Genome-wide association studies and candidate genes networks affecting reproductive traits using Iranian Holstein sequence data

Abstract

Background

The reproduction process in domestic animals is one of the most important challenges of animal husbandry. Fertility is an important trait that contributes to herd profitability and can be improved by genomic information. One of the best ways to investigate the association between single nucleotide polymorphisms (SNPs) and phenotypic performance is the genome-wide association study (GWAS). The aim of our study was to identify the genomic regions affecting reproductive traits, interval between first and last insemination (IFL), days open (DO), days from calving to first service (DFS), number of services per conception (NSPC), age at first calving (AFC) and age at first insemination (AFI) using SNP chip data in Iranian Holstein cows.

Results

GWAS analysis for all reproductive traits based on the significant-association threshold P < 1 × 10^–8, led to the identification of 55 single nucleotide polymorphisms (SNPs) for IFL (n = 3), DFS (n = 0), DO (n = 5), NSPC (n = 5), AFI (n = 33), and AFC (n = 9) traits. Based on the results of gene ontology analysis, 54 different candidate genes for reproductive traits were identified in this study. For IFL, NSPCC, DO, AFC, and AFI traits 4, 8, 11, 10, and 21 candidate genes were identified in the vicinity of significant SNPs, respectively. Key genes with biologically important positions for heifers (ATG7, PTPN5, STAC, GAD2, PLXDC2, KARS1, PRIM2, and ZNF597) and cows (LPL, SERP2, BIRC6, CENPU, PIK3C3, and MYLK3) can be mentioned.

Conclusions

Our results identified 55 marker-trait associations (MTAs) and 54 different candidate genes associated with reproductive traits. As a result, the SNPs and candidate genes discovered in this study can be used in genomic experiments to improve the reproductive performance of Iranian Holstein dairy cows and provide new information about the genetic architecture of these traits.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12864-025-11744-1.

Background

Following the increase and growth of the population in the world, there is a need to provide animal protein for people [1]. On the other hand, due to the limited land and water resources in the world, cattle management is necessary for higher productivity [2]. In this regard, paying attention to reproductive traits can have a great impact on production in dairy cows. Reproductive traits are among the most important traits in dairy cattle breeding [3]. Non-economic reproduction can lead to an increase in the calving interval, which reduces the replacement rate in dairy herds, and then the genetic improvement in the herd decreases, which ultimately results in a decrease in milk production as well as economic loss [4]. Improvement of reproductive traits, in addition to the effect they have on herd reproduction, can affect treatment and veterinary costs, animal welfare, and most importantly, the number of inoculations [3]. Today, reproductive problems are among the main reasons for the lack of profitability in dairy cattle herds [5]. The decline in reproductive efficiency in recent years has been significant in many studies. Research shows that the rate of stillbirth and mortality in cows, especially in dairy races, is increasing [6]. This issue has not only increased concern about the health and comfort of animals but has also caused a lot of economic damage to the dairy cattle industry. Therefore, the importance of fertility should be considered more in breeding programs [7]. The availability of data for the date of insemination allows animal breeders to calculate reproductive traits, e.g., AFI, AFC, IFL, DFS, DO, CI (calving interval), and GL (gestation length). In addition, fertility can be measured using categorical records such as NRR (non-return rate) and NSPC (number of services per conception) [3]. Reproductive traits are less responsive to genetic improvement through selective breeding due to their low heritability [3]. Reproductive traits are influenced by environmental factors and genetic factors. Environmental factors include season, temperature, nutrition, and management, which have a great impact on the length of the reproductive period [8].

During the past years, there have been studies on the identification of quantitative trait loci (QTL) on various traits in dairy cows, which has led to the identification of many QTLs affecting these traits on different chromosomes [9]. However, since measuring reproductive traits is not as easy as measuring productive traits [10], and due to the complexity of the structure of genes in these traits, fewer studies have been conducted in this field. Association mapping, or linkage disequilibrium (LD), is a relatively new and promising method to study quantitative traits. In this technique, the recombination events that have occurred during distant times until today are used for mapping [11–14]. In this technique, it is possible to identify quantitative trait control sites that are associated the desired trait or even causal polymorphisms within the gene that are related to two alternative traits [11, 15]. Association mapping is based on the hypothesis of linkage disequilibrium (LD), which can be useful for identifying the association between haplotype blocks and the desired trait [16–19]. Many studies using GWAS have been performed on reproductive traits in dairy cows and have identified candidate genes for these traits [3, 20–23]. The whole genome scanning study can help to better identify genes and variants related to economic traits, and this information can help our understanding of the mechanisms of the studied traits. Since the goal of breeding is to increase the economic efficiency of the system, it is necessary to investigate the reproductive traits and genes affecting these traits using the GWAS method and find the loci affecting these traits in different Holstein cattle populations. Ultimately, these data can help expand our understanding of the structure of genes, reproductive biology of female cattle and improve genetic predictions within a race in a specific region [24].

Interest in SNP markers has increased in recent years because they are the smallest unit of genetic variation and the basis of most genetic variation between individuals. On the other hand, in recent years, many single nucleotide mutations, or SNPs, have been discovered at the genome level, which has made it possible to prepare a genetic map with high density [25, 26]. These high-density SNP chips have provided an extraordinary throughput for genotyping and genome scanning of economically important traits in dairy cattle [27]. Abdulahi Arpanahi et al. [28] reported that the correlation between SNP effects for heifer and cow fertility traits ranges from 0.29 to 0.41. On the other hand, previous studies showed that the genetic correlation between fertility traits recorded on heifers and cows is less than one [29, 30]. Therefore, we strongly believe that genetic and biological mechanisms regulating reproductive traits in heifers and cows are not the same, and selection programs should be designed separately. The present study is the first to be conducted on the reproductive traits of Iranian Holstein cows based on the whole genome. While the previous studies have been conducted based on low-density chips, which are far less accurate, therefore, in the present study, the GWAS method was used to identify genomic regions affecting reproductive traits in Iranian Holstein heifers and cows. which can provide valuable information to breeders. Because there is a lack of information in this field in the dairy cattle herds of Iran, it is necessary to conduct a more comprehensive study.

Results

Descriptive statistics for reproductive traits are presented in Table 1. In the least squares analysis of the studied traits, the effect of contemporary group and age was statistically significant for all traits (p < 0.05). The results showed that the highest amount of variation was related to the trait IFL (126.99%), and the lowest amount was assigned to the trait AFC (7.19%).

