Allelic to genome wide perspectives of swine genetic variation to litter size and its component traits

Litter size is a complex and sex limited trait that depends on various biological, managemental and environmental factors. Owing to its low heritability it is inefficaciously selected by traditional methods. However, due to higher heritability of ovulation rate and embryo survival, selection based on component traits of litter size is advocated. QTL analysis and candidate gene approach are among the various supplementary/alternate strategies for selection of litter size. QTL analysis is aimed at identifying genomic regions affecting trait of interest significantly. Candidate gene approach necessitates identification of genes potentially affecting the trait. There are various genes that significantly affect litter size and its component traits viz. ESR, LEP, BF, IGFBP, RBP4, PRLR, CTNNAL1, WNT10B, TCF12, DAZ, and RNF4. These genes affect litter size in a complex interacting manner. Lately, genome wide association study (GWAS) have been utilized to unveil the genetic and biological background of litter traits, and elucidate the genes governing litter size. Favorable SNPs in these genes have been identified and offers a scope for inclusion in selection programs thereby increasing breeding efficiency and profit in pigs. The review provides a comprehensive coverage of investigations carried out globally to unravel the genetic variation in litter size and its component traits in pigs, both at allelic and genome wide level. It offers a current perspective on different strategies including the profiling of candidate genes, QTLs, and genome wide association studies as an aid to efficient selection for litter size and its component traits.


Introduction
Livestock sector has been an integral part of human civilization, fulfilling the cultural, economic and nutritional needs of man. The demand for livestock products are ever expanding being driven by population growth, urbanization and increasing incomes. The intensively raised livestock in the developed countries are the main source of the world's poultry and pig meat production, and such systems are being established in developing countries, particularly in Asia, to meet increasing demand. In particular factory farm based swine operations have huge scope for fulfilling the global meat demand. Profit gained from the swine industry is immediately reflected via improvement of reproductive traits enhancing efficiency of production. However, reproductive traits are complex traits exhibiting low heritability and strong heterosis thereby limiting the efficacy of selection [1]. Among various reproductive traits, litter size is of prime interest to pig producers and breeders and is the most important economic trait. Realizing its extreme importance, selection of replacement gilts having potential for larger litters has proven beneficial to the swine industry [2].
Litter size-a complex and sex limited metric trait is measured as the number of alive piglets born or total piglets born. Prior knowledge regarding determinants of litter size and their interaction is required to facilitate litter size improvement. Litter size depends on various factors such as rate of ovulation, rate of fetal survival, uterine capacity, management and environmental conditions as well as genetics [3]. However primary factors influencing litter size, also known to be the main components of the trait include rate of ovulation, fertilization, survival of embryo and fetus [4]. The average degree of embryonic losses has been documented as 20-30% whereas fetal losses as 10-20% [5]. Though the chance of occurrence of these losses is at any stage, the principal factor limiting porcine litter size is pre-implantation embryonic losses. Owing to its low heritability and being influenced by various biological factors that are difficult to intensely estimate, litter size is inefficaciously selected by conventional schemes and hence genetic improvement by marker assisted selection and genomic selection are proposed as a complimentary tools in order to amplify the economic performance of piggery. Furthermore, advancement in pig industry through the use of improved feeding and housing conditions has reached a limit therefore genetic improvement of economic traits via selection is of prime importance [6].
Marker assisted selection that allows division of litter size into its component traits can be utilized for greater progress as it optimizes different physiological mechanisms affecting litter size [7]. This has been further proven by selection experiments which produced insignificant response for increased litter size. A study on selection for ten generations found absence of response [8]. The reasons for failure of these experiments are multi causal such as maternal effect, inbreeding, management issues, small population size and within family selection. However, owing to comparatively higher heritability of ovulation rate and embryo survival, selection based on components of litter size is encouraged. Also index selection for viable embryo and litter size is more efficacious in improvement of litter size than single trait selection. Alternatively, litter size may be measured as minimum number of embryos allowed by uterine space or minimum viable embryos. Basis of this method was laid by measurement of uterine space via surgical procedures.
Genetic variability which exists for various reproductive traits is prerequisite for selection. However, choice of selection program should also involve keen observation of the predicted outcome. For instance, if selection is undertaken for increased ovulation rate then it would detrimentally affect early embryonic survivability. Likewise, selection for improved litter size would affect the birth weight. Additionally, uterine overcrowding 44 days post fertilization affects development of placenta along with fetal and post natal development [9]. Presence of genetic variability within breed is indicative of possible genetic improvement of a particular trait. Rate of improvement can be increased if genes that are responsible for physiology and biochemistry of the trait are manipulated directly. Since most of the genetic control is evident at the level of gene therefore gross chromosomal abnormalities severely affect reproduction. However, incidence of such abnormalities is relatively rare. In this review, studies in the domain of swine genetic variation to litter size and its component traits at allelic and genome wide level have been thoroughly reviewed. Further comparison of strategies for genetic improvement has been surveyed along with understanding allelic to genome wide perspectives. However, for wider application various strategies need to be merged across varied breeds and lines. In addition, advanced tools should be incorporated for profound understanding of genetics of reproductive traits in swine.

