Identification of genomic regions affecting nitrogen excretion intensity in Brown Swiss dairy cows Authors: Žan Pečnik, Daša Jevšinek Skok Published: 2024 ABSTRACT AbstractDairy cows with a lower nitrogen excretion intensity (Nexi) excrete less nitrogen, ammonia (NH3) and nitrous oxide (N2O), a highly potent greenhouse gas (GHG), per kg of milk produced and therefore represent a lower environmental impact while maintaining food security. To date, the genomics background of Nexi is unknown. Here we performed a genetic association study, overlap analysis and functional enrichment analysis for Nexi in 875 genotyped dairy cows with 2,147 lactations from 200 herds. We identified 1456 single nucleotide polymorphisms (SNPs) that significantly affect Nexi. We found 140 SNPs overlapping with 148 protein-coding genes. The MAN1A1 gene is a strong candidate gene for Nexi. Genotype CC of rs42786248, the most significantly associated SNP located in the MAN1A1 gene, had higher Nexi than genotypes AA (p < 0.001) and AC (p < 0.001). We identified 33 genes involved in biological processes related to nitrogen metabolism. Our results form the basis for further research on the genomics background of Nexi. The identified SNPs serve as potential targets for selective breeding programs, aimed at reducing Nexi and associated NH3 and N2O emissions in cattle production, thus contributing to more environmentally sustainable milk production. Introduction The excretion of various nitrogen (N) compounds from dairy cattle leads to environmental problems. As recently reported, the livestock sector urgently needs a global initiative to tackle N pollution while supporting food security.1 Cattle excrete nitrogen through feces and urea, which is further transformed into ammonia (NH3), and nitrous oxide (N2O), a highly potent greenhouse gas. In the latest IPCC assessment report2 the global warming potential (GWP100) for N2O was estimated at 273 ± 130, which is much higher compared to methane (CH4), where the GWP100 for non-fossil CH4 is 27.0 ± 11. In addition, CH4 remains in the atmosphere for 11.8 ± 1.8 years, while N2O accumulates for 109 ± 10 years2 and can be defined as long-lived. The livestock sector emits one-third of human-induced N emissions1 and N2O mitigation measures in dairy herds with declining or constant emissions will be more important than CH4 mitigation measures in the future as the warming effect of methane is often overestimated.3 N2O emissions originate from urine and manure nitrogen excretion (Nex) in processes of nitrification and denitrification. Most of the N-rich compounds are excreted in urine and are products of complex ruminant N-metabolism, where crude protein is digested by microorganisms to peptides, amino acids and NH3. Kohn et al.4 have reported a close relationship between the level of crude protein in the feed ration and the amount of N released into the environment. In this way, 75–85% of the excess proteins consumed by a dairy cow are excreted in urine, feces and milk.4,5 To obtain information on total Nex, measurements are laborious, expensive and impractical at large scales; therefore, models are needed to predict N excreted in urine and feces. Different models enable quantification of urinary, feces or total Nex and different phenotypes could be used.6 The total estimated Nex is subsequently used to calculate N2O emission at individual animal levels. Strategies to reduce Nex indirectly mitigate N2O emissions.7 Technical options for mitigation of Nex and N2O emissions were clearly defined,8,9 and breeding was recently recognized as a valuable contribution to the whole set of mitigation strategies that could be applied to achieve the temperature goals for 2050 set by the EU.10 Furthermore, farmers are actively seeking highly efficient cows concerning Nex and milk production.11 Efficient cows are those that, for the same amount of milk production, excrete lower amounts of N in both urine and feces, i.e., have lower nitrogen excretion intensity (Nexi). Nexi and N2O emission intensity are calculated by dividing the sum of Nex and total