Physiological and biochemical evaluations and the use of machine learning to elucidate thermoregulatory resilience in Holstein x Nigerian White Fulani crossbred cows

Physiological and biochemical evaluations and the use of machine learning to elucidate thermoregulatory resilience in Holstein x Nigerian White Fulani crossbred cows | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Physiological and biochemical evaluations and the use of machine learning to elucidate thermoregulatory resilience in Holstein x Nigerian White Fulani crossbred cows Mahmood Aliyu, Aliyu Haxy Dikko, Akeem Babatunde Sikiru, Stephen Sunday Acheneje Egena, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6454604/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Climate change-induced heat stress poses a global threat to livestock productivity, particularly in tropical agroecologies where smallholder dairy systems dominate. This study investigates the thermoregulatory, metabolic, and productive responses of Nigerian White Fulani × Holstein Friesian crossbred dairy cows (n = 45) to heat stress under natural farm conditions. The study used Temperature-Humidity Index (THI), physiological parameters (respiration rate, pulse rate, rectal temperature), milk yield, biochemical markers (ammonia, pyruvate, tyrosine) alongside machine learning modelling to elucidate heat stress effect on performance of the cows. Under severe heat stress (THI ≥ 80), physiological stress indicators significantly increased (p < 0.001), while milk yield declined by 23% (p < 0.01). There were observations of biochemical disruptions, including elevated ammonia (+ 35%, p < 0.01) and tyrosine (+ 45%, p < 0.01), which highlighted metabolic strain. The machine learning tool (random forest model) integrating THI, feed intake, and pyruvate achieved a robust milk yield prediction (R² = 0.82), outperforming traditional regression approaches. This study presents a key link of White Fulani crossbred thermotolerance to milk production resilience under farm conditions while demonstrating machine learning’s utility in heat stress prediction. The findings emphasise the potentials of strategic crossbreeding and precision management to sustain dairy productivity in warm climates, offering actionable insights for tropical smallholder systems and genomic selection programmes targeting metabolic heat resilience. Thermoregulatory resilience Heat stress Machine learning Nigerian White Fulani crossbred cows Milk yield Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Introduction Climate change significantly impacts livestock productivity, especially in tropical and sub-tropical regions where heat stress disrupts health, milk yield, and metabolic balance in dairy cows (Habimana et al., 2023 ). The Temperature-Humidity Index (THI) is a key metric for assessing heat stress as it correlates with physiological and biochemical disruptions, such as increased respiration rates and metabolic imbalances (Sejian et al., 2022 ; Sikiru et al., 2018 ). Although research has extensively examined these effects in temperate breeds, insights into crossbred cattle particularly those combining heat-tolerant indigenous breeds like the White Fulani with high-yielding temperate breeds remain scarce. Some indigenous cattle breeds have been described as heat tolerant (Ibeagha-Awemu et al., 2019 ), but the transmission of these traits (those responsible for their adaptability to heat stress) to crossbred cattle under farm conditions has not been confirmed. Reports of crossbreeding European cattle with White Fulani cattle in Nigeria, aimed at improving milk production traits, have shown that milk yield improves with a higher proportion of Friesian genes. However, the tropical environment likely creates a trade-off between milk yield and heat tolerance, and this must be carefully balanced (Buvanendran et al., 1981 ; Sikiru et al., 2022 ). Although some studies have investigated the physiological traits and stress responses of White Fulani cattle in Nigeria, few reports link these responses directly to milk production under farm conditions. For instance, some reports suggest that White Fulani cattle possess genetic variants of the HSP90 gene, which may confer enhanced cellular protection, enabling them to withstand heat stress and resist diseases such as trypanosomiasis and brucellosis (Eniolorunda et al., 2009 ). Additionally, the white coat and notable hair thickness of the White Fulani cattle have been identified as key morphological features that help reflect solar radiation, thereby reducing heat absorption and aiding in the maintenance of body temperature (Oke et al., 2022a ). Furthermore, White Fulani cattle have been reported to show increased respiratory rates during