Table 1.

Descriptive statistics of the studied reproductive traits in Iranian Holstein

Traits	N	Mean	Minimum	Maximum	Range	Variance	Coeff of Variation	Std Error
AFC	192	24.05	20.41	31.44	11.03	2.99	7.19	0.13
AFI	202	14.47	11.70	19.26	7.56	1.50	8.46	0.09
DFS	189	63.71	35.50	136.00	100.50	311.69	27.71	1.28
IFL	189	62.39	0.00	384.50	384.50	6277.70	126.99	5.76
NIPCC	190	2.63	1.00	15.00	14.00	4.02	76.21	0.15
OD	189	127.66	40.50	564.50	524.00	7653.66	68.53	6.36

Genome-wide association studies for reproductive traits in Holstein cows

The results of the GWAS analysis for all reproductive traits were reported based on two significant thresholds: P < 1 × 10^–6 and P < 1 × 10^–8. For the DFS trait, based on the significance threshold (P < 1 × 10^–8), no significant SNP markers were observed, but based on the significance threshold (P < 1 × 10^–6), 24 marker-trait associations (MTAs) were identified, which were respectively located on chromosomes BTA1, BTA9, BTA10 (two markers), BTA11, BTA13 (six markers), BTA25 (six markers), and BTA26 (seven markers) (supplementary 1 Fig. 1A). Three significant marker-trait associations (P < 1 × 10^–8) in BTA8, BTA20, and BTA24 regions and 15 significant marker-trait associations (P < 1 × 10^–6) in BTA4, BTA5, BTA8, and BTA14 regions. nine markers), BTA20 (two markers), and BTA24 for the IFL trait were detected (Fig. 1). For the OD trait, 5 SNP markers based on a significance threshold of P < 1 × 10^–8, were observed in the BTA2, BTA6 (two markers), and BTA14 (two markers) regions. In addition, based on the significance threshold of P < 1 × 10^–6, 59 MTAs were identified on the chromosomes BTA2 (six markers), BTA5, BTA6 (26 markers), BTA9, BTA14 (11 markers), BTA17 (three markers), BTA18, BTA25 (eight markers), and BTA26 (two markers) (supplementary 1 Fig. 1B). The results showed that five SNP markers had a significant relationship with the NSPC trait based on the significance threshold of P < 1 × 10^–8, which were located on the BTA24 (four markers) and BTA28 chromosomes. Based on the significance threshold of P < 1 × 10^–6, seven MTAs were identified, which were located at BTA24 (four markers) and BTA28 (three markers) positions (supplementary 2 Fig. 1C). All significant markers for reproductive traits in cows are reported in Supplementary 2.

Fig. 1.

Manhattan plot of the genome-wide p values of association for interval between first and last insemination trait in cow. The solid line represents the p < 1 × 10⁻⁶ and p < 1 × 10⁻⁸ significance threshold

Genome-wide association studies for reproductive traits in Holstein heifers

The results for the AFI trait, the number of 35 SNP markers, were able to cross the significance threshold line of P < 1 × 10^–8 in the regions BTA1, BTA2, BTA6, BTA7 (three markers), BTA11 (nine markers), BTA13 (three markers), BTA15 (two markers), BTA18, BTA20 (three markers), BTA22 (nine markers), BTA25, and BTA29 were located (Fig. 2). For the AFC trait, nine SNP markers were identified based on the significance threshold of P < 1 × 10^–8 in the BTA18, BTA25, and BTA26 regions (seven markers), respectively (supplementary 1 Fig. 2). Also, based on the significance threshold of P < 1 × 10^–6, 401 and 375 significant SNP markers were identified for AFI and AFC traits, respectively. All significant markers for reproductive traits in cows are reported in Supplementary 2.

Fig. 2.

Manhattan plot of the genome-wide p values of association for age at first insemination trait in cow. The solid line represents the p < 1 × 10⁻⁶ and p < 1 × 10⁻⁸ significance threshold

QTL regions for reproductive traits in Holstein cows

A summary of significant SNPs (P < 1 × 10^–8) associated with reproductive traits in cattle that are near the reported QTLs is presented in Table 2. The results showed that in the chromosomes BTA20 and BTA24 in the vicinity of the SNP affecting the IFL trait, there are important QTLs with related traits: milk yield (MY), milk fat yield (FY), milk protein yield (PY), muscularity (MUSC), body weight (BW), interval to first estrus after calving (CALEST), feed conversion ratio (FCR), stillbirth (SB), fertility index (FERIND), heifer pregnancy (HPG), first service conception (FSC), inseminations per conception (CONCEPT), finishing precocity (FPREC), calving ease (CALEASE), foot angle (FANG), pregnancy rate (PREGRATE), and calving index (CALVIND) were identified (Table 2). In the proximity of SNPs related to NSPC traits in chromosomes BTA24 and BTA28, QTLs related to traits: carcass weight (BW), body height (BW), milk protein yield (PY), udder height (UHT), interval to first estrus after calving (CALEST), interval from first to last insemination (INSINT), foot angle (FANG), and calving interval (CALVIND) were observed (Table 2). Further, for the DO trait, QTLs are associated with some traits: milk protein percentage (PP), milk protein yield (PY), body height (BW), milk fat yield (FY), calving ease (CALEASE), foot angle (FANG), interval from first to last insemination (INSINT), gestation length (GLENGTH), and rib thickness (RIBTH) was identified (Table 2).

Table 2.

