4.1. Heritabilities
Heritability estimates (
h2) are crucial for understanding the genetic influence on these traits and their potential for improvement through selective breeding. The heritability of MY305 (0.262 ± 0.005) suggests that a moderate proportion of variation in this trait is attributable to genetic factors. While genetic selection can enhance MY305, effective management strategies are necessary, especially in challenging environments, to achieve substantial progress (
Table 2). The high additive genetic variance (
) further supports the potential for genetic improvement in MY305. The heritability estimate for MY305 observed in this study aligns with values previously reported in the literature, ranging from 0.19 to 0.27 for Holstein cows [
38,
39,
40,
41] and from 0.24 to 0.27 for Guzerat cows [
42,
43]. Beyond breed differences, the lactation stage also influences heritability variations. Pangmao et al. [
44] reported that the heritability for MY305 was generally higher during the first lactation compared to subsequent lactations. Additionally, differences in management practices, climate conditions, and methodological approaches across populations can significantly affect MY305 estimates and heritability [
45].
However, the low heritability of DO (0.029 ± 0.0004) suggests that most of the variation in this trait within the population was due to management practices, with genetic factors playing a minor role. As a result, selective breeding may have a limited impact on improving DO, and alternative strategies, such as enhancing environmental conditions and reproductive management, may be more effective. The heritability of DO observed in this study was lower than the values reported for Holstein cows, which range from 0.04 to 0.07 [
46,
47,
48,
49], and for Russian Black-and-White cattle (0.07) [
50]. Additionally, it was lower than the values previously reported for the same population in 2015 (0.04 to 0.05) [
19], but comparable to values reported in a more recent study [
22]. The low heritability value may be due to the strong influence of fixed factors on DO and the high residual variance observed in this study (
). A previous study identified factors affecting DO, including Holstein genetic proportions and heat stress in hot climates [
51,
52]. Studies from other subtropical regions have suggested that environmental stress can increase genetic variability in dairy cattle [
49,
53]. In Thailand’s tropical climate, heat stress poses a significant challenge for livestock, potentially compromising reproductive performance. Additionally, variations in the month and year of calving across different populations may have contributed to the low heritability estimates of DO. To improve this trait, strategies such as better reproductive management, heat stress control, skilled artificial insemination, modern milking facilities, and optimized nutrition should be implemented [
54,
55].
Low heritability (0.102 ± 0.002) was observed for milk FPR, indicating that fixed factors and residual variance were the primary contributors to its variation within the population. Similar heritability values have been reported in Thai-Holstein crossbreeds (0.17) [
17] and Nordic cattle (0.14) [
19]. In contrast, higher heritability values have been observed in Holstein cows, ranging from 0.25 to 0.54 [
11,
56,
57,
58,
59]. The variation in milk FPR heritability estimates across studies may be due to differences in breed, estimation methods, and model effects. Genetic differences between cows and environmental factors such as heat stress have also been reported to influence milk FPR [
56]. Additionally, a study utilizing random regression test-day models reported higher milk FPR heritability values [
59]. Genetic improvement of milk FPR through selection requires careful management strategies to achieve meaningful progress, particularly in preventing metabolic disorders, such as acidosis and ketosis.
4.2. Genetic Correlations Between MY305, DO, and Milk FPR
Table 3 shows the positive genetic correlation between MY305 and DO (0.559), suggesting that cows with a higher milk yield over 305 days tended to have extended DO. Consequently, selection for increased milk production may compromise fertility. This finding aligns with previous research on Holstein cows [
60,
61,
62], showing that nutrient allocation prioritizes milk production over reproductive functions [
63]. Additionally, management strategies that delay the first insemination in high-yielding cows may further contribute to prolonged DO [
64]. Moreover, high-producing dairy cows often experience a postpartum negative energy balance (NEB), which can significantly impair fertility by delaying ovulation [
13,
65]. These results emphasize the need for balanced breeding strategies to improve milk production while maintaining reproductive efficiency.