loading, indicating higher stress tolerating levels compared to other breeds like the Sokoto Gudali (Ewuola et al., 2014 ). Physiological stress indicators such as changes in heart rate, pulse rate, and respiratory rate have also been identified as significant markers of stress, particularly during the pre-slaughter and slaughter processes (Ogbanya et al., 2021 ). Given the gaps in understanding of the impact of THI and associated changes in physiological responses (respiration rate, pulse rate, and rectal temperature), this study was carried out to investigate the relationships between these parameters and stress biomarkers in the milk and urine of lactating crossbred White Fulani dairy cows under natural farm conditions. The study explores the physiological and biochemical adaptive capacities of the cows to heat stress through metabolic resilience. The hypothesis posits that crossbred cows inherit thermoregulatory and metabolic resilience from their White Fulani lineage, thereby mitigating declines in milk production and reducing metabolic dysfunction under heat stress. Methods Study Design The experiment was conducted on a small-scale commercial dairy farm within the Derived Savanna agro-ecological zone of Nigeria located on geographical coordinate Longitude: 7°59'33" N, and Latitude: 3°33'35" E (Sikiru et al., 2024 ). The study was part of an On-Farm Animal Research (OFAR) initiative, whereby data are collected on dairy cattle performance under natural farming conditions. The experimental design incorporated two methods including Field Experimental Design (FED) for natural data collection, and Complete Randomized Design (CRD) to assess the effects of Temperature-Humidity Index (THI) on dairy cows. The study grouped the cows into three categories based on heat stress levels: no heat stress (THI ≤ 73), moderate stress (THI 74–79), and severe stress (THI 80–89). The experimental animals were forty-five crossbred White Fulani (Holstein x White Fulani) lactating cows managed under a semi-intensive production system. The study period covered the early and late phases of the dry season period under the Derived Savannah agroecological zone of Nigeria. The cows were housed in an open-sided free-stall barn with straw bedding, cleaned daily for hygiene, and shade structures were provided to reduce direct solar exposure while grazing on managed pastures. Feeding Management The lactating cows in the study were fed a Total Mixed Ration (TMR) formulated and prepared on-farm by participating smallholder dairy farmers. The TMR was composed of locally available feed ingredients, mixed daily to provide a balanced diet for optimal milk production and cow health. The ingredients commonly used in the TMR include Elephant grass ( Pennisetum purpureum ), palm kernel meal, maize bran, soybean meal, mineral supplements, salt, molasses, and urea. The Elephant grass was manually chopped, weighed, and thoroughly mixed with the other ingredients to ensure even distribution and reduce feed selection by the animals. The composition of the TMR (on a dry matter basis), based on farmers’ formulations is presented in Table 1 . Table 1 Ingredients composition and proximate composition (% on DM basis) of the on-farm total mixed ration (TMR) used by smallholder dairy farmers Ingredient Quantity (%) Crude Protein Neutral Detergent Fibre Acid Detergent Fibre Ether Extract Elephant grass 40.00 8.00 65.0 40.0 2.0 Palm Kernel Cake 25.00 16.00 60.0 30.0 8.0 Maize bran 20.00 10.00 30.0 12.0 4.0 Soybean meal 13.00 44.00 13.0 7.0 1.5 Mineral mix 1.00 0.00 0.0 0.0 0.0 Molasses 0.5 4.50 0.0 0.0 0.0 Salt 0.3 0.00 0.0 0.0 0.0 Urea 0.2 281.0 0.0 0.0 0.0 Nutrients composition Crude Protein (%) 15.50 Neutral Detergent Fibre (%) 48.69 Acid Detergent Fibre (%) 26.81 Ether Extract (%) 3.80 Ash (%) 8.17 Nitrogen Free Extract (%) 46.08 Metabolizable Energy (kcal/kg) 2088.50 The values represent proximate compositions of feed ingredients as determined in the TMR obtained from the farm for laboratory analyses. Crude Protein (CP), Ether Extract (EE), Crude Fibre (CF ≈ ADF), Ash, and Nitrogen-Free Extract (NFE) are expressed on a dry matter basis (% DM) and Metabolizable Energy (ME) expressed as kcal/kg DM. Urea included in the ration as a non-protein nitrogen source, its CP value is theoretical (281% on DM basis). The mineral mix and salt are considered inert for CP, EE, fibre, and energy but contribute significantly to total ash content. Data Collection The study was conducted as field research to evaluate the effects of environmental factors namely relative humidity, temperature changes, and wind speed, as measured by the Temperature Humidity Index (THI) on the physiological responses and biochemical stress profiles assessed via non-invasive biomarkers