Association of litter size with sow and piglet characteristics
Litter size is complex trait and positively associated with ovulation rate, uterine capacity and postnatal survival. It is also associated with stillbirths where probability of stillbirth increases with abnormally large or small litters [10]. Litter size being trait with low heritability requires ample time for genetic improvement. However manipulation of its component traits can promote significant improvement. Index selection for ovulation rate and embryo survival could be one such approach [11]. Ovulation rate has higher heritability and hence selection for ovulation rate can increase the no. of ovulating follicles thereby increasing the litter size [12]. However it often increases the risk of piglet mortality and therefore direct selection for postnatal survival is considered the most effective strategy [13]. The most crucial factors involved in survival of piglet from birth to weaning are individual and relative birth weight which are negatively correlated [14]. Further litter size is negatively correlated with birth weight [15] which affects number of piglets born alive. Litter size is also significantly affected by stillbirths. Studies have indicated that around 10% of piglets born are stillbirth at parturition [16]. Also significant association occurs between farrowing duration and number of stillbirth. Farrowing duration for over three hours leads to asphyxia due to rupture of umbilical cord or placental detachment. Risk of hypoxia and farrowing time also increases with large litters whereas in small litters, the oversized piglets are subjected to greater difficulties in gestation or farrowing and their birth is blocked leading to hypoxia and death. It has been reported that male piglets have higher incidence of stillbirth and lower birth weight [17]. In order to decrease stillbirth, increase in individual birth weight was recommended. Additionally, selecting individuals with higher birth weight did not increased pre weaning survival rate indicating the need for balanced selection. Therefore keen evaluation of component traits is ideal for increasing litter size.

Strategies for improvement of porcine reproductive traits
Improvement of reproductive traits has gained much importance in view of the fact that moderate increase results in large profits especially in swine [12]. This improvement is made through selection which is routinely based on performance of the animal but the procedure is tedious and expensive. These lacunae can be filled with the cohesive use of marker assisted selection, resulting in a faster rate of change and hence improvement [18]. Association of genes with trait of interest has been unveiled with advent of DNA technologies thereby necessitating designing of selection methodologies based on information obtained from DNA. Genes influencing these traits are determined by employing following two different strategies: 1. Linkage analysis wherein genomic regions accommodating the concerned genes are scanned to detect quantitative trait loci.

3
2. Candidate gene approach to identify polymorphism in genes causing phenotypic variation through their physiological and biochemical role or their location in genome, linked with the trait or with differentially expressed genes for the trait. It is pursued as direct evaluation.