QTLs located in close distance to the most significant single nucleotide polymorphisms (SNPs) associated with reproductive traits in Holstein cows

Trait	SNP name	CHR	SNP position	QTL trait	QTL symbol
IFL	20:35332061	20	35332061	Somatic cell score	SCS
				Sperm motility	SPMOT
				Milk yield	MY
				Milk fat yield	FY
				Milk protein yield	PY
				Stature	STA
				Muscularity	MUSC
				Stayability	STAY
	24:39426460	24	39426460	Body depth	BD
				Carcass weight	CWT
				Body weight	BW
				Sole ulcer	SULC
				Milk fat percentage	FP
				Interval to first estrus after calving	CALEST
				Bovine tuberculosis susceptibility	BTBS
				Milk fat yield	FY
				Feed conversion ratio	FCR
				Milk protein percentage	PP
				Average daily gain	ADG
				Body height	HEIGHT
				Stillbirth	SB
				Interval from first to last insemination	INSINT
				Fertility index	FERIND
				Milk protein yield	PY
				Residual feed intake	RFI
				Milk yield	MY
				Milk potassium content	MK
				Feed efficiency	FEEDEFF
				Heifer pregnancy	HPG
				Sperm motility	SPMOT
				First service conception	FSC
				Stillbirth	SB
				Body length	BL
				Milk yield	MY
				Milk protein yield	PY
				Inseminations per conception	CONCEPT
				Conception rate	CONCRATE
				Finishing precocity	FPREC
				Calving ease	CALEASE
				Foot angle	FANG
				Pregnancy rate	PREGRATE
				Feed efficiency	FEEDEFF
				Calving index	CALVIND
				Antral follicle number	AFOLN
				Rump width	RUMWD
NSPC	24:14252019,	24	14252019,	Body weight	BW
	24:14817979,		14817979,	Carcass weight	CWT
	24:14841115,		14841115,	Longissimus muscle area	LMA
	24:14847278		14847278	Calving ease	CALEASE
				Conception rate	CONCRATE
				Udder swelling score	USS
				Feed conversion ratio	FCR
				Stillbirth	SB
				Milk stearic acid content	MFA-C18:0
				Finishing precocity	FPREC
				Feet and leg conformation	FTLEG
				Dressing percentage	DRESSPERC
	28:43469861	28	43469861	Teat placement—front	FTPL
				Carcass weight	CWT
				Body height	HEIGHT
				Marbling score	MARBL
				Calving ease	CALEASE
				Milk protein percentage	PP
				Stillbirth	SB
				Body weight	BW
				Udder height	UHT
				Metabolic body weight	MBW
				Scrotal circumference	SCRCIR
				Conformation score	CONF
				Milk protein yield	PY
				Subcutaneous fat thickness	SUBFAT
				Abnormal flavor intensity	ABFLAV
				Foot angle	FANG
				Longissimus muscle area	LMA
				Interval to first estrus after calving	CALEST
				Palmitoleic acid content	FA-C16:1
				Pelvic area	PELAR
				Interval from first to last insemination	INSINT
				Heifer pregnancy	HPG
				Milk palmitoleic acid content	MFA-C16:1
				Milk C16 index	MC16
				Sole hemorrhage	SHEM
				Meat color	MCOL
				Average daily gain	ADG
				Udder swelling score	USS
				Calving interval	CALINTV
				trans-12-C18:1 fatty acid content	FA-C18:1n6
OD	2:15024774	2	15024774	Myristoleic acid conten	FA
				Immunoglobulin G level	SIGG
	14:80517355, 14:80519317	14	80517355,  80519317	Milk protein percentage	PP
				Milk protein yield	PY
				Tick resistance	TICKR
				Body weight	BW
				Clinical mastitis	CM
				Somatic cell score	SCS
				Milk fat percentage	FP
				Milk fat yield	FY
				Marbling score	MARBL
				Gestation length	GLENGTH
				Interval from first to last insemination	INSINT
				Dry matter intake	DMI
				Calving ease	CALEASE
				Foot angle	FANG
				Muscularity	MUSC
				Maintenance efficiency	MAINEF
				trans-Vaccenic acid content	FA-C18:1t
	6:15796637, 6:15814726	6	15796637, 15814726	Calving ease	CALEASE
				Rib thickness	RIBTH

QTL regions for reproductive traits in Holstein heifers

The results of significant SNPs (P < 1 × 10^–8) associated with reproductive traits in heifers that are near the reported QTLs are presented in Table 3. The results showed that in the BTA26 chromosome, in the vicinity of the SNPs affecting the AFC trait, there are QTLs with the traits: milk fat yield (FY), milk yield (MY), milk protein yield (PY), milk fat percentage (FP), stillbirth (SB), bone quality (BQ), interval from first to last insemination (INSINT) and udder attachment (UA) were observed (Table 3). QTLs associated with traits such as carcass weight (CWT), meat color (MCOL), and dry matter intake (DMI) were also observed in the vicinity of effective SNPs in the BTA18 chromosome. In the vicinity of significant SNPs in the BTA25 chromosome, there are QTLs related to traits: age at puberty (PUBAGE), milk fat yield (FY), milk yield (MY), conception rate (CONCRATE), bone percentage (BONEP), stillbirth (SB), calving ease (CALEASE), and number of embryos (EMBN) was discovered (Table 3). Near significant AFI SNPs in chromosomes BTA1, BTA2, BTA6, BTA7, BTA11, BTA13, BTA15, BTA18, BTA20, BTA22, and BTA25 and positions related to traits: gestation length (GLENGTH), milk protein percentage (PP), meat color (MCOL), longissimus muscle area (LMA), calving ease (CALEASE), number of embryos (EMBN), fertility index (FERIND), body weight (BW), feed conversion ratio (FCR), and mean corpuscular volume (MCV) were detected (Table 3).

Table 3.

QTLs located in close distance to the most significant single nucleotide polymorphisms (SNPs) associated with reproductive traits in Holstein heifers

Trait	SNP name	CHR	SNP position	QTL trait	QTL symbol
AFC	26:35254329,	26	35254329,	Milk fat yield	FY
	26:35260042,		35260042,	Milk yield	MY
	26:35261296,		35261296,	Milk protein yield	PY
	26:35261978,  26:35261830,		35261978,	Milk fat percentage	FP
	26:35263013,		35261830,	Stillbirth	SB
	26:35263174		35263013,	Bone quality	BQ
			35263174	Interval from first to last insemination	INSINT
				Udder attachment	UA
				Structural soundness	SOUND
	18:29660489	18	29660489	Carcass weight	CWT
				Meat color	MCOL
				Dry matter intake	DMI
	25:3032618	25	3032618	Age at puberty	PUBAGE
				Milk fat yield	FY
				Milk yield	MY
				Conception rate	CONCRATE
				Bone percentage	BONEP
				Stillbirth	SB
				Calving ease	CALEASE
				Number of embryos	EMBN
AFI	7:32022337,	7	32022337,	Gestation length	GLENGTH
	7:31973977,		31973977,
	7:31986904		31986904	Milk protein percentage	PP
	2:88625030	2	88625030	Meat color	MCOL
	15:40739864	15	40739864	Udder swelling score	USS
				Shear force	SF
	20:9880631,	20	9880631,	Longissimus muscle area	LMA
	20:9897995,		9897995,	Calving ease	CALEASE
	20:9907264		9907264	Immunoglobulin G level	SIGG
	25:41217739	25	41217739	Residual feed intake	RFI
				Number of embryos	EMBN
	13:32439203,		32439203,	Fertility index	FERIND
	13:32442132,		32442132,
	13:22165240		22165240	Body weight	BW
	11:12676017,	11	12,676,017,	Finishing precocity	FPREC
	11:12680526,		12680526,
	11:12683949		12683949
	18:41119642	18	41119642	Interdigital hyperplasia	IDH
				Somatic cell score	SCS
				Residual feed intake	RFI
				Body weight	BW
				Milk fat yield	FY
				Longissimus muscle area	LMA
	22:55540238,	22	55540238,	Milk fat yield	FY
	22:55540506,		55540506,	Milk energy yield	EY
	22:55540517,		55540517,	Milk yield	MY
	22:55540539,		55540539,	Feed conversion ratio	FCR
	22:55555856		55555856	Mean corpuscular volume	MCV