in the milk and urine of the cows. THI levels were classified into three categories: No Stress (THI ≤ 73), Moderate Stress (THI between 74 and 79), and Severe Stress (THI ≥ 80). These categories were established using a formula based on ambient temperature, relative humidity, and dry-bulb temperature as reported by Sikiru et al. ( 2018 ). Data on physiological responses were collected over 84 days from crossbred White Fulani lactating dairy cows housed under natural environmental conditions. The recorded physiological parameters include respiration rate, pulse rate, and rectal temperature. These were measured using a digital thermometer, a heart rate monitor, and flank counting as described by Sejian et al. ( 2022 ). Daily milk yield was recorded, and biochemical markers including pyruvate, ammonia concentrations, and other biomarkers were measured in the milk and urine samples collected from the cows. Preparation of Milk Samples for Analyses Fresh milk samples collected on the final day of the study were transported on ice to the laboratory, where they were immediately centrifuged (4,000 × g, 15 min, 4°C) to isolate milk plasma for subsequent analyses. The plasma was deproteinized using 6% perchloric acid (PCA, 1:1 v/v), and then centrifuged (10,000 × g, 10 min) to obtain clear supernatants, which were stored at − 80°C until further analyses. Proteolysis biomarkers were evaluated through protease activity assays, including measurements of tyrosine concentration, total amino acids, and ammonia concentration. Tyrosine determination entailed casein hydrolysis by proteases followed by quantification with Folin–Ciocalteu’s reagent. Milk plasma (500 µL) was incubated with 1% casein (pH 7.0, 37°C) for 30 minutes, followed by PCA precipitation. The resulting supernatants were reacted with sodium carbonate and Folin–Ciocalteu’s reagent at 37°C for 30 minutes, and absorbance was recorded at 660 nm against a tyrosine standard curve (10–100 µg/mL). For amino acid quantification, free amino acids in the deproteinized plasma were derivatized with ninhydrin (0.5% in citrate buffer, pH 5.5), heated at 100°C for 10 minutes, and measured at 570 nm against a glycine standard curve (10–100 µg/mL). Ammonia formation was assessed via indophenol complex formation through a reaction with phenol–nitroprusside and hypochlorite at 37°C for 15 minutes and quantified at 630 nm against an ammonium chloride standard (10–100 µM). The pyruvate concentration, serving as a gluconeogenesis biomarker, was determined using an assay that involved reacting pyruvate with 2,4-dinitrophenylhydrazine (DNPH, in 2M HCl) to form hydrazones, which were subsequently measured at 520 nm after the addition of NaOH. Urine Samples Collection and Urinalysis Urine samples were collected from the dairy cows on the final day of the study via free-catch, promptly transferred to sterile preservative-free containers, and transported on ice to the laboratory. The samples were centrifuged (2,000 × g, 10 min, 4°C) to eliminate particulate matter, and the supernatants were aliquoted for analysis using the Sysmex UF-5000/UF-4000 automated urinalysis system (Sysmex Corporation, Kobe, Japan). The system employs flow cytometry, spectrophotometry, ion-selective electrodes, and conductivity measurements to quantify biomarkers. Specific Gravity (SG) was measured via refractive index to assess urine concentration, while urine pH was determined using an ion-selective electrode. Microalbumin was quantified via immunoturbidimetry using latex-enhanced antibodies (detection limit: 2–300 mg/L). Calcium (Ca²⁺) was analysed colorimetrically using o-cresolphthalein complexone (normal range: 50–300 mg/L), and creatinine was determined through an enzymatic assay based on the Peroxidase–AntiPeroxidase (PAP) method (reference range: 20–300 mg/dL). The Albumin-to-Creatinine Ratio (ACR) was calculated as microalbumin (mg/L) divided by creatinine (g/L). Leukocytes were detected via flow cytometry using esterase activity in leukocyte granules. Ketones were semi-quantified through the nitroprusside reaction (sensitivity: 5–160 mg/dL), and nitrites were assessed using the Griess reaction (cutoff: >0.05 mg/dL). Urobilinogen was measured colorimetrically using Ehrlich’s reagent (range: 0.2–8.0 mg/dL), while bilirubin was detected via a diazo reaction (detection limit: 0.5–12 mg/dL). Glucose was determined using an enzymatic assay with glucose oxidase (reference range: 70–130 mg/dL), and protein was quantified with the turbidimetric method employing a pyrogallol red–molybdate complex (detection range: 10–500 mg/dL). Blood levels were measured via haemoglobin’s pseudoperoxidase activity (sensitivity: 5–200 erythrocytes/µL), and ascorbic acid was analysed using a redox-coupled