Nowadays scientists pay much attention to molecular genetic tools for assessing the breeding and productive qualities of farm animals. The use of certain DNA segments as genetic markers became widespread in the 1980s [1]. Later on, the development of genotyping and sequencing technologies made genome-wide genotyping of single nucleotide polymorphisms (SNPs) more available resulting in an impetus for research genetic basis of complex traits in farm animals as well.

With adopting genomic selection most of the purebred elite lines of farm animals are genotyped using biochips with high and medium density [2]. As a rule, such methods as the genome-wide association study (GWAS) and/or the Selection Signature Analysis [3] are used to evaluate the effects of selection based on the analysis of genomic data. These methods are successfully applied to identify genomic regions in rigorously selected farm animals.

For pigs, it was found that GWAS is a powerful method for identifying genetic variants associated with various breeding traits [4]. However, we should note that identification of reproducible significant associations across the whole genome, GWAS requires significant sampling.

The advantage of studies based on “selection signatures” consists in that they are applicable to relatively small populations under study. One of the “gold standards” for assessing differentiation between the groups is the Fst statistic. Fst is an integral part of descriptive statistics for population estimation, being used in evolutionary biology and clinical genetics, and also to identify genomic loci associated with complex traits [5].

One of the most important factors influencing the economic efficiency of pig breeding is reproduction. The task of identifying loci associated with the reproductive traits of pigs has not lost its relevance over the past decade, but with the advent of new technologies and methods it increasingly attracts researchers. Since genome scanning with further Fst evaluation can identify genome regions subjected to selection we decided to apply the Fst method to identify potential loci associated with reproductive qualities of pigs. The work mainly focused on searching common options related to the fertility of sows and the weight of the litter at birth, since today the urgent task is to prevent a decrease in the weight of piglets at birth with high fertility of sows.

The traits characterizing the fertility of sows and the weight of piglets at birth were of certain interest. For this research we divided the indicators for all features into low, medium and high ones, taking into account the quantiles of 0% to 25%, 25% to 75%, and 75% to 100%, respectively. In searching selection signatures we used only extreme phenotypes of L and H groups. The average values for all sows in the studied sample, and as well as in the groups with low and high productivity are presented in Table 1.

Based on the results of the F_st analysis we identified 724 outliers exceeding 0.99%. These variants represent selection signals for the BALWT, BALWT1, NBA, and TNB traits. In general, outliers are presented on all chromosomes; this is especially clearly seen for BALWT1. For other traits we can identify individual chromosomes with the largest number of outliers. Thus for BALWT and TNB the largest number of outliers is located in SSC8, for NBA – in Sus scrofa chromosomes 3 (SSC3) (Figure 1). The outliers are represented by various variants of nucleotide substitutions, to a greater extent these are intron and intergenic variants, but there are also upstream gene, non coding transcript exon, downstream gene and 3 prime UTR variants (Supplementary file).

For BALWT1 and BALWT, 197 and 201 variants, were identified respectively, of which 22 SNPs are common for both traits. For TNB and NBA (219 and 198) variants were found respectively, while 46 SNPs were common for these traits. Between the fertility traits (NBA and TNB) and piglet weights (BALWT1 and BALWT), common variants were also established (Table 2).

Two variants turned out to be common between BALWT1 and NBA: rs81379527 (SSC3: 27441840), an intron variant of the xylosyltransferase 1 (XYLT1), and rs80934876 (SSC4: 354634), an intron variant of the tonsoku like, DNA repair protein (TONSL). XYLT1 is expressed in chondrocytes during embryonic development and encodes xylosyltransferase 1, with its functions associated with the synthesis of proteoglycans. In the research of Bergfelder-Drüing et al [9] an association between the XYLT1 gene (rs81379421) and number piglets born alive was revealed. In addition, XYLT1 is a potential candidate gene for the short stature and dwarfism syndrome. TONSL variants have deleterious effects at multiple stages of embryonic and postnatal development.