GO for reproductive traits in Holstein cows and heifers

The results of gene ontology analysis identified more than 200 genes for reproductive traits in heifers and Holstein cows, among which 54 important genes related to the desired traits are around (500 Kb higher than SNP and 500 Kb lower than SNP) SNPs correlated with the studied traits were detected (Tables 4 and 5). For the IFL trait, four candidate genes were identified in the vicinity of SNPs 8:66,532,988 (two numbers), 20:35,332,061, and 24:39,426,460, which were effective on the activity of the LPL, CNTNAP3, DAB2, and EPB41L3 genes. Eight candidate genes were detected for the NSPCC trait, which affected the activity of the PIK3C3, ERCC6, PARG, VSTM4, ERCC6, and ARHGAP22 genes. For the DO trait, 11 different candidate genes were identified in the vicinity of 2:15,024,774 (7), 6:15,796,637, 6:15,814,726, 14:80,517,355 (4), and 14:80,519,317 (4), which affect the activity of SERP2, BIRC6, CYP8B1, CENPU, CDK10, MYLK3, ASH1L, SLC10A5, SNX16, IMPA1, and CHMP4C genes (Table 4). The results of gene ontology analysis for the AFC trait led to the identification of 10 candidate genes in the vicinity of SNP, 29:25,831,763, which affect the activity of DST, EEF1A1, NPM1, KARS1, PRIM2, ALDH9A1, MIS18A, ADCY9, ZNF597, and ZNF93 genes. For the AFI trait, 21 different candidate genes were observed in the vicinity of significant SNPs, which affect the activity of some genes such as TMEM86A, UEVLD, CSRP3, PTPN5, TSG101, CRTAM, SLC6A1, GRHL1, HPCAL1, etc. (Table 5).

Table 4.

The candidate or nearest genes to the most significant single nucleotide polymorphisms (SNPs) in significant regions based on P < 1 × 10⁻⁸ for reproductive traits in Holstein cows

Trait	SNP name	SNP position	CHR	Gene start	Gene end	Ensembl gene ID	Gene name
IFL	8:66532988	66532988	8	66032988	67032988	ENSBTAT00000017086,	lipoprotein lipase (LPL),
						ENSBTAT00000008454,	contactin associated protein-like 3(CNTNAP3)
	20:35332061	35332061	20	34832061	35832061	ENSBTAG00000016152	DAB adaptor protein 2(DAB2)
	24:39426460	39426460	24	38926460	39926460	ENSBTAG00000019251	erythrocyte membrane protein band 4.1 like 3(EPB41L3)
NSPCC	24:14252019	14252019	24	13752019	14752019	ENSBTAT00000077983,	phosphatidylinositol 3-kinase catalytic subunit type 3(PIK3C3)
						ENSBTAT00000067289,
						ENSBTAT00000037295
	24:14817979	14817979	24	14317979	15317979	ENSBTAT00000077983,	phosphatidylinositol 3-kinase catalytic subunit type 3(PIK3C3)
						ENSBTAT00000067289,
						ENSBTAT00000037295
	24:14841115	14841115	24	14341115	15341115	ENSBTAT00000077983,	phosphatidylinositol 3-kinase catalytic subunit type 3(PIK3C3)
						ENSBTAT00000067289,
						ENSBTAT00000037295
	24:14847278	14847278	24	14347278	15347278	ENSBTAT00000077983,	phosphatidylinositol 3-kinase catalytic subunit type 3(PIK3C3)
						ENSBTAT00000067289,
						ENSBTAT00000037295
	28:43469861	43469861	28	42969861	43969861	ENSBTAT00000046144,	ERCC excision repair 6, chromatin remodeling factor (ERCC6),
						ENSBTAT00000079760,	poly (ADP-ribose) glycohydrolase (PARG),
						ENSBTAT00000008254,	V-set and transmembrane domain containing 4(VSTM4),
						ENSBTAT00000083009,	ERCC excision repair 6, chromatin remodeling factor (ERCC6),
						ENSBTAT00000010363	Rho GTPase activating protein 22 (ARHGAP22),
OD	2:15024774	15024774	2	14524774	15524774	ENSBTAG00000051524,	stress associated endoplasmic reticulum protein family member 2(SERP2),
						ENSBTAG00000027932,	baculoviral IAP repeat containing 6(BIRC6),
						ENSBTAG00000034106,	cytochrome P450 family 8 subfamily B member 1(CYP8B1),
						ENSBTAG00000018864,	centromere protein U(CENPU),
						ENSBTAG00000033333,	cyclin dependent kinase 10(CDK10),
						ENSBTAG00000014818,	myosin light chain kinase 3(MYLK3),
						ENSBTAG00000003954	ASH1 like histone lysine methyltransferase (ASH1L)
	6:15796637	15796637	6	15296637	16296637	ENSBTAG00000003954	ASH1 like histone lysine methyltransferase (ASH1L)
	6:15814726	15814726	6	15314726	16314726	ENSBTAG00000003954	ASH1 like histone lysine methyltransferase (ASH1L)
	14:80517355	80517355	14	80017355	81017355	ENSBTAT00000027342,	solute carrier family 10 member 5(SLC10A5),
						ENSBTAT00000078298,	sorting nexin 16(SNX16),
						ENSBTAT00000015548,	inositol monophosphatase 1(IMPA1),
						ENSBTAT00000006291	charged multivesicular body protein 4C(CHMP4C)
	14:80519317	80519317	14	80019317	81019317	ENSBTAT00000027342,	solute carrier family 10 member 5(SLC10A5),
						ENSBTAT00000078298,	sorting nexin 16(SNX16),
						ENSBTAT00000015548,	inositol monophosphatase 1(IMPA1),
						ENSBTAT00000006291	charged multivesicular body protein 4C(CHMP4C)

Table 5.

