafon-Placette et al. 2018; Roth et al. 2018a, 2018b; Städler et al. 2021). To our knowledge, these types of comparisons are rare in mammals. However, in the Mus example above, M. musculus domesticus exhibited lineage-specific evolution relative to either M. m. musculus or Mus spretus. In this trio, M. m. domesticus is polyandrous and has a significantly larger effective population size than M. m. musculus (which is also polyandrous) or the monogamous M. spretus (Cassaing and Isaac 2007; Phifer-Rixey et al. 2012; Rodriguez-Caro et al. 2023). Differences in demography and mating system may have contributed to a stronger history of conflict in M. m. domesticus, resulting in a faster rate of lineage-specific evolution. Intriguingly, unlike in Capsella and Arabidopsis, where PEGs more readily show signals of positive selection, genes showing lineage-specific evolution in Mus were largely MEGs (Rodriguez-Caro et al. 2023). Overall, despite some evidence for rapid evolution of imprinted genes in both eutherians and angiosperms, an exciting avenue of future research would be to explicitly link rates of evolution in imprinted expression with factors that should influence the strength of conflict.

Parental conflict theory also predicts that antagonistic evolution should specifically target paternally derived resource-acquiring and maternally derived resource-repressive alleles. Direct evidence that antagonistic evolution has driven the fixation of paternally derived resource-acquiring and maternally derived resource-repressive alleles is limited. Within species, compelling evidence for natural variation in paternal resource-acquiring alleles comes from Pignatta et al. (2018). Here, the authors show that HDG3 (a class IV homeodomain leucine zipper transcription factor) is variably imprinted within A. thaliana, acting as a PEG in some accessions, but not in others. Moreover, when imprinted, this PEG is expressed specifically in the endosperm—including the chalazal endosperm—and its imprinted expression promotes seed growth (Pignatta et al. 2018). Intriguingly, imprinting and its downstream effects can be partially induced by changing the methylation status of HDG3, rather than the DNA sequence itself, suggesting that genetic variation in HDG3 may not determine its variable imprinting status. Instead, variation in cis—perhaps due to proximal TEs or associated small RNAs—may underlie natural variation in imprinting (Pignatta et al. 2018).

Although much evidence from both genetic mapping and hybrid expression studies is in line with parental conflict in angiosperm and placental mammal speciation, few studies show a direct connection between parental conflict, divergence in imprinting, and hybrid inviability (Kradolfer et al. 2013; Wolff et al. 2015; Butel et al. 2023). In both angiosperms and placental mammals, studies have associated parent-of-origin nuclear loci with inviability, in line with a role of imprinted loci in lethal maternal- or paternal-excess phenotypes (Zechner et al. 1996; Vrana et al. 1998, 2000; Hemberger et al. 1999; Josefsson et al. 2006; Duselis and Vrana 2007; Rebernig et al. 2015; Brekke et al. 2016, 2021; Garner et al. 2016; Lafon-Placette et al. 2018; Arévalo and Campbell 2020; Arévalo et al. 2021; Dziasek et al. 2021). For example, inviable hybrid seeds between A. thaliana and A. arenosa exhibit loss of imprinting and consequently overexpression of the PEG PHE1 (Josefsson et al. 2006). Disrupting PHE1 expression significantly increased seed viability, suggesting that PHE1 misexpression—whether caused by genetic variation at PHE1 or in a regulatory element upstream—confers seed inviability (Josefsson et al. 2006; Batista et al. 2019). Moreover, many studies show significantly disrupted gene regulation in hybrid placenta/endosperm, including misexpression of normally imprinted genes (Vrana et al. 1998, 2000; Josefsson et al. 2006; Wiley et al. 2008; Burkart-Waco et al. 2012, 2015; Brekke and Good 2014; Rebernig et al. 2015; Brekke et al. 2016, 2021; Florez-Rueda et al. 2016; Roth et al. 2019; Arévalo and Campbell 2020; Arévalo et al. 2021; Dziasek et al. 2021), elevated TE activity (Josefsson et al. 2006), reduced small RNA expression (Florez-Rueda et al. 2016, 2021), and decreased levels of methylation and chromosome condensation (Dziasek et al. 2021). In eutherians, disrupted placental development is often associated with misexpression of autosomal imprinted genes as well as genetic variants on the X-chromosome (Zechner et al. 1996; Vrana et al. 1998, 2000; Hemberger et al. 1999; Brekke et al. 2016, 2021; Arévalo et al. 2021), suggesting an X-autosomal incompatibility may underlie dysgenic placenta growth. Indeed, recent work in Phodopus showed that the degree of misexpression of autosomal imprinted gene networks depends on the identity of the inherited X-chromosome. For example, Tfpi2—an autosomal MEG associated with placental overgrowth when disrupted—was significantly downregulated when maternally inherited from P. campbelli and combined with P. sungorus X-chromosome (Brekke et al. 2021). This may hint at a genetic model where imprinted loci on the X-chromosome regulate autosomal imprinted expression, and mismatches between the X-chromosome and autosomes lead to inappropriate autosomal expression with lethal effects (Fig. 4c).