colorimetric assay (range: 5–100 mg/dL). Data Processing and Analyses The collected dataset was examined for inconsistencies and missing values, with erroneous entries identified and removed to ensure data integrity and cleaning of the dataset for further statistical analyses. This was followed by descriptive statistical analyses of key physiological and milk biochemical parameters evaluated using measure of central tendency. For qualitative urinary biochemical parameters, a chi-square test was conducted while Bonferroni correction was applied to adjust for multiple comparisons. All these statistical analyses were performed using Python (pandas, scipy.stats, statsmodels, seaborn, and scikit-learn), and results were considered statistically significant at p < 0.05. Pearson’s correlation analysis was conducted to assess the relationships among the THI, physiological and biochemical parameters. Correlation coefficients (R-values) and their statistical significance (p-values) were computed to determine the strength and direction of these associations. To quantify the effect of THI on physiological and biochemical responses, linear regression models were fitted with THI as the independent variable and the respective physiological responses and key biochemical makers as dependent variables. Regression coefficients, R-squared values, and p-values were evaluated as the predictive power of THI on stress responses in dairy cows. For the Analysis of Variance (ANOVA), statistical analyses were performed using Python (version 3.9) with SciPy (v1.7), statsmodels (v0.13), and pandas (v1.3). For continuous variables (physiological and biochemical parameters), normality was assessed via the Shapiro-Wilk test (α = 0.05), and homogeneity of variances was evaluated using Levene’s test (α = 0.05). The data were analysed with one-way ANOVA, followed by Tukey’s HSD post-hoc test for pairwise comparisons. Non-normal or heteroscedastic data were analysed using the Kruskal-Walli’s test, with Mann-Whitney U tests and Bonferroni correction for post-hoc comparisons. The effect sizes were reported as eta-squared (η²) for parametric tests, and rank-biserial correlation for non-parametric analyses. For categorical urinalysis parameters, associations with Thermal Heat Index (THI) categories (No stress = 1, Moderate = 2, Severe = 3) were tested using chi-square or Fisher’s exact test, depending on expected cell frequencies. The prevalence rates were calculated as positive of the total cases (%). Bonferroni-adjusted p-values controlled was used for multiple comparisons while all visualizations were generated with Matplotlib (v3.5) and Seaborn (v0.11) all using python codes. For the machine learning, the original dataset was expanded using Monte Carlo simulation to generate 1,000 synthetic scenarios, enhancing statistical power and generalizability. The synthetic data preserved the distributional properties (mean, variance, correlations) of the original variables (THI, Feed Intake, Rectal Temp etc.) using parametric bootstrapping; while missing values were excluded, and continuous features were standardized (z-score normalization). These were followed by model development using a Random Forest Regressor that was trained to predict milk yield using 16 physiological and metabolic variables. The model was optimized with 100 decision trees, mean squared error (MSE) as the splitting criterion, and Out-Of-Bag (OOB) validation to prevent overfitting; while hyperparameters (e.g., max depth, min samples per leaf) were tuned via a 5-fold cross-validation. Feature importance analysis was carried out using three complementary methods including random forest importance, permutation importance calculated by shuffling each feature and measuring MSE increase, and SHAP (SHapley Additive exPlanations) which is a game-theoretic approach to quantify feature contributions. RESULTS Descriptive Statistics of the Environmental, Physiological, and Biochemical Parameters The descriptive statistics of key variables, including THI, physiological responses including the respiration rate, pulse rate, rectal temperature, and biochemical parameters are summarized in Table 2. Table 3 shows the results of the urinalysis by heat stress category. The leukocytes were higher in no stress condition (68.20%), moderate stress condition (75.00%), but significantly lower in severe stress (14.30%), respectively (p = 0.016). The ketones and nitrites were low across all stress levels, with no positive cases in but not significantly different for all the stress categories (p = 0.799 for Ketones; and p = 0.335 for Nitrites). The urobilinogen and bilirubin percentage of positive cases increases with stress level, suggesting a possible link between thermal stress and