The candidate or nearest genes to the most significant single nucleotide polymorphisms (SNPs) in significant regions based on P < 1 × 10⁻⁸ for reproductive traits in Holstein heifers

Trait	SNP name	SNP position	CHR	Gene start	Gene end	Ensembl gene ID	Gene name
AFI	29:25831763	25831763	29	25331763	26331763	ENSBTAG00000048059,	transmembrane protein 86A(TMEM86A),
						ENSBTAG00000013568, ENSBTAG00000002488, ENSBTAG00000011869, ENSBTAG00000020257,	UEV and lactate/malate dehyrogenase domains (UEVLD), lactate dehydrogenase C(LDHC), cysteine and glycine rich protein 3(CSRP3), protein tyrosine phosphatase non-receptor type 5(PTPN5),
						ENSBTAG00000017446,	E2F transcription factor 8(E2F8),
						ENSBTAG00000013563	tumor susceptibility 101(TSG101)
	22:55681344	55681344	22	55181344	56181344	ENSBTAG00000000842, ENSBTAG00000005193	ubiquitin associated and SH3 domain containing B (UBASH3B), cytotoxic and regulatory T cell molecule (CRTAM)
	22:55540238	55540238	22	55040238	56040238	ENSBTAG00000005857, ENSBTAG00000035827	solute carrier family 6 member 1(SLC6A1),
							autophagy related 7(ATG7)
	22:55540506	55540506	22	55040506	56040506	ENSBTAG00000005857, ENSBTAG00000035827	solute carrier family 6 member 1(SLC6A1),
							autophagy related 7(ATG7)
	22:55540517	55540517	22	55040517	56040517	ENSBTAG00000005857, ENSBTAG00000035827	solute carrier family 6 member 1(SLC6A1),
							autophagy related 7(ATG7)
	22:55540539	55540539	22	55040539	56040539	ENSBTAG00000005857, ENSBTAG00000035827	solute carrier family 6 member 1(SLC6A1),
							autophagy related 7(ATG7)
	22:55555856	55555856	22	55055856	56055856	ENSBTAG00000005857, ENSBTAG00000035827	solute carrier family 6 member 1(SLC6A1),
							autophagy related 7(ATG7)
	15:33112811	33112811	15	32612811	33612811	ENSBTAG00000005857, ENSBTAG00000035827	solute carrier family 6 member 1(SLC6A1),
							autophagy related 7(ATG7)
	11:92438410	92438410	11	91938410	92938410	ENSBTAG00000001140, ENSBTAG00000001141,	isoamyl acetate hydrolyzing esterase 1 (putative)(IAH1), ADAM metallopeptidase domain 17(ADAM17), grainyhead like transcription factor 1(GRHL1),
						ENSBTAG00000007485,	TATA-box binding protein associated factor, RNA polymerase I subunit B(TAF1B), hippocalcin like 1(HPCAL1), cytidine/uridine monophosphate kinase 2(CMPK2), KLF transcription factor 11(KLF11),
						ENSBTAG00000007543,	radical S-adenosyl methionine domain containing 2(RSAD2)
						ENSBTAG00000004259, ENSBTAG00000019979,
						ENSBTAG00000046218,
						ENSBTAG00000016061
	11:87794544	87794544	11	87294544	88294544	ENSBTAG00000001140, ENSBTAG00000001141,	isoamyl acetate hydrolyzing esterase 1 (putative)(IAH1), ADAM metallopeptidase domain 17(ADAM17), grainyhead like transcription factor 1(GRHL1),
						ENSBTAG00000007485,	TATA-box binding protein associated factor, RNA polymerase I subunit B(TAF1B), hippocalcin like 1(HPCAL1), cytidine/uridine monophosphate kinase 2(CMPK2), KLF transcription factor 11(KLF11),
						ENSBTAG00000007543,	radical S-adenosyl methionine domain containing 2(RSAD2)
						ENSBTAG00000004259, ENSBTAG00000019979,
						ENSBTAG00000046218,
						ENSBTAG00000016061
	22:10534777	10534777	22	10034777	11034777	ENSBTAG00000006735	SH3 and cysteine rich domain (STAC)
	22:10539654	10539654	22	10039654	11039654	ENSBTAG00000006735	SH3 and cysteine rich domain (STAC)
	22:10547117	10547117	22	10047117	11047117	ENSBTAG00000006735	SH3 and cysteine rich domain (STAC)
	13:32439203	32439203	13	31939203	32939203	ENSBTAG00000009475	plexin domain containing 2(PLXDC2)
	13:32442132	32442132	13	31942132	32942,132	ENSBTAG00000009475	plexin domain containing 2(PLXDC2)
	13:22165240	22165240	13	21665240	22665240	ENSBTAG00000009475	plexin domain containing 2(PLXDC2)
	11:12676017	12676017	11	12176017	13176017	ENSBTAG00000009475	plexin domain containing 2(PLXDC2),

Gene networks

The results of gene network analysis for reproductive traits including IFL, DO, NSPC, AFI, and AFC are shown in Figs. 3 and 4. Based on the results of gene network analysis, there were contributions among candidate genes through co-expression, pathways, physical interactions, and shared protein domains for reproductive traits. Most of these contributions were related to physical interactions between genes, which indicated protein–protein interaction, and if two genes showed the same protein–protein interaction, their products were linked. Therefore, the candidate genes identified in our study had significant protein–protein interactions with each other or with related genes.

Fig. 3.

Gene networks analysis for interval between first and last insemination (A), days open (D), number of services per conception (C) and reproductive traits in Holstein cows (D). Dark circles with and without slash represent candidate genes and associated genes, respectively. Arrows in pink, blue, red and bone color represent co-expression, pathway, physical interactions and shared protein domains, respectively

Fig. 4.