There are four variants located in SSC8, turned out to be common between BALWT1 and TNB, of which rs321611489 and rs81404839 are intron variants of the potassium voltage-gated channel interacting protein 4 (KCNIP4) gene. The effect of KCNIP4 gene variants on porcine NBA was reported by He et al [10].

There are three common variants have been identified between BALWT and NBA, intron variants rs327523214 (SSC3: 36974365) of the RNA binding FOX-1 gene; intergenic variants rs80838609 (SSC14: 139576563); synonymous variant of the thromboxane A synthase 1 (TBXAS1) rs81248107 gene (SSC18: 9717477). The TBXAS1 enzyme is involved in several pathophysiological processes, including hemostasis, cardiovascular disease, and apoplexy. In addition, it is assumed that the enzyme is involved in the regulation of uterine and intrauterine blood flow.

Two variants common for BALWT and TNB have been identified: intron variant rs81256424 (SSC8: 65110831) of the ENSSSCG00000063524 gene and intron variant rs81330142 (SSC11: 18043223) of the emopamil binding protein-like (EBPL) gene. The relationship of the EBPL gene with the body weight of young bulls was highlighted in the work of Lindholm-Perry [11].

Besides 5 SNPs were noted for more than two traits (Table 2). Thus, intergenic variant rs334075913 showed a selection signal for BALWT1, NBA, and TNB. this variant is localized in SSC2: 138901377. QTL#106221 associated with piglet mortality was previously defined in this area. In close proximity to this variant, the secreted protein acidic and cysteine rich (SPARC) gene (SPOCK1 (osteonectin), SSC2: 139018177 – 139523601) is localized, which was identified as a new candidate gene for the Menarche age and is associated with the onset of female reproductive life, cattle, humans, and sheep fecundity [12] and ovulation rate in pigs [13].

Intergenic variant rs81476874 showed a selection signal for BALWT1, BALWT, and TNB traits. This variant is localized in SSC8:12351405. This area intersects with QTL #24282 associated with litter size and QTL #645 affecting the plasma follicle-stimulating hormone (FSH) and ovulation rate in females [14]. Intergenic variant rs328047631 (SSC8: 13883790) for BALWT1, BALWT, and NBA and intergenic variant rs81401114 (SSC8: 11570387) for BALWT1, NBA, and TNB were also found in the eighth chromosome. This option intersects with a number of QTLs, among which it is interesting to note QTL#24282 for Litter size; QTL#325 [15] and QTL:656 for Average daily gain; QTL#653, QTL#654, QTL#655, QTL #21254 and QTL#21253 [16] for Body weight. Besides, the transmembrane anterior posterior transformation 1 (TAPT1), (SSC8: 11365423-11415699) and lim domain binding 2 (LDB2), (SSC8: 11641060..12037388) genes are localized in close proximity to the rs81401114 variant (SSC8: 11570387). TAPT1 encodes the evolutionary conservative transmembrane protein anterior posterior of transformation 1. The studies [17] show the involvement of TAPT1 in the basis of a complex congenital syndrome clinically manifested by lethal skeletal dysplasias and ciliopathy. This syndrome is characterized by fetal death and multiple congenital malformations.

The LDB2 gene plays a regulatory role in retinal development and the cell cycle, but its biological role remains unclear. The association of the LDB2 gene is characterized by the body weight of chickens (of commercial broiler chickens) and birth weight in Cashmere goats [18].

Synonymous variant rs81211492 showed a selection signal for BALWT, NBA, and TNB. This variant is localized in the aminopeptidase puromycin sensitive (NPEPPS), (SSC12: 23804839..23892055) gene. This gene encodes the puromycin-sensitive aminopeptidase playing the neuroprotective role [19]. Besides, the protein is involved in the regulation of the cell cycle in mammals, and it is required for meiosis exit and anteroposterior polarity in single-celled Caenorhabditis elegans embryos and mice deficient in puromycin-sensitive aminopeptidase are smaller and have reproductive problems [20]. According to the QTLdb database, rs81211492 overlaps with 42 QTLs associated with various functions, but 5 QTLs (QTL:5261, QTL:5227 [21], QTL:4254 [22], QTL:6479, QTL:6472 [23]) are responsible for tea number.