A negative genetic correlation (−0.306) was identified between MY305 and milk FPR, indicating that cows with higher MY305 often have lower milk FPR. Similar genetic correlations between MY305 and milk FPR have been reported in Thai-Holstein crossbreeds [
17,
56] and Nordic Red cattle [
19]. Negussie et al. [
19] found that genetic correlations between test-day FPR and MY during early lactation at 15, 30, and 60 days in milk (DIM) were small but positive (0.01 to 0.13), and after 60 DIM, these correlations became predominantly negative (0.01 to −0.22). These positive genetic correlations in early lactation suggest that high-producing cows mobilize body reserves to meet the energy demands of peak milk production, leading to a higher FPR. However, as lactation progresses, the genetic correlation between FPR and MY declines or becomes negative, indicating recovery from NEB. This trend is consistent with the findings in Canadian Holsteins [
18] and German Holsteins [
66]. In Thai dairy cattle, this condition may arise because of less intensive management during late lactation compared to early lactation, particularly among dairy farmers who continue to use traditional practices. As a result, these animals often experience inadequate nutrition management [
67].
One notable finding from this study is the negative genetic correlation between DO and milk FPR, which indicates that cows with higher DO tend to have a lower milk FPR. These results contradict previous studies that reported low-to-moderate positive genetic correlations between DO and milk FPR [
17,
19]. This may be attributed to a lack of awareness regarding the nutritional needs of dairy cows during lactation, particularly during the late lactation phase. An extremely low milk FPR indicates acidosis, a condition often caused by excessive concentrate feeding without sufficient high-quality forage. Overfeeding with concentrates leads to excessive starch fermentation, which increases lactic acid production and lowers rumen pH. This can negatively impact reproductive performance, resulting in infertility, prolonged calving-to-conception intervals, and increased incidence of lameness [
12,
15,
68]. Alternatively, the negative genetic correlation between DO and milk FPR may be influenced by differences in the time from calving to first service between trained officers and traditional farmers, which affects DO variation among cows [
67].
In our study, the phenotypic and environmental correlations among MY305, DO, and milk FPR were very low, with both positive and negative values close to zero, indicating that these traits share minimal common influences at the observable and environmental levels. This suggests that changes in MY305 are unlikely to have a direct impact on DO and milk FPR, and shared environmental factors do not significantly affect these traits together. As a result, improvements in milk yield management or genetic selection may not necessarily affect fertility and animal health.
4.3. GEBVs for MY305, DO, and Milk FPR
The breed group (BG) effects indicated that a higher proportion of Holstein genetics was associated with increased MY305, longer DO, and lower milk FPR across both methods (
Table 4). Holsteins have superior genetic potential for milk production compared to many other breeds, which explains why animals in BG1 exhibited the highest MY305. Studies have consistently shown that Holsteins outperform other breeds in terms of milk production [
69,
70,
71]. However, animals with higher Holstein genetics also tend to have a longer DO, possibly because of their reduced adaptability to tropical environments. This finding aligns with the study by Pongpiachan et al. [
69], who reported that purebred Holsteins exhibited lower reproductive efficiency, even when specialized management strategies were employed to mitigate the effects of tropical climates and enhance their diet. Consequently, crossbreeding has been proposed as a strategy to counteract reproductive decline associated with “Holsteinization” [
72]. A previous study found that crossbred cows exhibited improved reproductive performance, including shorter DO and calving intervals (CIs), higher conception rates at 28 days, and reduced incidences of mastitis, ketosis, and endometritis [
73]. Additionally, high-production Holsteins often experience greater energy deficits, leading to increased mobilization of body fat. This metabolic shift can alter the fat-to-protein ratio (FPR) in milk, often resulting in a lower FPR [
67].
