A genome-wide association study of mare fertility in the Pura Raza Español horse N. Laseca a,⇑, S. Demyda-Peyrás b,c, M. Valera d, M. Ramón e, B. Escribano f, D.I. Perdomo-González d, A. Molina a aDepartamento de Genética. Facultad de Veterinaria, Universidad de Córdoba, Campus de Rabanales, CN-IV km 396, 14071 Córdoba, España bDepartamento de Producción Animal, Facultad de Ciencias Veterinarias, Universidad Nacional de La Plata, Calle 60 y 118 s/n, La Plata 1900, Argentina cConsejo Nacional de Investigaciones Científicas y Tecnológicas (CONICET) La Plata, La Plata 1900, Argentina dDepartamento de Agronomía, Escuela Técnica Superior de Ingeniería Agronómica. Universidad de Sevilla, Ctra. Utrera, Km 1, Sevilla, Spain eCentro Regional de Selección y Reproducción Animal (CERSYRA), Av. del Vino, 10, 13300 Valdepeñas, Ciudad Real, Spain fDepartamento de Fisiología, Universidad de Córdoba, Campus de Rabanales, CN-IV km 396, 14071 Córdoba, Spain a r t i c l ei n f o Article history: Received 10 December 2021 Revised 25 January 2022 Accepted 27 January 2022 Available online 2 March 2022 Keywords: Association analysis Candidate genes Equine Reproductive traits Single-nucleotide polymorphisms a b s t r a c t Despite the economic importance of fertility for the horse industry, few efforts have been made to achieve a better understanding of the genetic mechanisms underlying its control. This is probably due to the difficulty of obtaining reliable phenotypes and the complexity of modelling the environmental and management factors. This work is novel in that we propose to use reproductive efficiency (RE) as an indicator of mare fertility. To achieve this, we performed a genome-wide association study in the Pura Raza Español horse aimed at identifying genomic variants, regions, and candidate genes associated with fertility in mares. The dataset included 819 animals genotyped with the Affymetrix AxiomTM Equine 670 K single-nucleotide polymorphisms (SNPs) Genotyping Array and the deregressed breeding values for RE trait, obtained using a ssBLUP model, employed as pseudo-phenotypic data. Our results showed 28 SNPs potentially associated with RE, which explained 87.19% of the genetic variance and 6.61% of the phenotypic variance. Those results were further validated in BayesB, showing a correlation between observed and predicted RE of 0.57. In addition, 15 candidate genes (HTRA3, SPIRE1, APOE, ERCC1, FOXA3, NECTIN-2, KLC3, RSPH6A, PDPK1, MEIOB, PAQR4, NM3, PKD1, PRSS21, IFT140) previously related to fertility in mammals were associated with the markers and genomic regions significantly associated with RE. To our knowledge, this is the first genome-wide association study performed on mare fertility. Ó2022 The Author(s). Published by Elsevier B.V. on behalf of The Animal Consortium. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Implications Reproductive traits are a critical factor for the profitability of equine farms. However, the information available is limited due to the difficulty in obtaining reliable phenotypes and the complex- ity of modelling the environmental factors. This is the first attempt to perform a genome-wide analysis study focused on mare fertility using a large cohort of horse genotypes and a pseudo-phenotypic value obtained by analysing more than 344 000 reproductive records of Pura Raza Español horses. We determined (and vali- dated) the existence of several candidate regions and genes that might provide insights for genomic or marker-assisted selection in the mares reproduction. Introduction Fertility is a key factor in the economic success of livestock pro- duction systems. Nevertheless, the horse is probably the domestic species in which natural and artificial selection has had the least influence in fertility. Since its domestication about 8 000 years ago (Moazemi et al., 2020; Orlando, 2020), the horse has been used for work, warfare, leisure or sport, activities in which fertility is often considered less important than other traits (Palmer and Chavatte-Palmer, 2020). For that reason, it is not often included in breeding programmes as a selection objective. In addition, the difficulty of accounting for a criterion with sufficient heritability as a measure of reproductive aptitude in mares also hinders its inclusion in breeding programmes as well as scientific efforts. However, it has been demonstrated that fertility is still a critical factor for the profitability of horse farms (Gómez et al., 2020). In addition, the reproductive management and environmental condi- https://doi.org/10.1016/j.animal.2022.100476 1751-7311/Ó2022 The Author(s). Published by Elsevier B.V. on behalf of The Animal Consortium. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). ⇑Corresponding author. E-mail address:ge2lagan@uco.es (N. Laseca). Animal 16 (2022) 100476 Contents lists available at ScienceDirect Animal The international journal of animal biosciences
tions have a crucial effect on the reproductive efficiency of mares. Even though fertility is considerably lower in horses than in other domestic species (Perdomo-González et al., 2021), the existence of ageneticcomponentaffectingthistraitiswell-demonstrated (Gómez et al., 2020; Mantovani et al., 2020; Todd et al., 2020). To date, few studies have been carried out which allow us to better understand the genomic mechanisms underlying fertility in mares (Laseca et al., 2021a; 2022). This is probably due to the lack of large, reliable phenotypic datasets, which impairs the mod- elling of environmental and management factors (e.g. age, nutri- tion, training, temperature at mating and breeding season, etc.) affecting this character. Furthermore, fertility is a complex poly- genic trait with low heritability (Mahon and Cunningham, 1982) and is influenced by a large number of genes, each with a small absolute effect, and due to the high level of linkage disequilibrium between genomic variants, shown particularly in the Pura Raza Español (PRE) breed (Poyato-Bonilla et al., 2022), is it difficult to pinpoint causal variants for complex traits. For this reason, most of the reproductive traits usually employed to evaluate the fertility of mares (total foaling number, age at first foaling, average interval between first and second foaling, average inter-foaling interval, age at last foaling, and productive life (Gómez et al., 2020; Perdomo- González et al., 2021) tend to have low heritabilities. Therefore, its