Classical models for how differences in imprinted expression could cause hybrid lethality via a Dobzhansky–Muller incompatibility invoked a more “tit for tat” evolution of imprinting—wherein the evolution of a resource-soliciting PEG is compensated for by the evolution of a resource-restrictive MEG (i.e. Fig. 4; Haig and Westoby 1989; Zeh and Zeh 2000, 2008; Wilkins and Haig 2001; Babak et al. 2015; Fishman and Sweigart 2018; Coughlan 2023a, 2023b; Reifová et al. 2023). Under this model, hybrids would exhibit both a loss of imprinting and transgressive expression and dosage imbalances of maternal:paternal gene products (Fig. 4a). Yet, other genetic models are possible (i.e. Brekke et al. 2021). Another class of models posits that parent-of-origin growth defects in hybrids stem from divergence in any genes that underlie dosage-sensitive processes, wherein hybrids exhibit dosage imbalances of maternal:paternal expression (Dilkes and Comai 2004; Städler et al. 2021; Satyaki and Gehring 2022). Dosage imbalances could manifest due to divergence in the extent to which imprinted genes are expressed, TE positioning, genes controlling small RNA activity, X-linked genes that regulate autosomal imprinting, or bi-parentally expressed regulators of imprinting (Fig. 4b–d) In all cases, this allelic divergence would precipitate dysregulated imprinted expression and an imbalance of maternal:paternal expression in hybrids, ultimately causing inviability (we note, however, that few models predict complete loss of imprinting; Dilkes and Comai 2004; Josefsson et al. 2006; Roth et al. 2019; Städler et al. 2021). Understanding the relative importance of these genetic models will require integrating traditional genetic mapping with expression studies and functional validation.

Criterion 4: The possibility of viability restoration via induced genomic perturbation

If dysgenic hybrid growth is caused by an imbalance of maternal:paternal gene expression, then genetic manipulations that restore balance should rescue inviable hybrids and reinstate proper development (Johnston et al. 1980; Johnston and Hanneman 1982; Haig and Westoby 1989, 1991). The dosage imbalance model further suggests that the extent of transgressive hybrid phenotypes should be predictable; crosses between taxa with similar conflict histories should have a more balanced dosage of maternal:paternal expression, leading to minimal parent-of-origin effects on growth, and consequently, lower levels of inviability. These two predictions—that hybrid growth phenotypes should be predictable based on differences in the parents’ genome strength, and that genetic perturbations that restore imbalances of maternal:paternal expression should rescue hybrid development—are the two observations that sparked EBN theory (Johnston et al. 1980; Johnston and Hanneman 1982). Although EBN theory has been well tested (and supported) in angiosperms, genome-engineering work in mice may also provide early evidence that genetic perturbations can restore appropriate growth phenotypes in mammals.

Given the feasibility of collecting naturally occurring polyploid populations as well as creating viable synthetic polyploids or mutant lines, angiosperms offer a unique opportunity to investigate viability restoration via genomic perturbation. Many studies find that hybrid seed inviability between species with different histories of conflict can be reversed if the species with the stronger conflict history is instead crossed with a tetraploid (Lafon-Placette and Köhler 2016; Lafon-Placette et al. 2017, 2018; Coughlan et al. 2020; Sandstedt and Sweigart 2022). These tetraploids may be naturally occurring or synthetically created from the species with the weaker history of conflict (Lafon-Placette and Köhler 2016; Lafon-Placette et al. 2017, 2018; Coughlan et al. 2020; Sandstedt and Sweigart 2022). In either case, doubling gene expression of the “weaker” parent is sufficient to counteract the expression of the “stronger” parent, in line with a model that dosage imbalances mediate dysgenic hybrid phenotypes. An alternative approach to test the dosage imbalance model is to leverage gene knockouts to re-establish dosage equilibrium. For instance, the paternal-excess phenotype displayed by fis2 mutants in A. thaliana can be reversed by downregulating AGL62, a MEG whose loss of function normally causes precocious endosperm cellularization (Hehenberger et al. 2012). Similarly, the viability of triploid seeds exhibiting paternal-excess phenotypes can be restored via loss-of-function mutations in PEGs (Kradolfer et al. 2013; Wolff et al. 2015; Jiang et al. 2017).