The GEBV results presented in
Table 4 demonstrate that wssGBLUP consistently produced higher GEBVs than ssGBLUP, particularly for MY305 and DO, indicating that wssGBLUP captured genetic effects in the dataset more effectively. In the all-animal dataset, both methods yielded negative GEBVs for MY305, but the wssGBLUP values (−8.983 to −8.259) were closer to zero, suggesting enhanced GEBV accuracy. A similar trend of improvement was observed in the bull dataset, whereas in the dam dataset, the improvements were less pronounced for DO and milk FPR than for MY305. The GEBV for MY305 in this population was generally negative, in contrast to that of Russian Black-and-White cattle, which was reported to be 0.88 [
50]. A negative average GEBV for MY305 suggests that the population’s genetic potential for milk yield was below the baseline or the genetic mean defined by the reference population. This may have resulted from factors such as population structure or the genetic merit of the reference group used for comparison. Research has shown that the composition of the reference population plays a crucial role, as it directly affects the estimated breed composition and subsequent GEBVs [
74]. In this study, the reference population consisted of high-performing breeds, which might have led to lower GEBVs for animals from breeds with a lower genetic potential for MY305.
The positive GEBVs for DO indicate that animals have a higher genetic potential for longer reproductive intervals, which may not be desirable based on breeding goals favoring shorter calving intervals. However, the mean GEBVs for DO in the all-animal dataset were lower than those reported in Russian Black-and-White cattle (3.25 to 4.14) [
50] but comparable to recent findings of 0.266 to 0.274 [
22]. This suggests that the Thai-Holstein population has better genetic values for the DO trait than the Russian Black-and-White cattle population. This improvement may be attributed to the genetic advantages of crossbred cattle, which exhibit superior reproductive efficiency and overall performance compared with purebred Holsteins. For milk FPR, the GEBVs remained consistently negative or near zero across both the ssGBLUP and wssGBLUP methods, indicating that both models provided similar GEBV accuracy for this trait.
The wssGBLUP method consistently demonstrated a higher accuracy than ssGBLUP across all traits and groups (
Figure 1). For MY305, the accuracy increased by 27.54%, improving from 0.138 with ssGBLUP to approximately 0.176–0.177 with wssGBLUP. This highlights both the higher heritability of MY305 and the enhanced ability of wssGBLUP to capture key genetic markers compared to other traits. In contrast, DO and milk FPR exhibited smaller accuracy gains, with the averaged GEBV accuracy increasing from 0.112–0.113 to 0.124–0.125 (10.71%), and from 0.094 to approximately 0.110–0.112 (17.02%), respectively. These findings suggest that wssGBLUP improved accuracy by emphasizing key genetic regions, with notable gains for highly heritable traits, such as MY305. These results align with the findings of previous studies, where wssGBLUP provided greater accuracy for production traits, while yielding smaller gains for traits with lower genetic influence. Zhang et al. [
75] reported that wssGBLUP enhances accuracy by assigning different weights to SNPs, making it particularly effective for highly heritable traits in dairy cattle. The wssGBLUP method has been shown to effectively identify SNPs linked to traits such as protein content and to enhance the accuracy of GEBVs in Canadian Holstein cattle [
76]. Additionally, a study on Hanwoo beef cattle found that wssGBLUP improved the prediction accuracy for carcass traits such as carcass weight, eye muscle area, and yearling weight [
77]. In pig breeding, a study revealed that wssGBLUP offered improved estimation reliability compared to ssGBLUP for meat, fattening, and reproductive traits [
78].
wssGBLUP is superior to ssGBLUP because it can assign weights to single-nucleotide polymorphisms (SNPs), optimize predictions for specific traits, and iteratively refine accuracy. According to Teissier et al. [
79], wssGBLUP assigns different weights to SNPs based on their estimated effects, enabling the more precise identification of quantitative trait loci (QTLs) with significant impacts on traits. By emphasizing SNPs associated with major genes, wssGBLUP enhanced the accuracy of GEBVs for traits strongly influenced by these genes. Additionally, the iterative weighting process of wssGBLUP allows it to adapt to the genetic architecture of different traits, making it highly flexible in modeling both polygenic traits and those controlled by a few major genes [
77].