use as a selection criterion will only produce moderate results intermsofphenotypicimprovement.However,Perdomo- Gonzalez et al., 2020 have recently developed a new phenotypic trait (reproductive efficiency,RE), which is able to estimate the fer- tility of a mare with great accuracy based on the analysis of pedi- gree records. This trait, which was validated using a large sample population of 300 000 foaling records, showed a moderate to high heritability (h2 = 0.23) but of a greater magnitude than the classics trait previously used as reproductive criteria in mares (Sairanen et al., 2009; Wolc et al., 2009; Gómez et al., 2020). This parameter, which could be easily estimated in all the mares existing in a stud- book, is an interesting candidate to be included in a breeding pro- gramme with the aim of improving the maternal fertility of the breed. However, it also allows its use in genomic studies aimed at determining genomic regions associated with fertility in mares. Nowadays, advances in high-throughput genotyping technolo- gies (medium and high-density chips) and sequencing technolo- gieshaveenabledustoidentifymanysingle-nucleotide polymorphisms (SNPs) associated with phenotypic traits (Laseca et al., 2021a). This genomic revolution has produced significant advances in our understanding of complex traits. Most of these associations were determined by performing genome-wide associ- ation studies (GWAS), which are now considered the most power- ful tool to screen and determine (at least partially) the genomic architecture of qualitative and quantitative traits, thus improving theaccuracyandpersistencyofgenomicselection (VanRaden et al., 2017). In horses, several GWAS studies have recently been reported for several traits, such as conformation (Al Abri et al., 2018), racing performance (Pereira et al., 2018), gait (Fonseca et al., 2017), or diseases (Shrestha et al., 2020). However, despite the existence of GWAS in female reproductive traits in several live- stock species, such as cattle (Keogh et al., 2021), pigs (Wang et al., 2018), sheep (Smołucha et al., 2021), and goat (Islam et al., 2020), these kinds of reports in horses have only been performed in stal- lions (Gottschalk et al., 2016; Gmel et al., 2021), and so far not in mares. The Pura Raza Español is one of the oldest European horse breeds. Its breeding programme, which nowadays includes over 260 000 individuals from 63 countries, is managed by the Aso- ciación Nacional de Criadores de Caballos de Pura Raza Española (ANCCE). Historically, PRE horses have been selected based mostly on aesthetic criteria. Nevertheless, since three decades ago, several additionalcriteriarelatedtoaptitudeforsportivedisciplines (mainly in classical dressage), as well as morphology (in relation to functionality through linear morphological qualification) have been included as selection objectives in the breeding programme. However, our research group has recently analysed several new traits focused on the fertility of individuals (Gómez et al., 2020; Perdomo-Gonzalez et al., 2020; 2021) which are nowadays being taken into consideration in the breeding selection criteria for fertil- ity. This fact, together with the long, reliable pedigree (kept by the ANCCE since 1912), and the existence of a reliable, well-developed breeding scheme, makes this breed an interesting model to analyse the fertility of the horses from a genetic point of view. The aim of this study was therefore to perform the first GWAS for reproductive efficiency traits in mares to identify genetic vari- ants, genomic regions, and candidate genes associated with fertil- ity,usingderegressedbreedingvaluesforthereproductive efficiency trait as pseudo-phenotypic data. We hope this study will contribute to a better understanding of mare fertility in horses. Material and methods Phenotypic recording of the mare fertility The fertility of 78 986 Pura Raza Español mares was determined using the reproductive efficiency trait defined as the percentual deviation between the optimal and actual parity numbers of the mare at each age. RE includes all the mare’s recorded foaling throughout her lifetime, i.e. up to her last known age. This trait, which was developed and validated as an indicator of fertility in PRE mares, has produced the highest heritability (0.23) among sev- eral traits recently analysed in this breed (Perdomo-Gonzalez et al., 2020). Reproductive efficiency was estimated individually by analys- ing all the information available in the PRE studbook (344 707 foal- ing records from 78 986 PRE breeding mares bred in 63 countries collected between 1970 and 2020). As a first step, we selected all the records of the mares belonging to studs whose main activity was the production of foals (more than 10 foals produced per stud per year). Secondly, we pruned out all the mares employed mainly for leisure or sports activities and kept only the individuals who had had their first foal between 4 and 7 years old, and who had intervals between first and second foaling and last and penultimate foaling of less than 5 years. The final dataset employed in RE esti- mation included 23 899 mares belonging to 8 133 studs. Estimation of genetic parameters for reproductive efficiency Reproductive efficiency breeding values (REEBV) were estimated using an Single Step REML Animal model including the stud, year of birth and size of the herd of birth of the mare as fixed effects, age and inbreeding of the mare as lineal covariates, and additive effect of the mare and residual effect as random effects. The extended pedigree of the mares (with all known genera- tions) included 87 227 animals and was obtained from ANCCE (AsociaciónNacionaldeCriadoresdeCaballosdePuraRaza Español) studbook. The number of maximum known generations of these mares was 16, with an average of 5.26 complete genera- tions and 8.81 equivalent generations. In addition, a pedigree- genomic-basedhybridrelationshipmatrix(matrixH)was constructed. In the final step, REEBV showing an accuracy higher than 0.5 were deregressed toREdEBV(pseudo-phenotypes) following the procedure described by Garrick et al., 2009, in order to avoid the double imputation of the genetic effect on the GWAS. These values were used as pseudo-phenotypes in GWAS for fertility. All the genetic estimations, including REEBV and REdEBV, were obtained N. Laseca, S. Demyda-Peyrás, M. Valera et al.Animal 16 (2022) 100476 2