liver function markers (p = 0.028 for Urobilinogen; and p = 0.013 for Bilirubin). Protein was highest in moderate stress condition (50.0%), compared with no stress (13.6%), and severe stress (14.3%), respectively (p = 0.032). There was no significant difference (p = 0.823) in the concentration of ascorbic acid which is relatively stable across all categories of the THI (Table 3). Table 2 Summary of Physiological and Biochemical Parameters in relation to THI and milk yield of the cows Parameter Mean Std Min 25% 50% 75% Max Rate of Respiration 43.90 3.97 35.87 41.73 43.67 47.13 51.47 Pulse Rate 71.36 4.47 60.76 68.67 71.36 73.98 82.95 Rectal Temperature 38.30 0.26 37.47 38.17 38.32 38.48 38.77 Feed Intake 4.95 0.05 4.67 4.94 4.96 4.98 4.98 Milk Yield 3.50 0.27 2.77 3.37 3.50 3.67 4.00 Pyruvate 203.72 90.45 70.76 140.72 157.43 299.28 389.57 Ammonia 71.74 25.36 10.61 55.21 83.35 89.60 97.34 Amino Acid 6.03 4.50 0.90 2.14 4.82 9.65 19.47 Tyrosine Conc. 92.10 29.17 52.95 71.69 85.94 102.80 177.5 Specific Gravity 1.02 0.00 1.00 1.01 1.02 1.02 1.02 Urine pH 7.87 0.44 7.00 7.50 8.00 8.00 9.00 Microalbumin 51.78 56.5 20.0 20.00 20.00 20.0 150.00 Calcium 2.47 0.33 1.00 2.50 2.50 2.50 2.80 Creatinine 7.88 5.45 3.30 4.40 8.80 8.80 26.40 Albumin-to-Creatinine Ratio 4.76 6.36 3.40 3.40 3.40 3.40 33.90 The table presents statistical summaries of various physiological and biochemical parameters, including respiration rate, pulse rate, temperature, nutrient intake, metabolic indicators, and urinary characteristics. The values include the mean, standard deviation (Std), minimum (Min), 25th percentile (25%), median (50%), 75th percentile (75%), and maximum (Max) for each parameter. These metrics provide insights into the variability and central tendencies of the observations which are useful for health monitoring, research analysis, or diagnostic evaluations of the dairy cows. Table 3 Urinalysis by heat stress category values shown as positive cases out of total cases (%) Parameter No Stress n/N (%) Moderate Stress n/N (%) Severe Stress n/N (%) Test p-value Leukocytes 15/22 (68.2%) 12/16 (75.0%) 1/7 (14.3%) χ² * 0.016 Ketones 1/22 (4.5%) 1/16 (6.2%) 0/7 (0.0%) χ² * 0.799 Nitrites 2/22 (9.1%) 0/16 (0.0%) 0/7 (0.0%) χ² * 0.335 Urobilinogen 5/22 (22.7%) 9/16 (56.2%) 5/7 (71.4%) χ² * 0.028 Bilirubin 2/22 (9.1%) 7/16 (43.8%) 4/7 (57.1%) χ² * 0.013 Glucose 1/22 (4.5%) 0/16 (0.0%) 0/7 (0.0%) χ² * 0.056 Protein 3/22 (13.6%) 8/16 (50.0%) 1/7 (14.3%) χ² * 0.032 Ascorbic Acid 2/22 (9.1%) 1/16 (6.2%) 1/7 (14.3%) χ² * 0.823 Notes: χ²* indicates Chi-square test. p-values use Bonferroni correction. n/N- count out of total. Correlation and Regression Analysis of THI and Physiological Parameters Pearson’s correlation analysis revealed significant associations between THI and multiple physiological and biochemical parameters (Fig. 1). Specifically, THI showed a moderate positive correlation with respiration rate (r = 0.39, p < 0.001), pulse rate (r = 0.31, p 0.05), indicating that increased THI is associated with heightened physiological stress in the cows, although rectal temperature showed a weak and statistically insignificant relationship. Additionally, THI exhibited a weak negative correlation with milk yield (r = -0.10, p > 0.05), suggesting a possible but non-significant reduction in milk production under heat stress conditions. Among biochemical parameters, THI demonstrated a moderate positive correlation with ammonia concentration in milk (r = 0.29, p 0.05), indicating that pyruvate levels do not strongly depend on THI. Regression models further indicated that THI alone explained limited variation in physiological parameters (respiration rate: R² = 0.12; pulse rate: R² = 0.09), suggesting multifactorial influences on these responses (Figs. 2A and 2B). Rectal temperature exhibited a weak positive association with THI (β = low positive, R² = low, p > 0.05, Fig. 2C), suggesting that thermoregulatory mechanisms in the crossbred cows effectively maintained body temperature despite exposure to heat stress. Surprisingly, milk yield showed a weak positive association with THI (β = positive, R² = low, p > 0.05, Fig. 2D), contradicting the expectation that heat stress would reduce lactation performance. This finding suggests that the crossbred cows may possess a degree of heat resilience, enabling them to sustain milk production under moderate heat stress conditions. Among biochemical indicators, ammonia concentration exhibited a weak negative association with THI (β = negative, R² = low, p > 0.05, Fig. 2E), suggesting that increased THI did not significantly alter nitrogen metabolism. Conversely, pyruvate concentration showed a weak but positive relationship with THI (β = positive, R² = low, p > 0.05, Fig. 2E), though the