Phenotypic statistics and genotypic data
Descriptive statistics for TNB across parities are presented in Table 1, including breed, sample size, mean, standard deviation (SD) and coefficient of variation (CV). The number of individuals decreased progressively with increasing parity, from 1,898 sows in the first parity to 400 in the sixth parity; subsequent parities were excluded owing to insufficient sample sizes. Mean TNB values for parities 1-6 were 13.256, 13.809, 14.581, 14.305, 13.995 and 14.196, respectively. The average TNB per parity was highest in Yorkshire, followed by Landrace and Duroc across breeds. Overall, the SD of the TNB across the six parities was within an acceptable range, with CV below 30%. These results indicated that the phenotypic data were relatively stable.
A total of 40,021 qualified SNPs were retained for analysis after quality control and imputation. The genomic distribution of the SNP markers, visualized using the R package rMVP, is shown in Fig 1A. The PCA results for the post-quality control genotyping data generated using the R package rMVP are presented in Fig 1B. The analysis revealed three distinct principal components consistent with the sampling design of the three pig breeds included in the study. Thus, principal components need to be incorporated into the GWAS analysis to control for false positives.
Duroc, Landrace and Yorkshire pigs are the major commercial breeds used in modern swine production and play critical roles in global pork supply chains. Among the porcine reproductive traits, TNB is a key determinant of pig farming efficiency, Although high TNB may increase the incidence of stillbirth and mummified piglets
(Raguvaran et al., 2017). Given the low heritability of TNB, it is difficult to improve this trait through conventional breeding.
GWAS results for TNB
Through a GWAS of TNB across six parities we identified 14 genome-wide significant SNPs associated with TNB (the significant SNP locus on the X chromosome for the TNB6 trait was excluded from the analysis). In the case of TNB1, one significant SNP was mapped: TNB2 and TNB3 each harbored two significant SNPs and TNB4, TNB5 and TNB6 each had three significant SNPs (Fig 2, Table 2). Among these 14 SNPs, chromosomes 1 and 2 each contained three significant SNPs; chromosome 8 contained two and chromosomes 5, 7, 13, 14 and 17 each contained one SNP.
Notably, significant SNPs varied across parities and no common candidate genes were shared among litters. These results align with previous suggestions that sow litter size may be influenced by transient genetic effects and that physiological traits differ across sow ages
(Wu et al., 2018), thus enhancing our understanding of the genetic architecture underlying TNB in domestic pigs.
Haplotype analysis
Haplotype analysis within a 500-kb window upstream and downstream of the significant SNPs revealed three haplotype blocks in the vicinity of the significant locus ASGA000138, with sizes of 11, 120 and 14 kb, respectively (Fig 3). Among these haplotype blocks, the 120-kb block was annotated to
AKAP12, whereas the remaining two blocks did not map to any annotated genes. Haplotype analysis of the remaining sites failed to identify any haplotype blocks.
(Kang et al., 2021) reported that a 3’UTR region insertion mutation in
AKAP12 is significantly associated with first litter size and TNB in goats, where the mutation disrupts miR-181 binding and alters
AKAP12 spliceosome expression. In the present study of swine, the annotation of
AKAP12 suggests its potential role in regulating litter size in sows, although functional validation through further experiments is warranted.
QTL mapping and gene annotation of significant SNPs
Identification of the QTLs containing significant SNPs was conducted using AnimalQTLdb; The results revealed that their mapped QTL regions were predominantly linked to reproductive traits, growth traits and immunobiochemical indicators. Specifically, three SNPs ALGA0006774, ASGA0001308 and WU_10.2_2 _137721029 on chromosome 1 were localized to QTL regions with overlapping functions, primarily associated with gestation length, teat number, weaning weight and TNB. SNP WU_10.2_5_8582550 on chromosome 5 was mapped to a QTL region linked to first birth weight and stillbirth rate. On chromosome 8, SNP WU_10.2_8_133554012 was physically proximal to ALGA0049505, with both residing in overlapping QTL regions associated with nipple number, litter size, corpus luteum count and live births. The remaining SNPs were located in QTL regions predominantly related to growth and immune traits.In this study, the QTL regions where the significant loci identified by GWAS are located were characterized, thereby inferring the effects of these loci on the trait. Notably, most of these loci were also mapped to QTL regions associated with reproduction, which validates their value.
Annotation of the GWAS associated SNP loci using the Ensembl database identified nine SNPs mapped to genes on chromosomes 1, 2, 5, 7, 8, 14 and 17 (Table 3), including the genes
TSPAN18,
ADAMTS19,
RAC2,
UNC5C,
SSTR3,
SLC24A4,
ADK and
PTPRT that showed significant associations with total number born (TNB). Based on QTL mapping and literature review,
SSTR3 and
UNC5C were considered promising candidate genes strongly related to TNB.
The
SSTR3 gene encodes Somatostatin Receptor 3, which belongs to the G-protein-coupled receptor family. Somatostatin is an important neuroendocrine regulatory peptide that regulates various physiological processes, including hormone secretion, neural signaling and cell proliferation, by binding to its receptor. In a study of differential gene expression in gonadotropin-releasing hormone neurons in female mice, the
SSTR3 gene was identified
(Vastagh et al., 2015) and in another study on the effects of growth inhibitors on growth and reproduction in tilapia, the
SSTR3 gene was found to be enriched in FSH and LH cells
(Mizrahi et al., 2024), suggesting that
SSTR3 may affect sow reproductive performance by regulating hormone secretion; however, the specific mechanisms require further exploration.
UNC5C encodes a Netrin-1-dependent receptor protein that regulates axon guidance during embryonic development in collaboration with Netrin-1 (
Yu and Bargmann, 2001;
Round and Stein, 2007). In a GWAS of Holstein cow conception rate, a single nucleotide polymorphism (SNP, A+169G) in the 3' untranslated region (3’UTR) of
UNC5C was found to be significantly associated with conception rate. This SNP correlates with
UNC5C expression levels and influences preimplantation embryonic development
(Sugimoto et al., 2015), suggesting a critical role for
UNC5C in embryonic survival. In the present study, an intronic SNP in
UNC5C was mapped to a QTL region associated with litter size, implying that the mutation may affect the litter size of sows by altering
UNC5C expression. The experimental validation of this mechanism is warranted in future studies.
The remaining genes identified in our analysis, namely
TSPAN18,
ADAMTS19,
RAC2,
SLC24A4,
ADK and
PTPRT, may participate in embryonic angiogenesis
(Li et al., 2021), premature ovarian failure
(Fonseca et al., 2015), uterine inflammation
(Doye et al., 2024), Ca
2+ transport
(Jalloul et al., 2016), energy metabolism (
Boison and Jarvis, 2021) and embryonic survival
(Chen et al., 2024; Pan et al., 2020), which may collectively influence sow reproductive performance.
In this study, each parity was analyzed as an independent trait, which revealed that significant SNPs and candidate genes associated with reproductive traits varied across parities. To further dissect the genetic architecture of sow reproduction, future GWAS should be conducted separately for different parity groups with expanded sample sizes, which may uncover more accurate loci and deepen our understanding of the genetic mechanisms underlying sow reproductive traits.