Unlike angiosperms, mammalian tetraploids are extremely rare and difficult to synthetically create as polyploidy is generally lethal (Niebuhr 1974; Orr 1990; Andreassen et al. 2001; Eakin and Behringer 2003). This limits the possibility of testing viability restoration via genome doubling. Nonetheless, several studies bypassed this limitation by using early embryonic tetraploid cells to rescue diploid mutant mice with fatal extraembryonic defects (Nagy et al. 1993; Eakin and Hadjantonakis 2006). For example, mice that paternally inherited knockouts of Peg10 showed lethal growth reduction in the embryo and placenta (Ono et al. 2006). However, aggregating diploid mutant cells with wild-type tetraploid embryonic cells rescued viability by restoring normal placental development (Ono et al. 2006). These experiments do not establish a direct connection between restoring dosage balance and rescuing viability since the tetraploid cells are restricted to the ephemeral extraembryonic tissues and do not contribute to the formation of the embryo proper (Nagy et al. 1993). However, viability restoration is thought to be possible as these tetraploid cells carry the correct dosage of genes controlling placental growth that the embryo lacks (Nagy et al. 1993; Guillemot et al. 1994; Eakin and Behringer 2003). More direct evidence, albeit scarce, is in mutant mice carrying a paternal duplication of the imprinted distal chromosome 2 (Ball et al. 2013). These mice are inviable with paternal-excess phenotypes and overexpress the PEG Gnasxl. Upregulating the MEG Gnas restored viability and normal development (Ball et al. 2013). In another example, although the viability was not restored, the knockout of the PEG Igf2 partially rescued placental overgrowth in mice with null-mutations of the MEG Phlda2 (Frank et al. 2002). Overall, genetic perturbations can serve as evidence that dosage-sensitive processes underlie hybrid transgressive phenotypes in both angiosperms and eutherians, although this work is much less established in eutherians. Despite the strengths of these approaches, identifying the precise genes involved in restoring dosage balance is challenging when the whole genome is duplicated. Therefore, future research to identify the precise genetic interactions that mediate dosage-sensitive processes can offer insight into the exact mechanisms that underlie hybrid transgressive growth, and reveal evolutionary models of divergence between species in these dosage-sensitive processes.

Conclusions and moving forward

Despite their myriad differences, eutherians and angiosperms are united in the convergent evolution of extraembryonic nutritive tissues that are essential for embryo development, and the subsequent conflict that can arise as a byproduct of these key innovations. There are many differences in the timing and spatial extent of imprinting between angiosperms and placental mammals. In eutherians, imprinting manifests in many tissues and persists into adulthood, while in angiosperms imprinted expression is largely restricted to the endosperm and imprints must be reset every generation. Yet, several aspects of the molecular machinery and genes involved are shared across taxa; a striking example of convergence at the molecular level. Using four criteria, we examine if/how parental conflict has shaped evolution within both angiosperms and eutherians: 1) The strength of selection via parental conflict should be driven by the extent of sibling relatedness. Differences in mating system, demography or life history factors can influence the extent of multiple paternity in nature, thereby shaping the strength of parental conflict. Although some of these hypotheses have been well tested, an integration of more life history knowledge may provide valuable tests of these predictions and reveal if/how these factors differentially affect parental conflict in angiosperms versus eutherians. 2) Crosses between populations with different histories of parental conflict should result in reciprocal F1 phenotypes. In plants, these manifest largely in the endosperm, specifically in tissues that mediate nutrient transfer at the maternal:filial interface. In mammals, both the placenta and embryo exhibit these patterns. One intriguing avenue for future research is to more thoroughly determine if parental conflict is more apparent in specific tissues within the placenta/endosperm that more directly function in nutrient transfer. 3) Differences in parent-of-origin growth effects should be caused by differences in imprinted expression between populations. Specifically, these imprinted alleles should function to moderate nutrients between mothers and offspring, and should show signals of positive selection. We find that evidence for this prediction is mixed. While a handful of studies have found evidence for positive selection on imprinted loci in both angiosperms and eutherians, direct comparisons between closely related species are rare. Moreover, work showing that divergence in imprinting directly causes reciprocal intra-ploidy F1 size differences has yet to be shown; reciprocal F1s frequently show transgressive expression of TEs, small RNAs, and imprinted genes at a genome-wide level as opposed to being confined to specific pairs of dosage-sensitive interacting loci. Future mapping will be necessary to resolve which genetic models best describe dysgenic expression in hybrids and its connection with hybrid inviability. 4) Genetic perturbations that restore dosage imbalances should rescue offspring development. We find that this prediction is supported by manipulations of overall ploidy in angiosperms or via gene knockouts in Mus. Nonetheless, future work that targets specific genes underlying dosage-sensitive processes will be more powerful. Moreover, while we find support for the role of parental conflict in generating divergence in imprinted expression, other selective pressures likely also play a role in shaping imprinting diversity and divergence. Nonetheless, while many questions remain, the striking convergence between angiosperms and eutherians suggests that parental conflict is likely a ubiquitous force in evolution with biological differences between these taxa highlighting how/when/where/ and to what extent parental conflict manifests.

Acknowledgments

We are deeply grateful to Lila Fishman for organizing the AGA Presidential Symposium “Selfish Evolution: Mechanisms and Consequences of Genomic Conflict” as well as organizing this special issue. We would also like to thank the Coughlan lab, especially Megan Frayer, for helpful conversations on the topics discussed herein.

Funding

This work was funded by the National Institutes of Health (grant ID 1R35GM150907) to JMC.

Conflict of interest statement

The authors declare no conflict of interest.

Data availability

The manuscript does not present any new data.

References