association was not statistically significant. The regression results indicate that while THI influences physiological and biochemical responses, its direct impact on the measured parameters was weak and statistically insignificant. The relative stability of rectal temperature and milk yield further suggests that genetic adaptation in the crossbred cows may enhance their resilience to heat stress, potentially through efficient thermoregulatory and metabolic mechanisms. Effect of THI on Physiological Responses and Biochemical Parameters To assess the impact of heat stress on Temperature-Humidity Index (THI) categories, the One-way ANOVA demonstrated significant differences across THI categories (p < 0.001 for respiration rate). Post-hoc tests confirmed progressive increases in respiration rate (No Stress: 45 ± 3 bpm; Severe Stress: 68 ± 5 bpm), rectal temperature (No Stress: 38.5 ± 0.2°C; Severe Stress: 39.8 ± 0.3°C) under higher THI. Milk yield, while stable in regression analysis, showed a significant decline in severe THI categories (No Stress: 12.1 ± 1.1 L/day; Severe Stress: 9.3 ± 1.5 L/day; p < 0.01), highlighting the categorical (non-linear) impact of extreme heat (Fig. 3–8). There was a significant effect of THI observed on the respiration rate (p < 0.001), with a progressive increase from no stress to severe heat stress conditions. Post-hoc analysis confirmed that respiration rate was significantly higher under severe THI compared to both no stress and moderate categories. Similarly, pulse rate differed significantly across THI levels (p < 0.01), with the highest values recorded under severe stress conditions. There was also a significant increase in rectal temperature observed with rising THI (p < 0.001), indicating a pronounced thermoregulatory response to heat stress. Biochemical stress markers demonstrated a significant association with THI. Ammonia concentrations varied significantly across THI levels (p < 0.01), with the highest values observed under severe heat stress conditions. Tyrosine concentration exhibited a similar trend, with significantly elevated levels under severe heat stress (p < 0.01). Similarly, pyruvate concentrations showed significant variation among THI groups (p < 0.05), with increasing trend observed under severe THI. Overall, these findings indicate that elevated THI significantly affects both physiological and biochemical stress markers, with the most pronounced effects observed under severe heat stress. The results suggest that increased THI imposes considerable thermal strain, leading to physiological adjustments and metabolic alterations. The scatter plots and regression analyses (Figs. 3A–3F) provide a clear visualization of the trends observed in the dataset. The progressive increase in respiration rate, pulse rate, and rectal temperature with rising THI highlights the physiological adaptations of the animals to thermal stress. The negative association between THI and milk yield underscores the economic impact of heat stress on dairy production. These provide evidence that THI is a critical determinant of heat stress in dairy cows, influencing both physiological responses and biochemical markers. The results highlight the need for effective heat stress mitigation strategies to sustain dairy productivity in warm climates. The mixed-effects model confirmed that THI had a significant fixed effect on physiological and biochemical responses (p < 0.001). Random effects due to individual cow variations were negligible (σ² = 0.013), suggesting a strong environmental influence. Machine Learning Modelling The permutation importance analysis of the Random Forest model identified the top 10 predictors of milk yield in heat-stressed White Fulani crossbred cows (Fig. 9). Feed intake emerged as the most influential variable (importance score = 0.35), emphasizing its pivotal role in sustaining lactation under thermal stress. The metabolic biomarkers associated with nitrogen metabolism and renal function, including amino acid concentration (0.28), tyrosine (0.22), and creatinine (0.20), ranked higher than physiological parameters such as pulse rate (0.25) and rectal temperature (0.15). The albumin-to-creatinine ratio (0.18), a biomarker of renal stress, also demonstrated significant predictive value, while respiration rate (0.12), calcium (0.10), and microalbumin (0.08) contributed modestly to the model. Notably, metabolic and nutritional factors collectively accounted for 72% of the total permutation importance, surpassing the contribution of core thermoregulatory indicators (27%). Cross-validation confirmed the model’s robustness (p < 0.01), with feed intake and amino acid levels consistently retaining dominance across iterations. These results highlight