Gene networks analysis for age at first insemination (A), age at first calving (B) and reproductive traits in Holstein heifers (C). Dark circles with and without slash represent candidate genes and associated genes, respectively. Arrows in pink, blue, red and bone color represent co-expression, pathway, physical interactions and shared protein domains, respectively

Discussion

Identification of genomic regions affecting reproductive traits can provide new opportunities to improve fertility in dairy cows. However, this study introduced genes and biological networks affecting reproductive traits in Holstein cattle [3].

Based on our GWAS results, three SNP markers in BTA8, BTA20, and BTA24 regions with the IFL trait, five markers in BTA2, BTA6 (two markers), and BTA14 (two markers) regions for the DO trait, and five markers on chromosomes BTA24 (four markers) and BTA28 showed a significant relationship with the NSPC trait based on the significance threshold of P < 10^–8. The results obtained from this study are consistent with the results of other studies [3, 21, 31, 32]. Chen et al. [31] reported that the SNP markers that had a significant relationship with the IFL trait were in the BTA8, BTA7 and BTA20 regions. Mohammadi et al. [3] reported that several SNP markers located in BTA19 and BTA20 had a significant effect on the IFL trait. Eight QTLs located on BTA2, BTA1, and BTA26 chromosomes were significantly correlated with the DO trait [32]. Most SNPs affecting the NSPC trait were located on chromosomes BTA24, and BTA28 [21]. A marker affecting the age at first insemination (AFI) trait was observed in the cM 25 region of chromosome number two [33]. A SNP in the BTA29 region affecting this trait was also reported in Swiss-bred cattle [31]. In another study, Alves et al. [34] reported significant regions in BTA7 for AFI. In our study, significant SNP markers were observed the AFC trait in the BTA18, BTA25, and BTA26 regions (seven markers). In a recent study, 19 significant SNP markers for AFC were in the BTA13, BTA25 and BTA26 regions [35]. In the last research conducted by Nascimento et al. [36] on Nellore heifers, it was found that significant markers on the AFC trait were mostly located in the BTA18, BTA21, and BTA26 regions.

The results of QTLs identified in this study affecting the IFL trait were consistent with the results of Chen et al. [31] and Mohammadi et al. [3] on Holstein cows. Considering that the structure of the udder in cattle is one of the most important traits in profitability, the more suitable the structure of the udder is, the better the milk production will be, and the probability of udder infection in livestock and the number of somatic cells will decrease. The results of studies have proven the relationship between breast health traits, milk production [37], and reproductive traits [21]. The results of a study showed that five candidate genes with high performance, which were in the form of haplotype blocks, had an effective effect on stillbirth and fertility traits [38]. The results of this study show that the decrease in fertility rate can be affected by the residual properties of the consumed feed [39]. Because the negative energy balance affects the amount of insulin produced and reduces the production of growth hormone, it will eventually lead to a decrease in hepatic IGF-1 synthesis [40]. In this study, around the observed SNP, QTLs were observed that were correlated with milk quantity traits and udder infection, which can indicate that the animal became pregnant later at the peak of lactation. Around the SNP observed in this study, several QTLs affecting multiple births were observed. In the study of Widmer et al. [41] on chromosome 11, many QTLs in the form of haplotype blocks had a significant relationship with fertility and calving traits. In this study, it was also reported that the correlation between calving and fertility traits and the trait of multiple birth was negative and unfavorable. In the vicinity of SNPs related to IFL, DO, and NSPC traits, QTLs were observed in this study, which are related to body weight traits, which can be a proof of the claim that heifers should reach their ideal weight at the time of insemination. That this weight will be different depending on different breeds, for example, in Brahman breed cows, cows with higher weight had more inappropriate fertility [39]. In this study, QTLs related to calving ease were identified. The results of the study show that calving ease in the Brown Swiss breed is affected by body weight, so cows with higher weight have more difficulty calving and calving interval will be higher [42]. Around the SNP associated with the number of services per conception trait, a QTL was also effective on the fertility index trait. Fertility index is the result of multiplying the capture index by the estrus index, and since the inverse capture index is the trait of the NSPC, the relationship between these two traits is not far from expected [22]. The identification of QTLs affecting stillbirth and puberty age in this area shows the important role of the identified markers on the age at first calving trait. In general, and experimentally, heifer body weight, calf size, Keppel's angle, and age at the first insemination can have an effect on the ease of reproduction trait. The results of a study proved the moderate genetic correlation between calving ease traits, calving weight, and fetal survival [43]. In the last study by Nascimento et al. [36] that was conducted on Nellore heifers, it was found that in the vicinity of significant markers on the trait of age at first calving, there were QTLs with traits of milk yield, milk protein yield, milk fat percentage, non-return rate, and stillbirth. Finally, the QTLs identified in the margins of significant SNPs confirm the fact that the age of the first insemination and the age of the first calving can be effective on production traits, reproduction, longevity, and finally on the profitability of the herd in these conditions. It can be said that herd management plays an important role in advancing breeding goals [20]. Also, the identification of the QTLs related to traits such as stillbirth, calving ease, number of embryos, length of pregnancy, percentage of milk protein, fertility index, and body weight in these areas has already been confirmed by other researchers [20, 33, 34].

Four key LPL, CNTNAP3, DAB2, and, EPB41L3 genes were identified for the IFL distance trait. The Lipoprotein lipase (LPL) gene is a key enzyme for fat metabolism and plays an essential role in fat synthesis in adipose tissue and milk. LPL is also expressed in other places, including mammary gland, nervous system, heart, liver, pancreatic islet cells, and lungs [44]. Several key genes were identified for the DO trait. The SERP2 gene plays a role in udder infection. The effect of this gene on camel udder infection has been confirmed. Udder infection is among the things that affect the days open traits [45]. The MDFI gene is a gene that inhibits muscle growth, this gene can also affect growth traits and prevent the proper growth of livestock [46]. The BIRC6 gene causes fetal growth at the beginning of pregnancy. Embryo absorption in early pregnancy increases the number of inseminations per pregnancy [47]. In the BTA2 chromosome, two genes, CYP7A1 and CYP8B1, are involved in the conversion of cholesterol into bile acids, which play an important role in nutrient absorption, digestion, fat, and glucose regulation. These genes also play a role in the positive regulation of cholesterol biosynthesis in heifers and increasing milk production two weeks after giving birth, especially in multiple births [48]. The CENPU gene in BTA2 was effective on embryo survival in dairy cows. This gene has an abundance of 0.08% in French Holstein cows and can be among the genes that cause embryo death in cows [49]. The CDK10 gene plays an important role in fat distinctness. This gene is one of the 185 genes in the CDK family. In this study, another gene from this family (CDK2) in BTA13 had a SNP effect on the age of first insemination [50]. Eight key genes were identified for the NSPC trait, among them, the PIK3C3 gene was observed in BTA24 in the margins of SNPs affecting NSPC trait, which plays a role in growth factors, embryo growth during pregnancy, and protein synthesis. Also, PIK3C3 plays a role in the development of skeletal muscle 21 days after delivery, and this protein plays a role in the transition period and negative energy balance in protein balance. Postpartum energy balance is effective on the number of inseminations per pregnancy [51].