N2O emissions, respectively, by the total milk production over the selected timescale. Genetic selection for total Nex could be conflicting with other desirable traits in breeding programs, such as total milk production. This is because cows with higher milk production tend to excrete more nitrogen4 and selecting for such animals may result in lower overall milk production. Hence, we are interested in the genomics background of Nexi. Moreover, determining Nex and Nexi is important for assessing and managing N2O emissions in subsequent stages of manure management. If we identify genomic regions associated with Nexi, we can indirectly mitigate N2O emission intensity. Mitigation potential for N2O with genetic selection is uncertain due to limited research or lack of data.12 Some genetic studies aimed to reveal the genetic background for Nex-related traits. Honerlagen et al.13 identified genomic regions and candidate genes influencing nitrogen-excretory metabolites in the urine of Holstein-Friesians. These genes were ITPR2 and MYBPC1 (BTA 5), STIM2 (BTA 6), SGCD (BTA 7), SLC6A2 (BTA 18), TMCC2 and MFSD4A (BTA 16). Chen et al.14 identified 16 candidate genes associated with NEI, a trait that includes N intake, milk true protein N and milk urea N yield. These genes were mainly expressed in the milk cell, mammary and liver tissues, with several of them being involved in N metabolism. Examples include CA1, CA2, CA3, CA8, CA13, LOC784254 and LOC100847874, all of which are located on BTA14. Recently, Kong et al.15 reported that the length of cilia in kidney cells is associated with urine concentration in mice kidneys. The length of cilia is linked to the CFAP77 gene, which was recently also associated with urea N content in Holstein cattle.16 Pegolo et al.17 identified genomic regions associated with milk N fractions and highly interconnected genes in them. It was shown recently that high genetic merit Holstein cows in Northern Ireland utilize feed N for milk production more efficiently than lower genetic merit cows.18 Consequently, modern dairy cows can partition more ingested N into milk and excrete less N in feces and urine, per kilogram of standard milk.18 Manzanilla-Pech et al.19 conducted a genomic study on CH4 emission intensity, expressed as concentration and production. They linked three SNPs to both traits, namely Hapmap59221-rs29014908 and Hapmap44201-BTA-114510, both on BTA4, and Hapmap52436-rs29009653 on BTA6. However, to the best of our knowledge, there have been no studies investigating the genomic background of Nexi. Our study aims to fill the existing knowledge gap of the genomic background of Nexi in dairy cows. We focus on (1) the identification of single nucleotide polymorphisms (SNPs), candidate genes and gene ontologies associated with Nexi; (2) conducting association analysis, overlap analysis and functional enrichment analysis; (3) estimation of differences in N2O emissions between genotypes of most significant SNP associated with Nexi. Our ultimate objective is to establish a breeding strategy aimed at minimising Nexi, thereby making a significant contribution to the reduction of N2O emissions while sustaining milk production. Material and methods Dataset We used the data of 875 genotyped Brown Swiss dairy cows from the Slovenian cattle population, with phenotypes collected in the Central cattle database20 of the Agricultural Institute of Slovenia. Briefly, animal ID, calving ID, calving date, urea content and milk production in full lactation were recorded during the regular monthly milk recordings. As this was an in silico study, approval from the animal care and use committee approval was not required. The inter-calving period was calculated as the sum of days between two consecutive calvings. We used 2,147 lactation records collected in 200 herds from 2008 and 2022. Phenotype determination We estimated Nex from milk recording data for all 875 genotyped dairy cows. To calculate Nex (expressed in kg N per animal per year), we modified the following equation:21 129.9 + 0.0089 ∗ (milk − 7,744) + 1.7 ∗ (urea − 26), where ‘milk’ represents milk production (in kg per animal per year) and ‘urea’ represents urea content (in mg/dl). The values 0.882 for slurry and 0.800 for solid manure in the21 equations were redundant for our purpose. Our goal was to estimate the phenotype of the animals, so the volatilisation factors 0.882 or 0.800 were not included in our estimation. Furthermore, our milk recording data provided information on milk production throughout the lactation period (milk_f). Therefore, to obtain annual values, we adjusted to account for the inter-calving period, resulting in the final Nex equation: Nex = 129.9 + 0.0089 ∗ (milk_f ∗(365/IP) − 7,744) + 1.7 ∗ (urea − 26),where IP represents the inter-calving period in days. To assess the Nexi, we calculated Nexi as the ratio of Nex to milk yield during the full lactation period, adjusted for the inter-calving period (IP), using the formula: Nexi= Nex/(milk_f ∗ (365/IP)) This approach allowed us to provide a comprehensive estimation of Nexi while considering the variables of milk production, urea content and the inter-calving period. Ammonia and nitrous oxide emissions We assessed the emissions of NH3 and N2O from housing and storage areas, respectively, based on Nex to determine the total emissions per kilogram of milk and about candidate genes significantly associated with Nexi. We utilized the methodologies of the European Environment Agency22 and the Intergovernmental Penal for Climate Change.23,24 Due to the absence of data on farming practices, it was assumed that excreta were collected in the form of slurry without bedding residues, animals did not graze and no emission reduction measures were implemented on the farms. Since all excreta are collected in housing facilities, according to the EMEP/EEA22 guidelines, the annual N excreted in-house (mhous_N) is represented by the equation mhous_N = Nex holds. Additionally, given that all excreta are collected in slurry and housing facilities (mhous_slurry_TAN, where TAN stands for total ammonia nitrogen), the equation mhous_slurry_TAN = mhous_TAN holds. First, we calculated total ammonia nitrogen (TAN) as the amount of TAN deposited during housing: mhous_slurry_TAN = xTAN ∗ mhous_TAN. Next, we estimated ammonia emissions (NH3–N) as kg NH3-N by multiplying mhous_slurry_TAN by the emission factor EFhousing = 0.24.22 Next, we calculated TAN (entering store = TAN transferred from housing to the storage areas) as the difference between ammonia emissions and TAN deposited during housing ammonia (NH3–N) emissions by multiplying TAN (entering store) by the emission factor EFstorage = 0.25.22 Next, we assessed the indirect emissions of N2O from NH3 emissions in housing by considering that 1% of the emissions are converted into N2O as N2O–N,23 i.e., we multiplied indirect NH3 emissions by 0.01. To estimate the indirect emissions of N2O from NH3 emissions in storage, we multiplied TAN (entering the store) by 0.02.22 We then considered converting N2O-N to N2O by multiplying emissions by conversion factor 44/28 for both direct and indirect emissions. Finally, for conversion into CO2 equivalents, a GWP100 factor of 26525 was applied. Association study To conduct an association study, we combined genotypic datasets with phenotype data. A total of 875 cows were genotyped using SNP microarray ICBF International Dairy and Beef v3 (ICBF, Ireland) for a total of 53,262 SNP markers. Polymorphisms were annotated by SNPChimp v3.26 Polymorphisms without a mapped rs ID or commercial ID in the current genome assembly were assigned placeholder coordinates (chromosome 99, position 0) to differentiate them from SNPs originally located on chromosome 0 in platform files. Quality control (QC) of the genotype data was done using PLINK.27 For the details of QCs in this study, we refer to Anderson et al.28 Finally, a total of 38,427 SNPs were retained. The association analysis, using one SNP at a time, was performed with a linear mixed model using the ‘lme4’ package in R software.29 The single SNP regression analysis was performed using the following statistical model: (1)yijklmn= μ+Pi+Gj+Sk+b(xijklm− x¯)+cl+hm+eijklmn where yijklmn = Nexi (g/kg milk), µ = overall mean, Pi = parity (i = 1–3), Gj = genotype (j = 1–3), Sk = season (k = 1–4), xijklm = age at calving, cl = random additive effect for sire l (l = 1, 2, …, 227), hm = random effect for herd m (m = 1, 2, …, 200) and eijklmn = a random residual. To account for multiple comparisons and control the false discovery rate, we applied the false discovery rate (FDR) correction method. The FDR correction was used to adjust the p-values obtained from the mixed model analysis, ensuring a more stringent control of Type I errors. The significance threshold after FDR correction was set at 0.01. Genomic overlap analysis Genomic overlap analysis was performed comparing the location of SNPs associated with Nexi, and protein-coding genes. Genes were obtained from Ensembl Biomart Release 112, while gene nomenclature was based on the HUGO Gene Nomenclature Guidelines (http://www.genenames.org). Throughout the text, genomic positions of SNPs and genes were based on the ARS-UCD1.3 Bovine Genome Assembly. Views (graphical overview of the chromosomal locations) were constructed using the visualisation tool Flash GViewer (http://gmod.org/wiki/Flashgviewer/) developed by the GMOD project. Functional analysis Functional analysis for a set of SNPs associated with Nexi was conducted by the public web server GeneTrail3, where the analysis for 12 model organisms, including cattle can be done. The statistical significance threshold was set to p < 0.05. Over-representation analysis (ORA) as an enrichment algorithm with the FDR adjustment method was used to determine the number of statistically significant categories.30 We included the Ontology and Phenotype section (GO—Biological Process, GO—Cellular Component, GO—Molecular Function), Pathways (KEGG-Pathways, Reactome—Pathways and WikiPathways), Genomic positions from NCBI Assembly and analysis of protein families based on Pfam database. Results and discussion Phenotypes: nitrogen excretion and nitrogen excretion intensity in dairy cows The variability in milk production within the studied Slovenian population of Brown Swiss dairy cows was substantial. On average, dairy cows produced 7,337 kg of milk throughout the entire lactation period, with differences between the most and least productive exceeding fourfold (Table 1). The shortest inter-calving interval was 291 days, whereas the longest one was nearly two and a half years. The urea content was, on average, optimal, although certain individuals showed values that were either too low or too high. On average, the cows excreted 105.6 kg of N annually, translating to 17.52 g of Nexi and 314.04 N2O emissions expressed in kilograms of CO2 equivalents. Table 1.Descriptive statistics for phenotype traits of 875 Brown Swiss dairy cows.TraitaMeanStandard deviation Minimum Maximum Milk production in a year (kg)6,206 1,641 2,724 12,539 Milk production in full lactation (kg)7,337 2,343 3,063 16,637 Inter-calving period (day) 436 100 291 912 Milk urea (mg/100 ml)19.77 5.22 6.00 37.58 Nex (kg/year)105.6 18.9 61.0 190.7 Nexi (g/kg milk)17.52 2.59 12.40 28.21 NH3 emissions (kg/year)15.210 2.719 8.778 27.464 N2O direct housing emissions (kg CO2 eq/year)63.34 11.32 36.56 114.37 N2O direct storage emissions (kg CO2 eq/year)200.57 35.85 115.76 362.16 N2O indirect storage emissions (kg CO2 eq/year)50.14 8.96 28.94 93.94 N2O emissions (kg CO2 eq/year)314.04 56.14 181.26 588.39 N2O emission intensity (g CO2 eq/kg milk) 52.107.7036.8796.25 aNex: nitrogen excretion; Nexi: nitrogen excretion intensity; NH3: ammonia; N2O: nitrous oxide; SD: standard deviation. Descriptive statistics for phenotype traits of 875 Brown Swiss dairy cows. aNex: nitrogen excretion; Nexi: nitrogen excretion intensity; NH3: ammonia; N2O: nitrous oxide; SD: standard deviation. Nex and Nexi are dependent on milk production. Nex was 20 kg lower and Nexi higher 5.5 kg per kg of milk than values from a meta-analysis of 86 nitrogen balance trials in ref.31 They found that the average nitrogen content in feces and urine was 186 g/day and 192 g/day. When