4.4. Identification of Genomic Regions and Candidate Genes
Manhattan plots of the SNP effects from GEBVs for MY305, DO, and milk FPR are shown in
Figure 2. The peaks of the Manhattan plots highlight impactful SNPs, with positive values at the top and negative values at the bottom.
Figure 3 presents Manhattan plots illustrating the percentage of additive genetic variance accounted for by the five SNP moving windows. Using a threshold of at least 5% of the total genetic variance, 14 SNPs were significantly associated with MY305 across seven chromosomes (BTA 4, 6, 13, 18, 24, and 25), 19 SNPs with DO across five chromosomes (BTA 2, 5, 13, 18, and 25), and 16 SNPs with milk FPR across five chromosomes (BTA 5, 18, 19, 21, and 25), all within a distance constraint of less than 50 kb (
Table 6 and
Table 7).
After comparing the results with NCBI databases, we identified 24 candidate genes associated with MY305, 46 genes associated with DO, and 33 genes associated with milk FPR. However, four genes from MY305 (
LOC618297,
LOC101906304,
LOC529511, and
LOC786948), nine genes from DO (
LOC101905166,
LOC107133069,
LOC101902869,
LOC100138951,
LOC618463,
LOC100141212,
LOC101905312,
C25H16orf58, and
LOC783313), and six genes from milk FPR (
LOC101905166,
LOC100138951,
LOC618463,
LOC100141212,
LOC100196898, and
LOC786948) have not been characterized (
Table 6 and
Table 7). We also identified seven genes from MY305 (
LOC618297,
LOC101906304,
PPARGC1A,
VSIG10L,
PARD6G, and
LMF1), ten genes from DO (
TNS1,
SOX10,
LOC101905166,
BAIAP2L2,
PDYN,
LOC618463,
VSIG10L,
LMF1,
ITGAD, and
SEPT14), and nine genes from milk FPR (
SOX10,
LOC101905166,
BAIAP2L2,
LOC618463,
VSIG10L,
LOC100196898,
MGAT5B,
MGAT5B, and
LINGO1) at the target SNP location. Hajihosseinlo et al. [
80] found that the r
2 value tends to decrease as the distance between SNP pairs increases, implying that SNPs positioned within 1 Mb are more likely to exhibit strong and consistent associations with QTLs.
Candidate gene analysis revealed genes in the same genomic region across all three traits, suggesting potential pleiotropic effects (
Figure 4). Specifically, 14 genes were shared among MY305, DO, and milk FPR; 17 genes were shared between MY305 and DO; 14 genes were shared between MY305 and milk FPR; and 26 genes were shared between DO and milk FPR. According to Gratten and Visscher [
81], pleiotropy is a genetic phenomenon whereby a single DNA variant influences multiple traits. This indicates that when selection targets one trait, other traits often change over generations. This response is driven by genetic correlations that reflect the combined genome-wide effects of pleiotropy at shared genetic loci. Identifying pleiotropic genes associated with MY305, DO, and milk FPR is challenging, owing to the complexity of these traits. Although limited research has directly connected these traits to pleiotropic genes, some studies have examined genes that affect multiple production traits [
35,
82] and multiple fertility traits in dairy cattle [
83,
84].
4.5. Pleiotropy and Candidate Genes for MY305, DO, and Milk FPR
Pleiotropy occurs when one gene affects multiple phenotypic traits. This study identified genes influencing MY305, DO, and milk FPR simultaneously. The pleiotropic effects of these genes are shown through their SNP variance and effect relationships (
Figure 5,
Figure 6 and
Figure 7).
Figure 5A,
Figure 6A, and
Figure 7A show positive relationships between SNP variance for MY305 and DO, FPR and DO, and MY305 and FPR, respectively. Most SNP variances are clustered near the origin, indicating that most SNPs have low variances across all traits, while a few SNPs exhibit strong pleiotropic genetic influences. Influential SNPs can be found in all figures, indicating significant pleiotropic SNPs for all three relationships. It can be observed that SNPs affecting high MY305 variation also affect high DO variation (
Figure 5A). SNPs influencing high FPR variation affect moderate DO variation (
Figure 6A). However, SNPs influencing high DO and FPR variation might contribute independently (
Figure 7A).