Some key genes were identified for the AFI trait, e.g., in BTA15, the CRTAM gene was observed. This gene is a regulator of T cells, and it has been reported that this gene has a relationship with the immune response in cattle [52]. It seems that the ATG7 gene in BTA22 is effective for muscle and bone growth. The results of a study on Japanese black beef cattle showed that the expression of ATG7 increased during obesity in skeletal muscles [53]. In BTA29, the E2F8 gene is one of the essential genes for the development and proliferation of the cell cycle. This gene is also effective in the growth and development of the cow follicle. On the other hand, this gene affects the growth of the fetus and the development of the uterine system in the first pregnancy [54]. The Tsg101 gene is required for the maintenance of uterine epithelial cells, and its deletion may cause the breakdown of these cells and lead to embryo damage at the beginning of implantation [55]. The GRHL1 gene affects the process of yolk sac formation at the beginning of pregnancy and plays an important role in placenta development [56]. In general, the function of this gene on reproductive traits, including abortion and stillbirth, has been proven [57]. The KLF11 gene in BTA11 is a member of a gene family called KLFs. which affects muscle fat content [58]. Muscle fat content is one of the determinants of meat quality traits in livestock. There are also several different KLFs that regulate cell growth and differentiation in mammals and, like the previous gene, have a great effect on pregnancy traits [58]. In BTA22, the STAC gene (SH3 and cysteine rich domain) plays an important role in the growth and contraction of skeletal muscles. STAC3 expression is observed during the differentiation of bovine satellite cells during weight gain and muscle growth [59]. The MLLT10 gene in BTA22 is associated with fertility traits. It plays a role in the biological processes of mammary gland morphogenesis, uterine growth, morphogenesis and mesoderm formation and blood vessel growth. The expression of this gene may be affected in the reproductive traits of goats under natural or artificial selection [60]. Among other identified genes, we can mention that the PLXDC2 gene in BTA11 is effective on the weight trait. The results of a study conducted on Lori Bakhtiari sheep showed that this gene can be considered a candidate gene for body weight traits after weaning [61]. PHYH is another gene that is effective on fatty acid oxidation and fatty acid metabolism affects successful fertility, which shows that there is a relationship between immune traits and reproduction [62]. In BTA11, the sperm associated antigen 6 (SPAG6) gene, sperm motility, the percentage of sperm with forward progress, is one of the characteristics of semen quality, which is highly related to male fertility potential in farm animals. Low sperm motility disrupts reproductive efficiency. Studies have shown that SPAG6 protein is a positive regulator for testis growth, spermatogenesis, and regulation of sperm movement in chickens [63].

For the AFC trait in BTA13, the DST gene plays an important role in fighting the virus, especially the herpes virus in cattle, and is one of the genes that improves the immune system. The results of a study showed that the DST gene is effective on the survivability of female bulls [64]. The EEF1A1 gene, which encodes one of the most abundant plasma proteins, plays an important role in the translation machine. In a study conducted on dairy cows, it was found that there is a high correlation between body weight gain, muscle growth, and feed residue. It can be concluded that this gene plays a role in tissue and muscle growth, and the presence of this gene next to the SNP associated with the AFC trait confirms the effect of this SNP [65]. In BTA26, the PRIM2 gene has been effective on carcass weight traits. The results of a study conducted on dairy cows showed that this gene was effective on carcass weight, daily weight gain, and maturity weight [66]. The ADCY9 gene influences the improvement of reproductive organs in heifers. In general, the reproductive efficiency of heifers directly depends on the health of the reproductive system, especially the ovaries and uterus. In the last research conducted on the genes of the ADCY family, it was found that this gene group has a great effect on the formation of uterine tissue and ovaries in Xiangxi cattle [67]. The bovine ZNF597 gene is another maternally expressed gene, and its expression is regulated by DNA methylation. It was shown that this gene was effective in the formation of placenta tissue and breast cell structure in dairy cows [68]. However, since cows were selected based on the estimated breeding value of milk yield in this study, QTLs and genes identified around significant SNPs may increase milk or fat production and thus improve reproductive traits in cattle.

Conclusions

The use of GWAS facilitates the combination of genomic data in the genetic evaluation of reproductive traits in Holstein cattle. We have identified several SNPs, significant regions in different BTAs, and a list of locus candidate genes (new and known) that may contribute to variation in reproductive traits in Holstein cattle. The most important QTLs identified included: interval from first to last insemination (INSINT), age at puberty (PUBAGE), calving ease (CALEASE), fertility index (FERIND), and body weight (BW). Key genes with an important biological position for heifers (ATG7, PTPN5, STAC, GAD2, PLXDC2, KARS1, PRIM2, and ZNF597) and cows (LPL, SERP2, BIRC6, CENPU, PIK3C3, and MYLK3) can be mentioned. Which can be used as key genes in future programs. In general, these results can help to better understand the studied traits and develop and improve them.

Materials and methods

Phenotypic data

After receiving permission from two companies affiliated with Ferdous Pars Agriculture and Animal Husbandry Development Holding for sampling and receiving company information from Iran's Livestock Breeding and Improvement Center, information about the cows of these companies was received. It is worth mentioning that the breeding value of the milk production trait for these cows was estimated by that center through the following equation (Eq. 1).

$${\text{y}}_{\text{ij}}=\mathrm{\mu }+{\text{hys}}_{\text{i}}+{\text{a}}_{\text{ij}}+{\text{e}}_{\text{ij}}$$ 1

where y_ij is milk yield (adjusted to 305 days and twice a day milking); μ is the population mean; hys_i is the effect of herd-year-season group i; a_ij is the animal breeding for j^th animal and i^th herd-year-season group, and e_ij is the random residual effects.