Moreover, the intron variant rs80883327 (SSC17: 1204066) of the deleted in liver cancer 1 (DLC1) gene showed a selection signal for all four traits. DLC1 is considered a tumor suppressor gene in various human cancers. The functional features of DLC1 in pigs are not studied enough. According to an open-access pig expression map [24], the DLC1 expression is enhanced in retina, fallopian tube, testis and upper respiratory system. In other species, evidence of the role of DLC1 in embryogenesis has been obtained. The DLC1 gene plays a critical role in the regulation of cellular functions during the early development of mice. According to Durkin et al [25] homozygous mica with DLC1 knockout died around 10.5 days of embryonic development. The role of DLC1 in spermatogenesis and male fertility in mice is presented in Okitsu [26]. The DLC1 protein is also thought to play an important role in the development of the placenta [27].

Reproductive traits play a leading role in the economics of pig production. In this regard, the interest in the biology of these traits has not waned for decades. The main indicators of the reproductive performance of sows are the total number of piglets at birth (TNB) and the number of live piglets at birth (NBA), since these signs are measured ones and there are no particular difficulties in taking them into account. However, these traits have a very low inheritance rate and an extremely complex biology. In fact, these signs combine all the processes associated with the reproductive cycle of sows (ovulation, fertilization, implantation, prenatal survival, uterine capacity, etc.).

In addition, along with the fertility of sows the matter of piglet weight at birth is arising. On the one hand, the increase in the number of piglets at birth is believed to be directly related to the decrease in the weight of piglets. On the other hand, there appears more evidence that the decrease in the weight of piglets at birth does not depend directly on their number, but more related to the capacity of the uterus and the body of the uterus, the ability to provide energy costs for the full development of offsprings in the embryonic period.

Since these traits have low heritability, their genetic architecture is rather interesting. The growth of GWAS research on the genetic architecture of reproduction was observed at the end of 2017 after the publication of an updated version of the pig genome [10]. A large amount of data obtained today on the basis of GWAS showed a low reproducibility of the results, which can be explained by the individual characteristics of the genetic structure of populations (allele frequencies, linkage disequilibrium), but also by the need for a large sample of animals, this is especially important for traits with low heritability.

In our work to assess the genetic architecture of sow reproduction we decided on four main traits TNB, NBA, BALWT1, and BALWT, and chose the selection signature search approach based on the F_st method as a method for identifying genetic variants. As a result, we found 724 variants for all features. Selection signals for two traits was shown in 79 variants. To a greater extent the variants overlapped between TNB and NBA, which is quite expected since the biology of these traits has so much in common. Generic variants between BALWT1 and BALWT require more effective analysis in the future, as they may provide some clarity on the relationship between litter weight and average birth weight per piglet. Besides, a number of variants responsible for both the mass of the nest and the number of piglets were identified. These variants are of interest as potential markers for assessing correlations between the number of piglets and their weight. Most of them are localized in genes (but nearby genes are also considered in the case of intergenic variants). The functional characteristics of genes and their associative relationships with productivity traits presented in the literature indicate their connection, to a certain degree with the reproductive process, either in pigs, or the data are presented for other species. The functions of genes are associated with the embryonic process, the survival of embryos, as well as various pathologies, further associated with a growth decrease. All selection variants directly related to the number and weight of piglets intersect with different QTLs, including for average daily gain and body weight. Some of these genes were previously directly noted in associative studies with sow fertility and are reviewed by Bakoev et al [28].

Thus, our work resulted in identification of variants associated with the reproductive characteristics of pigs. Moreover, we identified variants which are potential markers for both the number of piglets at birth and the weight of piglets at birth, which is extremely important for breeding work to improve reproductive performance in sows.