QTL analysis
Mapping of genomic regions in linkage analysis is based on co-segregation of marker alleles with phenotypic traits. These genomic regions which are linked to polygenic phenotypic traits are known as quantitative trait loci (QTL). Microsatellites owing to their high polymorphism are usually employed in QTL analysis as an indirect strategy. QTL analysis is costly and time taking, since it requires at least three generations. Also they are comparatively difficult to use in MAS as they often span more than 10-20 cM hence fine mapping of QTL is necessary [6]. However, increase in number of alleles and uniform distribution in the population, enhances information content of markers. QTL analysis is pursued with the aim of identifying genomic region significantly affecting a desirable trait. Crosses between Landrace, White composite, Large White Yorkshire and Chinese Meishan have been frequently utilized as resource population in order to undertake genomic scan for litter size and its component traits. For instance several studies indicated presence of QTLs for ovulation rate on SSC8 [19][20][21]. In another study on Meishan X Large White cross population, QTL for prenatal survival, litter size, teat number on SSC8 was evident [22]. However the number of individual animals measured in the analysis was low than that of mice primarily because of longer generation turnover in pigs or in few candidate gene analyses because genotyping for numerous markers has to be performed unlike in candidate gene approach wherein few genes are studied. Population size therefore limits the mapping resolution of these studies which can be improved by conducting joint analysis of several experimental designs [23]. In Yorkshire sows, association analysis revealed a QTL harboring genes encoding various components of TGF (transforming growth factor) beta signaling pathway that is likely to participate in increased fecundity [24]. In another study on Erhualian pigs which is acclaimed for its high litter size, one SNP was significantly and 11 SNPs were suggestively associated with TNB and NBA. These SNPs were identified on chromosomes which harbored QTLs for litter size or uterine length [25]. Genome wide association analysis using linear and Bayesian model has also revealed presence of QTL for litter size and component traits. In a study on Danish Landrace and Yorkshire pigs, GWAS using Bayesian approach revealed QTL for teat number, TNB, corpus luteum number, age at puberty, litter weight, embryo weight on SSC 1, 2, 3, 6, 7, 9, 13 and 14 [26]. These results were in accordance to other studies [27][28][29][30].
Owing to the fact that several genes contribute for only limited magnitude of phenotypic variance in fecundity parameters and the QTL regions are wide enough to pose difficulty in defining the gene effect, no gene with a causative mutation has been detected in pigs through linkage analysis. Moreover, standardization of environmental influences for longer duration is difficult. Therefore fine mapping at early stage of QTLs should be advocated for economic traits and must be extended in future.

Candidate gene approach
Characterization and genotyping of candidate gene is done by using intragenic DNA variants. These variants differ in single base pair and referred as SNPs which are usually biallelic resulting in three distinct genotypes. The effect of candidate gene is evident by significant linkage disequilibrium studies in large population. To validate the significant role of candidate gene, results are tested in various populations of a breed or even on several breeds. Sometimes, there may be inconsistency of results which does not essentially indicate that gene is not potential candidate for the respective trait. Rather discrepancy may arise due to different allelic and genotypic frequency, epistatic effects or different linkage phases between marker and causal mutation (Table 1). Prior knowledge of gene having probable role in target trait is necessary in candidate gene analysis [34]. This is an attractive approach for animal breeders as positive results can be readily applied. Inconsistency of genetic effects across varied population and breeds poses limitation to its use. Before application of the approach, various key points have to be taken into account such as polymorphisms in candidate genes are intragenic markers rather than causative mutation therefore the recombination and mutation events may not strictly be as with causal mutations. Furthermore, low allelic frequency of polymorphism results in insignificant result. With these approaches, successful identification of genes influencing economic traits has been carried out with ESR gene being earliest proof of association between gene and litter size [35]. Mapped porcine genes are proportionately small, limiting the total positional candidate genes. Principally every animal can be studied from any population and hence feasible genotyping of candidate genes for particular trait by various research groups is possible. Testing of gene variants in various populations is pre-requisite in candidate gene strategy to detect its general effects. However, variation in resource population and number of individuals in different studies has been witnessed, some studies using reference families while some using commercial pig population. Owing to differences in housing, feeding and other management conditions, results are difficult to compare ( Table 2).
Candidate gene analysis in pigs has mostly focused on litter size or its component traits. First candidate gene that was significantly associated with litter size was ESR1 (estrogen receptor 1) gene [35]. Another candidate gene for litter size is PRL (prolactin) on SSC16 [36]. EPOR gene (erythropoietin receptor) regulates the development and differentiation of fetal erythrocytes. The gene was significantly associated with litter size and uterine capacity in Yorkshire, Landrace and Duroc pigs [37] indicating candidate gene effect. In a study conducted on Chinese Erhualian pigs various SNPs were significantly associated with reproductive process. These SNPs were corresponding to genes such as UCHL1 (Ubiquitin carboxy-terminal hydrolase L1) that regulates meiotic maturation of pig oocyte and CASP6 (Caspase 6) that activates mitosis and hence regulates ovulation rate. The results indicating these genes as vital candidate genes for litter traits [38]. In another study on Erhualian pigs, a SNP (rs81447100) was significantly associated with litter size. Interestingly, RBP1 (Retinol binding protein1) and RBP2 that regulates uterine development was in close proximity of the SNP [25]. In study on DIV pig population, a SNP in miR-27 gene was associated with TNB and ovulation rate [39]. Associative analysis on Large White pigs revealed significant effect of SNP rs80956812, rs81471381, rs81421148 and rs81399474 on litter traits and AKT3 (AKT serine/threonine kinase 3) gene was identified as potential candidate gene [40]. Various other genes have been studied in different population for association analysis with litter size and component traits which have been reviewed in Sect. "Candidate genes for litter size in pigs" and summarized in Table 3.
Candidate gene approach can be directly employed when gene has appreciable effect on physiological trait and is known in advance (physiological candidate), or the gene is distinctly located in QTL region (positional candidate) or influences the trait in other species (comparative candidate gene). For instance QTL region on SSC3 in Yorkshire pigs harbored Karyopherin alpha 7 gene which is known to be expressed in ovarian tissues and embryos for cleavage development in porcine embryos [41]. However in a study, region on SSC13 in Landrace included HES1 (hairy and  enhancer of split-1) gene which determines differentiation and fate of the cells [26]. Economic traits are influenced by genes via two possible pathways [6]: 1. Type of encoded protein may be changed due to mutation in the coding region of particular gene. 2. Amount of gene transcripts may alter in the cell, changing the quantity of functional protein due to mutation in regulatory region of particular gene.
These two gene effects cannot be differentiated with linkage and association studies because both pathways lead to appreciable phenotypic changes. Nevertheless, major phenotypic changes are often caused by change in structure of primary protein which results in absence of functional protein whereas minor changes are expected due to mutation in regulatory regions and neutral substitution of amino acids.