we recalculate these values and consider the average milk production of 31.4 kg of milk per day, we conclude that Nex in their study was 138 kg over an average year of 365.25 days, which amounts to a Nexi of 12 g/kg milk. Therefore, Nex was 20 kg higher, while Nexi was approximately 5.5 g/kg of milk lower. This is understandable because the animals had a significantly higher daily milk yield, and the excreted nitrogen was distributed over the total milk production. Association analysis and overlap analysis To identify SNPs associated with Nexi we carried out an association study on the cows with calculated Nexi. We performed an association study between Nexi data and 38,437 SNPs. While considering effects of parity, season, age at calving, random additive effect for sire, random effect for herd and residual in the mixed model, we found 1,456 SNPs associated with Nexi (adjusted p-value < 0.01) (Supplementary Table S1). Genotype contrasts were computed to more accurately assess the influence of an individual SNP. We discerned distinctions between heterozygous (AB) and homozygous individuals (AA, coded as 0 for homozygous reference, and BB, coded as 2 for homozygous alternative) by analysing the estimated values of Nexi. We found 803 SNPs with a significant difference between genotypes AB–AA, and 885 SNPs between genotypes BB and AB, while between genotypes BB and AA 749 SNPs were found. There were 294 SNPs with differences between heterozygous and homozygous genotypes. We overlapped the genomics regions of this set of 294 SNPs with genomics regions of protein-coding genes (Figure 1). (A) Bovine chromosomes with 140 SNPs associated with nitrogen excretion intensity (Nexi), overlapping 148 protein-coding genes. (B) Enlargement of chromosome 9 showing an overlap between gene MAN1A1 and two SNPs (rs42786248 and rs42785019) associated with Nexi. Most of the 294 SNPs are intergenic variants (132), followed by SNPs located within introns of protein-coding genes (123), 21 SNPs are located upstream of the gene region and 19 downstream of the gene region. Seven SNPs are located within coding regions of genes, five (rs109623009, rs109816907, rs110710907, rs110862887 and rs43712268) of them are missense variants and two of them are synonymous variants rs110501981, and rs43498404 within genes RTL1, and USHBP1, respectively (Supplementary Table S1). Notably, SNP rs42786248 stands out as the most significant SNP, exhibiting a strong association with Nexi, as evidenced by the lowest adjusted p-value (1.412243e-07). The MAN1A1 gene is involved in the protein N-glycosylation processing phase (mammalian) and is overexpressed in the liver.32,33 The 140 SNPs are distributed across 148 different genes (Supplementary Table S1). The ten most significantly associated SNPs (and their overlapping genes) with Nexi are rs42786248 (MAN1A1), rs42805044 (ADAMTSL1), rs110905653 (ITFG1), rs110905653 (PHKB), rs109746439 (RNF44), rs109040251 (PDE8B), rs42436662 (NEIL3), rs109645596 (RXRA), rs29013388 (NEIL3) and rs42570718 (ENSBTAG00000066084). The SNP rs110905653 overlaps with two genes, namely ITFG1 and PHKB. Moreover, two SNPs rs42436662 and rs29013388 overlap with the NEIL3 gene. NEIL3 was previously associated with renal clear cell carcinoma34,35 and hepatocellular carcinoma35 in humans. The nuclear factor 1 A (NFIA) gene was involved as a transcription factor that plays a role in the development of the liver in humans.36 To obtain a better understanding of the functions of genes associated with Nexi, we performed a functional enrichment analysis. Functional enrichment analysis: biological processes, cellular components, molecular functions and pathways associated with nitrogen excretion intensity The analysis of gene ontology (GO) and biological pathway was performed for 148 genes resulting from the overlap analysis. Our gene sets were compared with the GeneTrail3 GO knowledgebase, which contains 875 biological processes (BP), 124 cellular components (CC), 193 molecular functions (MF) and 23 Reactome Pathways. We identified 31 significant BP, 19 CC, 51 MF and 8 