Figure 5B,
Figure 6B, and
Figure 7B demonstrate strong relationships between SNP effects for MY305 and DO, FPR and DO, and MY305 and FPR, respectively. Positive collinearity can be observed in SNPs affecting MY305 and DO, indicating that SNPs with large effects on MY305 also extend to DO (
Figure 5B). A negative relationship can be found in SNPs affecting FPR and DO, showing that SNPs with large negative effects on FPR tend to increase DO (
Figure 6B). Similarly, SNPs influencing MY305 and FPR have strong negative correlations, suggesting that high MY305 effects decrease the FPR (
Figure 7B). This implies that selecting cows for high milk yield may lower the FPR and prolong DO, posing a challenge, as a greater value for DO is linked to extended calving intervals and reduced fertility. These finding confirm previous studies reporting antagonistic genetic correlations between milk yield and fertility [
85,
86], indicating that higher milk yield might reduce the fat-to-protein ratio, affecting milk composition and metabolic efficiency.
These findings highlight the need for multi-trait selection strategies to balance milk yield, fertility, and metabolic efficiency. Breeding programs should use optimal selection indices to improve MY305, DO, and milk FPR. High-variance SNPs in these traits suggest strong genetic influences, making them key targets for marker-assisted selection (MAS). Identifying beneficial SNPs for all three traits can refine breeding strategies. Additionally, pleiotropic genes associated with all three traits highlight the genetic connections among production, fertility, and health traits.
The results reveal fourteen genes within the same region, indicating pleiotropy among MY305, DO, and milk FPR (
SIGLECL1,
IGLON5,
VSIG10L,
ETFB,
NKG7,
CLDND2,
LIM2,
SOX8,
SSTR5,
TEKT4,
C1QTNF8,
CACNA1H,
LOC786948, and
TPSB2). Some genes have been reported to be pleiotropic and influence production, reproduction, and health traits across various species. Recent studies highlight
SIGLECL1’s role in animal reproduction and immune regulation, with its expression detected in the male reproductive tracts of mice, rats, and bovine sperm [
87,
88]. Additionally, SIGLEC family polymorphisms have been associated with milk yield, DO, and calving intervals in cattle [
89,
90].
The following genes play a role in reproduction in some species.
LIM2 plays a role in spermatogenesis, with studies in mice showing that
Limk2-deficient individuals exhibit impaired testicular development and increased germ cell apoptosis [
91]. This gene also protects spermatogenic cells from stress-induced damage [
91].
TEKT4 is expressed in male germ cells, and is essential for sperm motility in mice [
92,
93]. Additionally,
TEKT4 has been identified as a key gene associated with sperm motility in Brahman cattle, based on proteomic studies [
94].
Three genes,
SSTR5,
ETFB, and
SOX8, have been linked to animal production and growth traits. SSTR5 polymorphisms in sheep are associated with growth traits, making them potential molecular markers for selective breeding [
95]. At the same time,
ETFB has been identified as a key gene influencing meat quality in Qinchuan cattle [
96]. Copy number variations (CNVs) of
SOX8 genes in yaks are significantly associated with growth traits such as withers height and chest girth [
97].
Among the genes related to immunity and behavior,
IGLON5 plays a critical role in immune function in cattle [
98,
99].
NKG7 enhances cytotoxic activity in CD8+ T cells and NK cells, promoting T cell accumulation [
100,
101].
CLDND2 plays a role in immune responses in cattle [
102].
CACNA1H knockout in mice leads to autistic-like behaviors [
103].
TPSB2 is linked to immune cells in adipose tissue in Holstein–Friesian cows [
104]. Additionally, three genes (
PDYN,
SIRPA, and
LMF1) were found to be shared between MY305 and DO.
PDYN shows signs of positive selection in dairy cattle, indicating its role in reproductive traits [
105].
