Controlled experiments in mice and investigation of placental pathologies in humans have made it possible to understand the need for a genome derived from specific male and female parental contributions, which must be spatio-temporally regulated for normal development (Hall, 1990; Reik and Walter, 2001). Such a conception of heredity, apparently contravening Mendel’s laws, has fueled the emergence of a specific field of research dedicated to characterizing the underlying phenomenon, known as genomic imprinting, and identifying the molecular mechanisms behind the non-equivalence of parental genomes (Ferguson-Smith and Bourchis, 2018; Tucci et al., 2019).

Our understanding of how regions affected by imprinting have evolved and function has expanded considerably (Llères et al., 2021; Kaneko-Ishino and Ishino, 2022; Richard Albert et al., 2023) and we today have increasingly sophisticated resources and tools (Hubert and Demars, 2022; Jima et al., 2022; Akbari et al., 2023) to identify and accurately report the links between nuclear architecture, Imprinting Control Regions (ICRs), imprinted genes, other genes including different RNA species and, finally, phenotypes. While cytosine methylation is key to almost all regulations that affect it, including transgenerational maintenance, imprinting involves multi-scale mechanisms (Monk et al., 2019), giving grounds for multi-omics approaches (Hubert and Demars, 2022). A diversity of non-exclusive patterns, such as the recruitment of CCCTC-Binding Factor (CTCF) (Noordermeer and Feil, 2020), the transcriptional interference of long non-coding RNA (lncRNA) (MacDonald and Mann, 2020) or the presence of microRNA (miRNA) clusters (Malnou et al., 2019), have, for example, been highlighted at imprinted loci. In addition, histone marks affect the methylation at ICRs, representing another layer of regulatory complexity (Sanli and Feil, 2015). Interestingly, the possibility of DNA methylation-independent, histone-based imprinting has been identified in rodents and termed as noncanonical imprinting (Raas et al., 2022; Inoue, 2023; Richard Albert et al., 2023). The rapid initial accumulation of knowledge on imprinting has been largely enabled by the mouse model (Swain et al., 1987; Li et al., 1993; Greally et al., 1998). However, livestock species have also led to very illuminating results (Cockett et al., 1996; Feil et al., 1998; Jeon et al., 1999; O'Doherty et al., 2015), due to their presence in many vertebrate clades, with a particular importance in the Cetartiodactyla order that includes ruminants and pigs, and the availability of experimental livestock resources suitable for research on imprinting. We therefore wish to emphasize livestock research on imprinting, as we believe it can provide valuable basic knowledge and models to link multi-level molecular variations to complex phenotypical changes (Magee et al., 2014; Daigneault, 2022).

In the present article, we first summarize some of the key contributions regarding imprinting in livestock species, primarily around the identification of imprinted genes, from a broad and historical perspective. We then discuss in more detail molecular mechanisms linking variability at imprinted domains, including IGF2–H19, CDKN1C–KCNQ1 and DLK1–GTL2, to developmental effects on muscle and growth phenotypes, as established in livestock species. We finally review several current topics in livestock research aiding deeper understanding of the role of imprinting in complex traits related to energy supply or intake, with broad significance for mammalian biology and health.

Initially uncovered in the 80s through iconic murine pronuclear transplantation studies, the presence of imprinting in mammalian genomes led to new methodological developments to detect the parent-of-origin (PofO)-specific expression of genomic regions and their associated functions. They notably consisted of designing specific genetic constructs (Cattanach, 1986; Lee et al., 1993) then molecular biology approaches (Hatada et al., 1993) in mice, allowing the identification of the first imprinted genes in the early 90s (Igf2, Igf2r and H19). In humans, expression studies using in particular fetal tissues allowed determination of the imprinting status of many candidate genes for imprinting (Zhang and Tycko, 1992; Hatada et al., 1996; Blagitko et al., 2000; Murphy et al., 2001) leading to the identification of dozens of imprinted genes but also evidence for differences compared to mice (Kalscheuer et al., 1993; Monk et al., 2006).

In order to check theoretical predictions about the link between imprinting and viviparity, a subset of genes systematically identified as imprinted in placental mammals was found to be imprinted as well in different marsupial species (O'Neill et al., 2000; Suzuki et al., 2005), but not in monotremes (Killian et al., 2001) or in birds (O'Neill et al., 2000). While shedding light on the ancestral origin of genomic imprinting, such single-gene isolation studies were limited to highly conserved imprinted genes, such as in the IGF2 (Insulin Growth Factor 2) pathway. In 2002, humans and mice became the only mammals with a reference sequence, allowing early genome-wide studies for novel imprinted genes (Luedi et al., 2005; Luedi et al., 2007), in agreement with estimates suggesting that from one to a few percent of all mammalian genes could be subjected to imprinting. Some livestock species soon experienced comparable developments in sequencing technologies, improving the phylogenetic-scale knowledge on imprinting. With pig, cattle and sheep displaying more than 30 experimentally-validated imprinted genes, Cetartiodactyla is one of the most documented mammalian orders on imprinting, alongside primates (Cheong et al., 2015; Chu et al., 2021; Jima et al., 2022) and rodents (Raas et al., 2022; Richard Albert et al., 2023) (Figure 1). Not only do livestock species include representatives from various mammalian lineages but also from key outgroups, i.e., birds and fish, providing opportunities for testing hypotheses on imprinting and epigenomic evolution in vertebrates.