Estrogen receptors (ESR)
Steroid hormones such as estrogen play a central role in the postnatal female physiology and their effects are exerted through its receptors. The function of the estrogen hormone is mediated by binding with its receptors namely ESR1 and ESR2, which have cooperative action as heterodimers. Knockout mouse model technique has been utilized to infer the functional similarity of both receptors [42]. The estrogen receptor gene (ESR) located on the pig chromosome 1 [43] is one of the most extensively investigated candidate genes for litter size traits in pigs. Embryotrophic role of the gene is considered crucial for growth and maturation of ovarian follicles and embryo as well as peri-implantation development [44]. Loci coding for ESR gene affects the litter size of sows and their specific genotypes show additive effect with the same hence it is considered as candidate gene for prolificacy traits.
Different studies have been conducted in purebred, crossbred and hybrid sows concerning litter size traits, especially total number born (TNB) and number born alive (NBA). In a study conducted on Meishan breed, a PvuII polymorphism of the gene was significantly associated with litter size [45]. Contrarily, in Chinese-European pigs non-significant association of PvuII polymorphism with litter size but significant association with TNB was observed [46]. Also, in Italian Large White population, genetic variation of the gene was not observed and hence was not associated with litter size in the target breed [47]. In a study on Yorkshire, Large White and crossbred animals, ESR genotype was significantly associated with litter weight of piglets born, litter weight of piglets born alive, fetal weight, ovulation rate [48]. In another study, a polymorphism of ESR2 at telomere of q arm of SCC1 was identified and no significant association with litter size of Iberian pig population was found [43]. In a study on population of autochthonous pigs, no significant difference was observed for ESR2 loci. Nevertheless, BB genotype showing an increment of 2 piglets per litter was noticeably superior (p < 0.001) to AA genotype [49]. Despite the variable results obtained in different findings due to difference in sample size and target population, the gene was found to be associated with litter size traits and therefore is a potential candidate gene for litter size in pigs.