Reactome Pathways involving the set of 148 genes (Supplementary Figure 1). The top 10 gene ontologies within individual GO classes were selected based on the 10 lowest p-values resulting from our functional analysis (Figure 2). Additionally, Figure 3 showcases both significant biological pathways derived from the Reactome database. Ranking of the top 10 biological processes (BP), cellular components (CC) and molecular functions (MF) based on p-value, identified by functional analysis. Gene ontologies in individual sets are ordered by P/FDR from smaller to larger values. Relationship between two GO terms associated with nitrogen term and genes included in these biological processes. Biological processes In total, 97 genes are involved in at least one significant biological process. We identified 31 significant biological processes in which our gene set played a crucial role. The ten most noteworthy biological processes were as follows: developmental process, metabolic process, cellular metabolic process, steroid hormone-mediated signalling pathway, hormone-mediated signalling pathway, phosphorylation, multicellular organismal reproductive process, anatomical structure morphogenesis, nitrogen compound metabolic process, and regulation of nitrogen compound metabolic process. Two significant processes were associated with nitrogen compounds. In addition, the biological process called cellular response to peptide was also involved in response to nitrogen compounds. We found 33 unique genes in processes associated with nitrogen. In the biological process ‘regulation of nitrogen compound metabolic process’, we found 13 genes, while the process called ‘nitrogen compound metabolic process’ included 33 genes. Genes associated with the biological process ‘nitrogen compound metabolic process’ are ABHD4, ADAMTSL1, AK2, AK8, APP, CAMKK2, CARMIL1, CIAO2A, CPQ, DUSP4, EPHA5, EPHB1, ESR1, FGB, GALNT16, GTF2A1L, HIBCH, MAK, MAST4, MCM8, NEIL3, PARP4, PHKB, PLRG1, PROZ, RARA, RBM15, ROR1, SCAI, ST8SIA4, TRIM37, TTLL11 and ULK4. Additionally, the genes APP, CAMKK2, DUSP4, EPHB1, ESR1, FGB, MAK, PARP4, RBM15, ROR1, SCAI, TRIM37 and ULK4 are associated with both nitrogen compound metabolic process and regulation of nitrogen compound metabolic process (Figure 3). Cellular components Nineteen out of 124 cellular components were significant in our gene enrichment analysis with 98 genes from the list of 148 genes. Significant cellular components were intracellular organelle, protein-containing complex, intracellular membrane-bounded organelle, cell projection, cytoplasm, cytosol, membrane, nucleoplasm, nucleus, nuclear chromatin, filopodium, nuclear speck, neuromuscular junction, lamellipodium, transcription factor complex, cytoplasmic vesicle, intracellular vesicle, neuron projection and chromatin. The most significant cellular component called ‘intracellular organelle’ contained 45 genes. We found 29 genes (APP, BPTF, CARMIL1, CIAO2A, CRYBG3, ELP3, ESR1, F2RL2, FGB, IKZF1, KNTC1, LTBP1, MCM8, MRPL47, NR5A2, NR6A1, PBX1, PLRG1, RBM15, ROR1, RUNX2, RXRA, SGIP1, SNX1, SPATS2L, TFDP1, TRIM37, UTRN and ZFYVE9) that are involved in cellular component functional pathway protein-containing complex. Some of these genes have been studied in liver and kidney function. The APP gene is highly expressed in the liver and is involved in liver metabolism.37,38 The ESR1 gene encodes for oestrogen receptor alpha, which is expressed in the liver and has been shown to play a role in liver regeneration.39 The KNTC1 is upregulated in liver cancer.40 The LTBP1 has also been found to be involved in the formation of extracellular matrix, which is important for liver development and function.41 Molecular functions In the realm of molecular functions, we found 89 genes involved in 51 significant molecular functions. The top ten MF molecular functions are cation binding, metal ion binding, organic cyclic compound binding, protein binding, steroid hormone receptor activity, lipid binding, transition metal ion binding, zinc ion binding, DNA binding and sequence-specific DNA binding. The RARA