SIRPA is a marker of spermatogonial stem cells, embryogenesis, and gametogenesis in mice [
106,
107,
108].
LMF1 is essential for the post-translational activation of lipoprotein lipase and other enzymes [
109,
110].
The results indicate that DO and milk FPR share 17 common genes (
MICALL1,
C5H22orf23,
SOX10,
POLR2F,
LOC101905166,
PICK1,
SLC16A8,
BAIAP2L2,
PLA2G6,
LOC100138951,
LOC618463, and
LOC100141212); most genes influence health and immunity, while others impact fertility.
BAIAP2L2 genes are essential for mechanotransduction [
111].
PLA2G6 encodes
iPLA2β, related to immunity and membrane homeostasis [
112].
LOC618463 is associated with livability in Holstein cattle [
113], and
LOC100138951 is linked to calf survival in Nordic Holstein cattle [
114]. The
SLC16A8 gene family includes 14 monocarboxylate transporters vital for metabolic processes [
115].
LOC618463 is also a target gene for calving traits [
116].
SOX10 is involved in sex determination [
117], and
PICK1 is essential for male fertility [
118].
Candidate genes in MY305 include
LOC618297,
SHFM1,
LOC101906304,
PPARGC1A,
LOC529511,
PARD6G, and
ADNP2. PPARGC1A polymorphisms are linked to milk fat yield [
119] and birth weight in Holsteins [
120], with specific SNPs affecting milk traits in Iranian Holsteins [
121].
PARD6 is associated with the Hippo signaling pathway [
122,
123], and
ADNP2 is highly expressed in embryonic brain tissues [
124].
Seventeen candidate genes influencing DO were identified, with key roles in reproduction (
TNS1,
MAFF,
LOC107133069,
LOC101902869,
ITGAD,
COX6A2,
ARMC5,
LOC101905312,
TGFB1I1,
SLC5A2,
RUSF1,
AHSP,
LOC783313,
OR7D4,
SEPT14,
ZNF713, and
MRPS17).
ITGAD is linked to fertility and tick resistance [
124,
125,
126], and
TGFB1I1 is important for ovarian development [
127]. Genes from the TGF-β pathway are essential for embryogenesis [
128]. The SLC5 family, especially
SLC5A1 and
SLC5A2, affects glucose transport and litter size in sheep [
129].
OR7D4 in horses and stallion testes plays a role in reproductive functions [
130].
MRPs like
MRPS17 are vital for embryogenesis [
131], and
MRPS17 and
SEPTIN14 are candidates for fertility traits in cattle [
83].
C25H16orf58 and
SLC45A2 are associated with clinical ketosis in Holsteins and heat tolerance in Chinese cattle, respectively [
132,
133].
Finally, five genes (
LOC100196898,
MGAT5B,
MFSD11,
LINGO1, and
ODF3L10) were associated with milk FPR. In dairy cattle, milk FPR is linked to NEB and the overall health condition.
MGAT5 is crucial for the synthesis of complex N-glycans and is essential for various biological processes.
MGAT5-deficient mice exhibit complex phenotypes, including susceptibility to autoimmune diseases and reduced cancer progression [
134,
135]. Additionally,
MGAT5 has been associated with mastitis resistance [
136].
MFSD11 is widely expressed in the brain and periphery, particularly in the neurons, and its expression is altered by changes in energy balance in mice [
137].
LINGO1, a protein selectively expressed in the central nervous system, functions as a negative regulator of oligodendrocyte differentiation, myelination, neuronal survival, and axonal regeneration and plays a key role in neural health [
138]. Research on agonistic behavior in Lidia cattle has also identified genomic regions containing
LINGO2 that are linked to behavioral traits [
139]. At the same time,
ODF3L10 (also known as Odf3) is primarily associated with sperm tail outer dense fibers (ODFs) in animals with internal fertilization [
140]. Most of these genes contribute to health-related traits in cattle, including neurological function, energy balance, immune response, and structural integrity.