Leptin (LEP)
Leptin is a 16 KDa polypeptide hormone encoded by LEP gene. It is secreted by white adipocytes and implicated in the regulation of reproductive functions in conjunction with feed consumption and energy homeostasis [50]. During puberty, the concentration of the circulating leptin hormone increases reaching a threshold that activates reproductive axis. Hence, it acts as a metabolic gateway for puberty [51]. The porcine leptin gene has been localized on chromosome 18 [52] and consists of three exons and seven single base polymorphism [53]. It interacts with its receptor LEPR mapped on chromosome 6 to mediate its effect at hypothalamic level [54]. High degree sequence homology of porcine LEPR genes with that of human and mice have been utilized in primer designing [55].
Gene polymorphism has been studied in different population to study its effect on litter size and its component traits. In a study conducted on cross of Polish Large White and Landrace sows, two distinct alleles of the LEP gene were recognized. However, non-significant differences for litter size traits were observed between animals of different genotypes [56]. Weak association between leptin SNP and mummified fetus in four line composite pig population was also reported [57]. In Large White population, LEP polymorphism was significantly associated with total litter size at birth and total litter size born alive [58]. Similar significant effect was also observed in Luchuan pigs [59]. Polymorphism in LEPR gene was also studied in Landrace, Yorkshire and Duroc breeds. However, association was evident between polymorphism in intron 2, exon 2 and exon 18 with litter size in two breeds namely Yorkshire and Duroc [55]. Another study in Suzhong sows revealed the beneficial effect of T allele of LEPR locus on litter size and litter weight as genotype TT had greater TNB and NBA as compared to sows with genotype CC [60]. In F2 population of divergent crosses, polymorphism in the gene was associated with total teat number [61]. These studies indicated the potential use of leptin gene and it's receptor in marker assisted selection.

Properdin (BF)
Properdin-a single chain glycoprotein of 93 kDa is encoded by complement factor B (BF) gene. Porcine BF gene has been mapped within dense gene cluster (MHC) harboring genes regulating reproductive physiology and is located on centromeric region of chromosome 7 [62]. The gene plays vital role in growth of uterine epithelium which is crucial in establishment and maintenance of pregnancy [63]. The first RFLP of the porcine BF gene was demonstrated using SmaI restriction enzyme [64].
Various investigations have been undertaken which showed significant association of the gene with components of litter size i.e. total number born and number born alive. However, noticeable association was evident in multiparous sows [65]. In crossbred pig population (Large White X Landrace X Leicoma), allele B of the gene was found favorable as sows with BB genotype had higher TNB and NBA as compared to other genotypes [66]. Similar results were observed in commercial pigs in Greece where TNB and NBA showed significant statistical difference (p < 0.05) from 2nd to 5th birth (high productive period) and BB sows had higher TNB and NBA per parity than AB sows [67]. However, when all births were examined, significance level was insufficient even though BB sows had higher litter size. Furthermore, the concept of removal of sows from the breeding herd once they reach 5th parity or attain 30 months age was supported by the finding that litter size decreased 5th parity onwards. In another population of autochthonous Greek breed higher beneficial effect of the BB genotype was observed for TNB and NBA traits compared to the other two genotypes [49]. Owing to its significant effect, the gene may be considered as a candidate gene for litter traits in pigs.

Insulin-like growth factor binding protein (IGFBP)
Fertility is regulated by interaction of various reproductive organs, among which uterus plays primary role in implantation and fetal growth by expressing numerous proteins. One such protein is IGFBP which participates in reproductive physiology by regulating ovulation, implantation, fetal development and pregnancy maintenance [68]. IGFBP is a family of six binding proteins that are potent modulators of IGF1 and IGF2. They are produced in ovarian follicles as signaling molecules and regulate growth and differentiation of endometrial cells [69]. During follicular growth, IGFs play key role in sensitizing the granulosa cells to FSH in addition to regulating proliferation, survival and migration of trophoectoderm during early gestation [70]. The mRNA level of IGFBP2 is linked with implantation and varies during different stages of endometrium secretion. During development of placenta, mRNA expression of the gene is significantly increased and the action of IGF is modulated thereby controlling embryonic development. Gene up-regulation was witnessed in endometrium on 14th day of gestation in pregnant mares and sows indicating its vital role in endometrium. Another most abundant member of IGFBP family i.e. IGFBP3 regulates the biological activity of IGF-1 in mammalian reproduction.
Considering crucial role of the genes of IGFBP family in reproductive physiology, various studies were undertaken. In a study on Berkshire pigs, SNP was identified in intronic regions of IGFBP2 and IGFBP3. The SNP in IGFBP2 was found to affect its own mRNA expression. Further it was witnessed that animals with AT genotype had largest TNB and NBA while TT genotype had highest transcript level, additionally suggesting nil correlation between the two traits. For IGFBP3 gene, GG genotype exhibited highest litter size traits [71]. Similar study on Finnish Landrace sows suggested that animals with minor allele for IGFBP1 and IGFBP2 gene had positive effects on litter size traits [72]. In a composite population of Landrace-Duroc-Yorkshire, the IGFBP3 gene significantly affected estrus interval and stillborn births [57] thereby regulating the litter size. These findings proposed that SNPs in IGFBP gene family could be utilized as biomarkers for litter size in pigs.