Mainly driven by the needs of different agri-food sectors and the acceleration in the development of sequencing technologies in the 2000s, specific genomic resources began to emerge in animal species of economic importance. In 2015, the NCBI Reference Sequence Database reported at least a chromosome-level assembly reference sequence for a total of 14 of such species, with chicken (Wallis et al., 2004), ruminants (The Bovine Genome Sequencing and Analysis Consortium et al., 2009; The International Sheep Genomics Consortium et al., 2010; Dong et al., 2013) and pigs (Humphray et al., 2007) among the first. At the same time, these species have been subjected to the development of increasingly dense microarray platforms (Sellner et al., 2007). These early annotated genomes and high-throughput genotyping tools have promoted the use of genetic information for breeding, establishing species-specific genetic evaluation procedures (Hayes et al., 2009; Legarra et al., 2011; Ibáñez-Escriche et al., 2014) and enabling functional genomics approaches for increasingly fine and complex traits (Hayes et al., 2010; Do et al., 2013). This also brought higher-resolution analyses of imprinting (Bischoff et al., 2009; Chen et al., 2016; Kenny et al., 2022), leading to the identification of dozens of imprinted genes in pigs, cattle and sheep, including large conserved imprinted domains affecting early development, such as IGF2–H19 and MEST (Rutkowska and Lukaszewicz, 2019).

Combined with valuable pre-existing livestock resources, the increasing availability of Next-Generation Sequencing (NGS) platforms has resulted in diversified studies expanding several areas of our understanding of imprinting. Especially, results from non-mammalian livestock species with both adapted genomic and experimental resources made it possible to raise the question of the presence or absence of imprinting in egg-laying vertebrates (Frésard et al., 2014; Fan et al., 2016) at the genome level. As exemplified in chicken, several genome-scale studies across different tissues and developmental stages are advisable to better understand PofO-specific expression mechanisms among all possible forms of allelic imbalance (Zhuo et al., 2017). Robust tools and procedures are indeed essential when investigating potential imprinting patterns (Edwards et al., 2023), and transcriptomic studies involving reciprocal crosses have so far collectively concluded there were allele-specific expression rather than imprinting in chicken tissues (Frésard et al., 2014; Zhuo et al., 2017). In livestock mammals, NGS-based studies allowed identifying additional imprinted genes and started addressing different related regulatory processes (Congras et al., 2014; Chen et al., 2016; Duan et al., 2018). In addition, a growing body of research explores local mechanisms involved at imprinted loci with evo-devo or biomedical significance (Yu et al., 2018; Ahn et al., 2020; Li J. et al., 2022; Ahn et al., 2023).

Following the path taken by the biomedical research community, international initiatives promote the generation of increasingly comprehensive and accessible tools and resources in epigenomics and transcriptomics (Giuffra et al., 2019; Tixier-Boichard et al., 2021). Recent large-scale cattle studies show the interest of multi-omics not limited to coding regions to capture large proportions of heritability and better predict the phenome (Xiang et al., 2019; Xiang et al., 2023). Efforts in establishing large catalogs of functional elements has resulted in the public availability of datasets from adult, fetal tissues and single cells, including RNAseq, methylation, histone mark, CTCF occupancy and chromatin accessibility data in pigs, cattle and chicken (Kern et al., 2021; Pan et al., 2023). All this novel information on regulatory elements from a diversity of biological contexts (i.e., species, populations, developmental stages, tissues) provides an integrative framework that is particularly helpful to supporting new livestock studies on imprinting (Ahn et al., 2022; Bruscadin et al., 2022).

In livestock species, the major role played by imprinted genes is documented as they are part of the molecular architecture of some agronomic traits (O'Doherty et al., 2015). In pigs and sheep, two mutations affecting a production trait have been identified, and–along with a syndrome linked to imprinting disturbances in cattle–are examples of imprinting contribution to phenotypic variability.

Most of our understanding of the phenotypic consequences of imprinting comes from mouse models and human imprinting disorders (Peters, 2014). Imprinting disorders result in a range of clinical features including aberrant pre- and/or postnatal growth and abnormal feeding behaviours (Carli et al., 2020). Imprinted disorders are a group of congenital disorders with common underlying molecular alterations, including genetic abnormalities as well as aberrant epigenetic landscapes, which target imprinted genes. They show that imprinted genes are key loci involved in traits’ variability through their genetic and epigenetic PofO effects. Besides imprinted disorders, more and more studies pinpoint the role of imprinted genes and their PofO effects in the variability of complex traits (Lawson et al., 2013). In livestock species, some traits of major agronomic interest overlap with phenotypes affected in imprinting disorders such as birthweight, postnatal resources through mammary gland development or feeding behaviours, making livestock resources relevant for a better understanding of the role of imprinting in complex traits.