gene is associated with the most molecular functions, with as many as 40, such as cation binding, protein binding, protein kinase binding, signalling receptor binding, transcription coregulator activity, etc. Pathway analysis In the analysis of Reactome pathways, we found eight significant pathways, namely: Glycogen breakdown (glycogenolysis), EPH-Ephrin signalling, Factors involved in megakaryocyte development and platelet production, EPH-ephrin mediated repulsion of cells, Ion transport by P-type ATPases, Nuclear Receptor transcription pathway, Regulation of actin dynamics for phagocytic cup formation, and VEGFA-VEGFR2 Pathway. In glycogenolysis, genes AGL, PGM2 and PYGL are involved, while genes KNTC1 and SCAI play an important role in RHO GTPases Activate Formins biological pathway. Additionally, we conducted an in-depth enrichment analysis of two genes MAN1A1 and NEIL3 (Table 2). Polymorphisms located within those genes were shown as the most significant in our association study. Gene MAN1A1 is involved in six cellular components and four molecular functions, while gene NEIL3 is involved in 11 biological processes, five cellular components and 11 molecular functions. Interestingly, gene NEIL3 is contained in the nitrogen compound metabolic process. Table 2.Gene ontology terms for most promising candidate genes MAN1A1 and NEIL3 detected for Nexi in dairy cows.GeneContained inGene ontology MAN1A1 Intracellular organelle Cellular Components Cytosol Cellular component Intracellular membrane-bounded organelle Integral component of membrane Intrinsic component of membrane Metal ion binding Molecular Functions Molecular function Calcium ion binding Cation binding NEIL3 DNA metabolic process Biological Processes Cellular aromatic compound metabolic process Cellular metabolic process Nitrogen compound metabolic process Response to stimulus Biological process Heterocycle metabolic process Metabolic process Nucleobase-containing compound metabolic process DNA repair Cellular response to DNA damage stimulus Intracellular organelle Cellular Components Nucleoplasm Cellular component Intracellular membrane-bounded organelle Nucleus Transition metal ion binding Molecular Functions Organic cyclic compound binding Metal ion binding Double-stranded DNA binding Catalytic activity Hydrolase activity Nucleic acid binding Molecular function DNA binding Cation binding Zinc ion binding Gene ontology terms for most promising candidate genes MAN1A1 and NEIL3 detected for Nexi in dairy cows. Effect of most significant SNP associated with Nexi on NH3 and N2O emissions The most significant SNP from our association study is rs42786248, located within the intron of the MAN1A1 gene on chromosome BTA9 (Figure 4). We calculated Nex, Nexi, NH3 emissions, N2O emissions and N2O emission intensity separately for cows with each rs42786248 genotype. Cows with genotype CC were significantly less efficient than animals with genotypes AA and AC, as the Nexi in genotype CC was 1.49 g/kg milk higher than in genotype AA (p < 0.001) and 1.03 g/kg milk in genotype AC (p < 0.001), respectively (Table 3). Cows with the genotype AA had a lower Nexi of 0.46 g/kg milk compared to individuals with the AC genotype, with statistical significance (p < 0.001). Cows with the lowest Nexi had the highest amount of milk, Nex and total emissions, but on the other hand, also the lowest N2O emission intensity. Cows with genotype CC had 4.43 g CO2 equivalent per kg milk lower N2O emission intensity than animals with genotype AA (Table 3). Further analysis should focus on confirming the relevance of these candidate genes in dairy cow populations and ensuring their suitability for marker-assisted selection in the context of Nexi-related traits. This should include assessing their functional role, evaluating their association with important phenotypic traits and verifying their consistency across different population subgroups. Effect of the rs42786248 SNP on nitrogen excretion intensity. Reference allele A, alternative allele C. (A) Distribution of raw data for Nexi among genotypes for SNP