Retinol binding protein 4 (RBP4)
RBP4 is a major component of uterine histotroph and secretory product of pig conceptus that plays key role in transportation of vitamin A to the developing embryo [73]. Vitamin A in turn have numerous effects with respect to establishment and maintenance of pregnancy as well as embryonic development. Porcine RBP4 gene has been mapped on chromosome 14 [74] and is suggested to be a strong candidate gene for litter size in pigs as supplementing pregnant sows with vitamin A significantly increases litter size [75].
Various studies have been undertaken to investigate the association of gene with litter size owing to its increased production during the critical stage of blastocyst elongation [76] as well as its role in reproductive physiology. In an investigation pertaining to Polish sows, the gene showed polymorphism and was found to be significantly associated with component traits of litter size such as TNB, NBA and number of piglets weaned. Sows with BB genotype showed large litter size as compared to AA and AB genotype. The difference was statistically significant in first parity [77]. However in Tibet pigs, sows with AA genotype had significantly larger TNB and NBA as compared to BB and AB genotype [78]. Similar results were obtained in commercial crossbred from Large White X Landrace sows [79] and in hyper prolific Landrace X Large White sows where animals with AA genotypes had higher TNB [80]. In Autochthonous pig population, sows with genotype AB showed largest TNB and NBA as compared to other genotypes [49]. As the gene was significantly associated with litter size in all these findings, therefore it can be used for selecting sows with large litter size.

Prolactin receptor (PRL)
PRLR was first cloned from rat liver comprising 310 amino acids and later from the ovary of same species comprising 610 amino acids. Multiple isoforms of the gene in the findings are indicative of alternate splicing [81]. Porcine PRLR has been located on SSC16 and encodes receptor for prolactin hormone [82]. The hormone plays vital role in reproduction and lactation as attested by large body of literature.
Various studies have been undertaken to study the association of gene with litter size traits in pigs. The gene was significantly associated with reproductive attributes in various breeds such as Large White, Duroc, Landrace, Chinese Meishan and Pietrain [83]; Landrace, Meishan and Large White synthetics [82]; and crossbred gilts of Meishan and Large White [84]. A study on crossbred pigs (Large White and Landrace) demonstrated that in first parity, AA genotype sows had significantly largest and BB genotype smallest litter size. However in later parities, though the results were same but the difference between genotype was not significant. This may be due to large residual variance in later parities [56]. Nevertheless, difference can be proven to be significant by increasing the sample size. Similar studies were conducted in Large White [85], crossbred (Large White and Meishan) population [84] and Beijing Black pig population [86]. In these findings, frequency for favorable allele A was similar. However, in a study lower frequency in Landrace and a higher frequency in Duroc was observed [36]. Study on synthetic Large White line showed that sows with AA genotype had 0.66 more NBA than others while in Landrace synthetics AA sows had one more TNB and NBA than BB sows. However, in Meishan synthetics, these parameters were highest for sows belonging to AB genotype [82]. These studies suggested evident effect of PRL gene on improved litter size.

CTNNAL1, WNT10B AND TCF12
WNT signaling is complex interactive pathway comprising of proteins, receptors and other regulatory elements. The pathway has clear and crucial role in embryonic development, implantation and various other reproductive processes such as formation of ovarian follicles, ovulation, maintenance of normal pregnancy and lactation [87,88]. Catenin alpha-like protein 1 (abbreviated as CTNNAL1), wingless-type MMTV integration site family member 10B (abbreviated as WNT10B) and transcription factor 12 (abbreviated as TCF12) are among the various pathway genes. Down regulation of CTNNAL1 gene in placental tissues of women suffering from pre-eclampsia, hemolysis, elevated liver enzymes and low platelets or HELLP syndrome indicated the role of gene in maintaining normal pregnancy [89]. Up regulation of WNT10B was evident in endometrium of horse in late diestrus period [90], indicating its effect on early embryonic development along with maintenance of pregnancy. The gene was also expressed in murine blastocyst [91]. TCF12 played vital role in oogenesis and sex determination in drosophila [92]. Differential expression of these pathway genes was witnessed in Chinese Taihu and Large White sows [93].
Owing to the observed expression pattern and close proximity of these genes with QTLs for TNB and NBA, they are regarded candidates for litter size. Association analysis of TCF12 gene was conducted on Large White pigs. Animals with GG genotype had more NBA in first parity than other genotypes. In Large White pigs, for CTNNAL1 at locus c.1878 G > C, CC homozygotes had 1.14 more alive piglets per litter than GG. However, in Chinese DIV difference between CC and GG sows were 2.07 and 2.62 pigs per litter for NBA and TNB respectively. For WNT10B gene, TC gilts had lower NBA than other genotypes of Large White [94]. These findings suggest that these genes are potential markers for pig selection and breeding.