This review aimed to show how livestock species could contribute to essential information on genomic imprinting, from evolutionary conservation to pathophysiological models for specific human disorders, including complex molecular imprinting patterns affecting trait variation.

Overall, there is very little imprinting studies on non-mammalian vertebrates, a first obstacle being the difficulty of accessing or funding adequate experimental resources. The scarcity of such studies may also be partly explained by the widespread possibility of parthenogenesis in these species, which is generally seen as incompatible with imprinting. Non-mammalian vertebrates however show a remarkable diversity of embryo-related adaptations, allowing detailed research on embryonic resource allocation (Furness et al., 2019; Whittington et al., 2022). In both birds and teleost fish, the absence of an orthologue of DNMT3L, a key methylation enzyme in mammals, supports the absence of imprinting. Analyses carried out on whole embryos, adult brains or embryonic tissues always concluded that imprinting was absent in chicken (Frésard et al., 2013). Conversely, studies performed in fish are more contradictory, with Ma et al. showing the ntl gene as differentially methylated between sperm and egg in goldfish and a differential timing of transcription between paternal and maternal alleles in the embryo (Ma et al., 2011). Recent data from both laboratory and cultured teleost fish suggest the possibility that some imprinted genes, such as IGF2 and DLK1, present homologues subjected to imprinting-like patterns outside mammals (Du et al., 2022; Zhang et al., 2023), highlighting the need for further research on this topic.

Improved annotation of livestock genomes supported by international initiatives such as FAANG (The FAANG Consortium et al., 2015) and FarmGTEx (https://www.farmgtex.org/) should allow building a more comprehensive phylogenetic tree of imprinting evolution, including investigating potential early differentiation in some clades. As an example, assessment of specific imprinted genes in cattle and pigs showed that SLC38A4 appeared to be species-specific, with a non-imprinted pattern in cattle in all adult tissues evaluated (Zaitoun and Khatib, 2006) compared to mice, which harbor tissue-specific imprinting (Zaitoun and Khatib, 2006). Moreover, a fine-tuned pattern of imprinting has been shown for paternally-expressed NAP1L5, with an expression distribution in pig tissues differing from those in mice and cattle (Jiang et al., 2011). Genome-wide comparative analyses suggest that changes in methylation are a potentially important source for shaping lineage- and species-specific innovations, as exemplified with imprinting (Hu et al., 2023). Livestock mammals such as cattle and pigs display epigenomic-specific features that make them particularly attractive to investigate the contribution of imprinting and epigenomic evolution to phenotypic variation (Lu et al., 2021).

Several traits of agronomic interest overlap with major features of imprinted genes, particularly due to their known role in growth, development and behavioral processes. Although molecular tools targeting imprinting in livestock still need to be developed (Hubert et al., 2024), experimental livestock populations may help to better understand how imprinted genes affect performance (Figure 2). Not only do these populations display a broad range of relevant phenotypes for imprinting (Hinde et al., 2014; Gilbert et al., 2017; Khanal et al., 2017; Canario et al., 2022), but also they allow access to a full pedigree with parental origins either through nuclear family datasets (Khanal et al., 2017; Canario et al., 2022) or divergent lines selected over several generations (Gilbert et al., 2017). Thus, considering parental origins in future genome-wide association studies performed in livestock would enable significant advances on the functional impact of imprinted genes in the variability of important traits.

Apart from the bovine LOS that mimics BWS (Rivera et al., 2022), other livestock species such as pigs might constitute relevant pathophysiological models for imprinted disorders. In mice, targeted mutagenesis has provided highly valuable clues on the mechanisms involved in the regulation of imprinting. However, mouse models and patient data show discrepancies on both growth traits and inheritance patterns (Shmela and Gicquel, 2013), suggesting that mice do not seem appropriate for further analysis of pathophysiology for therapeutic purposes, making pigs a potential alternative model for imprinted disorders. Indeed, pigs are considered as one of the best animal generators of human disease models, because they share similar features with humans in physiology, anatomy and genome organisation (Groenen et al., 2012). The emergence of the CRISPR/Cas9 genome editing technology is revolutionizing porcine genome engineering (Yao et al., 2016). Recent studies have introduced epigenome-editing strategies which target enzymatic activity to introduce or remove an epigenetic mark at a defined genomic site (Kungulovski and Jeltsch, 2016). Given the crucial role of ICRs into the regulation of local imprinted genes and the disruption of DNA methylation observed there in various imprinted disorders (Soellner et al., 2017), epigenome-editing technologies might be applied to perturb DNA methylation at a dedicated ICR in an allele-specific manner to mimic a specific disorder Supplementary Figure S1. Epigenome-edited pigs might be promising pathophysiological animal models to investigate the long-term dynamics of imprinted disorders.