rs42786248; (B) estimated differences for Nexi between genotypes (AA, AC and CC); (C) Bubble graph illustrating the influence of SNP rs42786248 genotypes (AA, AC and CC) on seven studied traits. Bubble size corresponds to trait value, with larger bubbles indicating higher values. Colours represent desirability: red for undesirability, green for high desirability and blue for moderate desirability. Table 3.Phenotypic characteristics of dairy cows with rs42786248 SNP within MAN1A1 gene. GenotypeAAACCCMilk production in a year (kg)6,3756,1415,628Milk urea (mg/100 ml)19.7619.7120.1Nex (kg/year)107.1104.9101.0Nexi (g/kg milk)17.1917.6518.68NH3 emissions (kg/year)64.2362.9360.58N2O direct housing emissions (kg CO2 eq/year)203.38199.28191.85N2O direct storage emissions (kg CO2 eq/year)50.8549.8247.96N2O indirect storage emissions (kg CO2 eq/year)318.45312.3300.39N2O emissions (g CO2 eq/kg milk)51.1052.4855.55Nex: nitrogen excretion; Nexi: nitrogen excretion intensity; NH3: ammonia; N2O: nitrous oxide. Phenotypic characteristics of dairy cows with rs42786248 SNP within MAN1A1 gene. Nex: nitrogen excretion; Nexi: nitrogen excretion intensity; NH3: ammonia; N2O: nitrous oxide. We identified differences in emitted N2O per kilogram of milk, with a difference of 4.45 g CO2 equivalents between rs42786248 homozygotes (Table 3). Considering these genotypes for two hypothetical farms, where animals on the first farm are exclusively bred cows with genotype AA and on the second farm animals are exclusively bred cows with genotype CC. In the case that both farms annually produce a million litres of milk, the second farm would emit 4.45 tons more CO2 equivalent (106 × 4.45 g CO2 equivalents) emissions. The total milk production in Slovenia in the year 2022 was approximately 625 kilotons of milk,42 which, based on the difference between genotypes, would mean that cows with genotype AA would produce 2781 tons of emissions (which is 4.45 g CO2 eq/kg milk multiplied with 625,000 kg of milk and divided with 1000 to obtain tons) less than those with genotype CC. More research is needed to study the impact of the genetic background of Nexi considering all SNPs that significantly and simultaneously affect both NH3 and N2O emissions via Nexi. We included only indirect emissions of N2O during housing and both indirect and direct emissions of N2O during storage for N2O emission calculation. Milk production also generates other N compound emissions, which were not considered in the calculation, and occur in the same and subsequent stages, such as during fertilisation. Additionally, other greenhouse gas emissions, such as enteric CH4 emissions and manure CH4 emissions, would be included if we assessed all animal-produced emissions. Marshall et al.43 found that selecting animals with breeding values for low urea content in milk is environmentally more efficient. Manzanilla-Pech et al.19 investigated the genetic background of CH4 emissions and, by calculating CH4 production per kilogram of energy-corrected milk, identified regions on the genome associated with the CH4 emissions intensity. Further research is needed to identify genomic regions associated with the emission intensity of both greenhouse gases, namely CH4 and N2O. This is essential to incorporate the selection of animals optimally efficient in terms of emissions of both greenhouse gases into breeding programs for environmentally efficient milk production. Conclusions This study provides evidence of the influence of genetic background on Nexi in dairy cows. The identified SNPs and genes emerge as potential targets for selective breeding programs aimed at reducing Nexi and associated NH3 and N2O emissions. Genotype AA of rs42786248 SNP within the MAN1A1 gene is a favourable genotype for the selection of animals aiming for higher milk production with reduced Nexi. This study contributes to the broader goal of promoting environmental sustainability in milk production, offering a pathway for the sector to address the pressing issue of nitrogen pollution. Supplementary Material SupplementaryTable1.xlsx Supplementary Figures.docx