Deleted in azoospermia like (DAZL)
DAZL is one of the members of DAZ gene family that comprises of three genes namely DAZ, DAZL and BOULE. The genes of the family encode proteins that show germ cell specific expression pattern regulating development and differentiation of germ cells [95]. This has been validated via disruption of DZAL gene in adult mice that resulted in complete loss of germ cells and follicles in ovary [96]. Additionally the gene has been found to be expressed throughout human peri-implantation period suggesting its crucial role in implantation as well as embryo survival [95]. Porcine DAZL gene has been mapped on SSC13 and is located near QTL for ovulation rate [20] and stillborn birth [97].
Owing to its role in germ cell development, embryo survival and other reproductive processes, role of the gene as potential marker for litter traits was realized and association studies were conducted. In one such study on Italian Large White pigs, the gene was non-segregating [98]. However in another finding, A/G mutation in intron 7 and a C/A mutation in intron 9 of the gene was observed. This study reported that in DIV pig line during first parity, the BB pigs varied from AB with dominance effect for NBA as 0.06 pigs per litter. However, for SNP C/A (intron 9) in Large White CC genotype had 0.716 more pigs per litter than CD for NBA in all parities. Additionally, in DIV pig line animals with CC genotype had 1.940 and 2.017 more NBA than DD and CD respectively during first parity whereas CC differed from genotype CD in DIV line in all parities [99]. Linkage disequilibrium of the C/A mutation with QTL affecting reproductive processes indicated the role of gene as potential marker for improved litter size.

Ring finger protein 4 (RNF4)
RNF4 gene encodes for ring finger protein that was originally termed as SNURF (small nuclear ring finger protein) in rats. It co-regulates the transcription via steroid receptor dependent and independent promoters [100]. Overexpression of this protein enhances the transcription of various other steroid receptors such as estrogen and progesterone [101]. Additionally, the protein coordinates activities of numerous transcriptional signals and acts as a bridging factor. Spatial and temporal expression analysis of SNURF mRNA revealed that the gene was expressed significantly in oocytes in stage-dependent manner and participated in folliculogenesis. Additionally it regulated the development of fetal germ cells along with maturation of oocytes and granulosa cells [102]. The RNF4 gene is located on SSC8 and QTL influencing ovulation rate has also been identified within p arm of SSC8 [23].
Owing to its role in reproductive physiology, association analysis of the gene with litter size traits was conducted in pigs. One such study was undertaken in three distinct population and it was observed that TT sows in Minpigs had more TNB as compared to CC sows. Similar results were observed in Qingping and Line DIV population, however the results were not significant [103]. In another study the gene showed non-significant association with reproductive traits in Qingping pigs and Line DIV sows during first parity. However, in subsequent parities in Qingping sows, CC genotype had more TNB and NBA than sows with TT genotype [104]. These studies indicate significant yet inconsistent relationship between genotypes and litter size traits indicating that upon further validation the polymorphism in the gene can be considered a potential marker for litter size traits in pigs.