The cows were sorted according to the breeding value of the milk production trait from large to small, and 210 female cows born between 2012–2013 were selected for this study (150 and 60 cattle, respectively, in herds no. 1 and 2) [69]. These 210 animals were divided into two subpopulations of 105 based on the highest and lowest breeding values of milk yield.

In addition to the mentioned cases, the following cases were also considered during the sampling: during the sampling, the pedigree analysis of the livestock was done using the CFC software [70] and tried livestocks with minimal kinship relationships were selected so that the diversity of livestock in both herds was high [71]. The selected animals had complete pedigrees and records, and it was ensured that the selected animals were not candidate animals for elimination.

In this study, the reproductive traits of Holstein heifers and cows (lactation 1 to 3), collected from 2013 to 2020, were used. The following traits, including AFI, AFC, DFS, DO, IFL, and NSPC, were analyzed. Attributes are presented and defined in Table 6.

Table 6.

Definition of the studied reproductive traits in Iranian Holstein

Traits	Definition
Age at first insemination (days)	1st insemination date – birth date
Age at first calving (days)	Calving date – birth date
Interval between first and last insemination (days)	Last – first insemination date
Days from calving to first service (days)	1st insemination date – previous calving date
Days open (days)	Last insemination date – previous calving date
Number of services per conception	Number of services / conception

Genotype imputation and quality control (QC)

Hair samples were collected from 210 Iranian Holstein cows in a breeding population (Firdous Pars dairy farm, Isfahan, Iran) and no anesthetic was used in this study. One-hundred forty and 60 animals from herd 1 and 2 were genotyped by the GGP-LD v.4 SNP panel (with 30,108 SNPs) and the Affymetrix Axiom Bovine Array-50 K (with 51,987 SNPs), respectively. To control the quality of genotyping, four criteria were used in the PLINK 2.0 software. In this criterion, animals with more than 5% of missing genotypes were excluded, SNPs with minor allele frequency (MAF) less than 5% were excluded, and SNPs that were not genotyped for more than 5% of animals and chi scores were less than 10–6. Chi-square < 10⁻⁶) were excluded from the Hardy–Weinberg equilibrium test. Minimac3 software [72]. was used for imputation to check the accuracy of imputation and identify and remove markers with lower accuracy and stepwise imputation.

First, 210 cows were genotyped based on 30 K and 50 K SNPs, and the target population was named. Then, based on the reference population including 234 cows (the 1000 Bull Genome Project database containing 129, 43, 15, and 47 key progenitors from the global Holstein–Friesian, Fleckoyer, Jersey, and Angus breeds, respectively) with BovineHD beadchip and the whole-genome, imputation was performed [33]. The same quality control was performed on BovineHD beadchip and the sequence data, and a total of 578,505 SNPs and 12,063,146 SNPs were retained for analysis after filtering, respectively. In addition, Eagle (version 2.3) was used to phase genotypes before using Minimac3 for reference and target populations separately [73]. Finally, genotypes assigned with an accuracy of less than 0.30 were excluded, and a total of 6,583,595 SNPs were retained for further analysis [69]. Finally, the remaining 6,583,595 SNPs were used for GWAS analysis.

Genome-wide association studies

Analysis of the relationship between the imputed genotypes and the studied traits was performed using a general linear model (GLM) in EMMAX software [74]. The equation used in GWAS is as follows (Eq. 2):

$$\text{y}=\text{Xb}+\text{Zu}+\text{e}$$ 2

In this model, y is the vector of phenotypic values, X is the matrix of coefficients for fixed factors, including mean, fixed effects, random effects, and SNP effects. b is the vector of fixed effects, Z is the matrix of coefficients of additive genetic random effects, u is the vector of additive genetic random effects, and e is the vector of residual effects. In addition, $Var\left(u\right)={\sigma }_g^{2}K$ and $Var\left(e\right)={\sigma }_e^{2}I$, where K is a genomic kinship matrix calculated using genotype data and then used in the model, and I is an identity matrix.

Significance levels

In this research, we used the following three methods to identify the threshold of significant association in the GWAS study: 1- The threshold of P < 1 × 10^–8 suggested by Reed et al. [75] 2- Benferoni correction threshold for the whole genome, which is equivalent to P < 1 × 10^–9 (total number of SNPs / P < 0.05). 3- Benferoni correction threshold per chromosome (total number of SNPs per chromosome/P < 0.05) has been proposed as an alternative solution [76]. Because the Benferoni correction for the whole genome does not consider the connection between SNPs, this approach increases conservatism and the false negative rate [77]. Two significant thresholds were considered: P < 1 × 10^–6 and P < 1 × 10^–8. To draw the Manhattan plot, the qqman package was used in the R 4.3.2 software [78].

Gene annotation

Ensembl annotation of the UMD3.1 genome version (http://www.ensembl.org/biomart/martview) was used to identify candidate genes around (within one megabase) SNPs that passed the threshold of P < 1 × 10^–8. DAVID Bioinformatics Resources version 6.7 (http://david.abcc.ncifcrf.gov/) was used for gene ontology analysis. In addition, to identify the QTLs around (within 1 Mb) SNPs that passed the threshold of P < 1 × 10^–8, the QTLdb of cattle (https://www.animalgenome.org/cgibin/QTLdb/BT/index) was used. Finally, gene networks were drawn by GeneMANIA (http://genemania.org/).

Authors’ contributions

NM performed the experiments and data analysis and wrote the article draft; MS and AJS supervised the project and provided editorial input on the writing. SRMA and MK contributed to data analysis and writing the article draft. All authors discussed the results and contributed to the final manuscript. The author(s) read and approved the final manuscript.

Funding

This research did not receive any specific funding.

Data availability

The datasets generated and analyzed during the current study are available in the Figshare repository [https://doi.org/10.6084/m9.figshare.28604060].

Declarations

Ethics approval and consent to participate

The samples collected from the studied animals were performed in accordance with animal ethics and approved by the Animal Use Committee of the University of Tehran and the National Animal Breeding Centre of Iran. In addition, permission for sampling was obtained from the farmers on site. Also, we have obtained the informed consent of the owner(s) to use the animals in our study in this area from the University of Tehran and the National Animal Breeding Center of Iran.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.