Genome wide association study of litter size
Genetic architecture of economic traits has remained a mystifying matter for researchers since many decades. However with advent of GWAS, new hopes have been pinned. GWAS is a technical approach for mining functional genes, SNPs, QTLs and other relevant genetic information regarding traits that are controlled by many genes and gene interactions [105]. It enhances our understanding of biological relevance and genetic background of economic traits thereby facilitating efficient MAS or genomic selection for genetic improvement. Currently, genome wide studies utilize SNP Bead Chip technology rather than NGS owing to its affordability. SNP Bead chips face several disadvantages such as rigid structure and uneven marker density across genome [106]. However they are being widely used in livestock species particularly in pigs for exploration of the genetic architecture of polygenic traits [107]. In piggery sector, GWAS has been invariably applied to unveil the genetics of various economic traits such as carcass quality [108], genetic disorders [109], coat color [110] etc. Several studies based on GWAS has also been conducted for reproductive traits especially litter size that plays crucial role for economic success. However due to low heritability, polygenic inheritance, maternal and environmental effects, the conventional breeding program is challenging.
Various findings have suggested that 35,384 QTLs are associated with 716 different traits (http:// www. anima lgeno me. org/ cgi-bin/ QTLdb/ SS/ index). Among these, 1274 QTLs are associated with litter traits [111]. In a study on two Duroc population i.e. U.S and Canadian, GWAS was conducted across two parities based on GeneSeek Porcine 50 K Chip data. Total 76 SNPs related to litter traits were identified in both population. However 10 significant SNPs were determined in Canadian population. Intriguingly, 13 pleiotropic SNPs for litter traits were found SSC 7, 9, 14 and 15 [111]. GWAS on Large White and Landrace breeds have also yielded notable results. In a study, samples were genotyped using Illumina Porcine SNP60 Bead Chip and quality control of SNPs was undertaken. Further GWAS analysis revealed that in Large White, significant markers were found on SSC5 and SSC10 [112]. PPARα gene was mapped near the marker on SCC5. Expression of the gene was significantly higher in endometrial tissues during early gestation and was lower during maternal recognition of pregnancy and after implantation in Polish LR and Pietrain breed [113]. This indicated the vital role of gene during pregnancy and thereby its close proximity with the marker associated with litter trait. Another gene integrin β 1 (ITGβ1) was closely associated with marker in SSC10. The gene is reported to significantly affect litter size in Large White and Landrace [114]. However in Landrace, significant SNPs were associated on SSC7, SSC9, SSC11 and SSC16 [112]. Another finding in Landrace and Large White identified 80, 227 and 187 SNPs affecting TNB, NBA and LWB (litter weight born alive) respectively. Of these 22 loci were shared by the three component traits of litter size. In addition four candidate genes affecting litter size across six parities were also suggested [115]. Twenty potential SNPs and several candidate genes underlying litter traits have also been identified in Duroc pigs through genome wide association study [116]. Another GWAS study in Duroc population identified 10 putative regions associated with litter traits such as NBA, number of stillborn (NS) and mummified piglets (NM). Within these regions seven candidate genes were identified. In addition it was inferred that the genome wide significant SNPs in the candidate genes identified were parity specific and therefore the effect of these genes might be temporal [117]. In Large White population, GWAS was undertaken using Porcine SNP80 bead chip and the results identified 29 significant SNPs within regions known to be associated with reproductive traits including litter size [118]. In Yorkshire pigs, five significant genes i.e. MSX1, spindlin 1 (SPIN1), VEGFA, FOXQ1, and LHFPL3 regulating TNB and NBA through different physiological pathways were identified [119]. In Bama Xiang pigs, 29 significant SNPs and 12 genes for litter size traits were identified [120].
These findings suggest that GWAS analysis advances our understanding regarding genetic mechanism of litter size traits and thereby aid in efficient selection for genetic improvement of pigs. However there are still few limitations which can be overcome by single traits meta-analysis for identifying SNPs across multiple population. This will improve the power of identification of genetic information and will provide thorough insight of litter traits to undertake selection and enhance genetic improvement.

Conclusion
Genetic improvement of litter size is of immense interest for pig producers owing to the fact that improvement in feeding regime and housing system is limited. Initial improvements were made in growth and carcass traits which led to constant or even decreased litter size in pigs. However, improvement in litter size is not only economically vital but also indicates breeding efficiency of pigs. Therefore, selection for litter size is greatly emphasized and carried out in most selection programs.
Litter size is sex limited trait with low heritability and phenotyping of the trait is time consuming hampering efficient improvement using conventional selection. In order to maximize profit from piggery various efforts have been made using different strategies. Marker assisted selection along with conventional selection is one such approach that is considered most effective. Notable success has been observed using candidate gene approach specifically in reproduction related markers. Several genes play biological role in improving litter size and polymorphisms in these genes serve as crucial markers for litter size selection. These genes can be identified with recent technical approach such   as genome wide association studies (GWAS). Further using GWAS, genomic selection can be applied for accurate and efficient selection response (Figs. 1, 2, 3).