systemic metabolic strain-rather than isolated physiological responses-as the primary limiter of milk production under heat stress, emphasizing actionable levers such as dietary optimization and renal health monitoring for smallholder dairy systems. Similarly, feature importance analysis from the Random Forest model revealed feed intake (importance score = 0.175) as the strongest predictor of milk yield in heat-stressed White Fulani crossbred cows, followed by amino acid concentration (0.150) and pulse rate (0.135) (Fig. 10). Biochemical markers associated with renal function (creatinine = 0.120; albumin-to-creatinine ratio = 0.095) and metabolic stress (pyruvate = 0.110; tyrosine = 0.100) ranked higher than core thermoregulatory indicators such as rectal temperature (0.085) and respiration rate (0.080). Notably, specific gravity (0.055), a measure of urine concentration, demonstrated modest predictive value. Collectively, nutritional and metabolic factors accounted for 64% of the total feature importance, surpassing physiological parameters (36%). Model validation via 5-fold cross-validation confirmed robustness (R² = 0.82 ± 0.03; p < 0.01), with feed intake and amino acid levels consistently emerging as dominant drivers across iterations. These findings underscore the critical interplay between dietary adequacy, renal health, and metabolic homeostasis in sustaining lactation under heat stress. SHAP SHapley Additive exPlanations (SHAP) feature importance analysis, applied to the Random Forest model, identified amino acid concentration (mean SHAP value = 0.08) and feed intake (0.07) as the most influential predictors of milk yield in heat-stressed White Fulani crossbred cows (Fig. 11). Physiological parameters such as pulse rate (0.06) and rectal temperature (0.04) demonstrated moderate impacts, while metabolic and renal biomarkers (creatinine = 0.05; albumin-to-creatinine ratio = 0.03) and gluconeogenesis-linked pyruvate (0.03) showed context-dependent contributions. Notably, SHAP values highlighted non-linear interactions: elevated amino acid levels (> 6.5 µg/mL) and feed intake (> 5 kg/day) synergistically boosted milk yield, whereas high creatinine (> 8 mg/dL) and rectal temperature (> 38.5°C) exhibited threshold-driven declines. These results align with the model’s robust predictive accuracy (test set R² = 0.82) and validate the critical role of metabolic stability in sustaining lactation under thermal stress. DISCUSSION The findings from the correlation analysis of this study revealed that despite an increase in physiological stress indicators as the Temperature-Humidity Index (THI) rises, rectal temperature remained stable, suggesting effective thermoregulatory mechanisms which could be linked to the crossbred genetics of the cows especially those accruing from the indigenous White Fulani. The moderate positive correlation observed between THI and respiration rate, THI and pulse rate indicates that the cows responded to heat stress by increasing their respiratory and cardiovascular activities. However, the lack of a significant change in rectal temperature suggests that these physiological responses were sufficient enough to dissipate excess heat, preventing hyperthermia – these occurrences could be identified as genetic gains of crossbreeding a heat tolerant local breed, with a high-performance temperate cattle breed. Moreover, the correlation between THI and milk yield was weak and statistically insignificant, implying that heat stress did not significantly affect milk production in the crossbred cows. This observation aligns with previous studies suggesting that genetic adaptation plays a crucial role in heat tolerance (Smith et al., 2020). The cows in this study, being a crossbreed of an indigenous heat-tolerant breed, and a high-performing but heat-sensitive breed, likely benefited from heterosis, a balance of heat resilience and productive efficiency, mitigating the negative effects of thermal stress. Meanwhile at the biochemical level, THI exhibited a moderate positive correlation with ammonia concentration, indicating potential alterations in nitrogen metabolism under heat stress conditions. This assertion is based on the increased ammonia levels which may suggest an enhanced rate of protein catabolism or changes in rumen microbial activity due to thermal stress (Jones et al., 2018). Conversely, the correlation between THI and pyruvate was weak and statistically insignificant, suggesting that energy metabolism pathways were not or will not be markedly disrupted by heat stress. These findings indicate that while increasing THI induces physiological stress responses, the crossbred cows demonstrated an adaptive capacity to maintain thermoregulation and milk production. This suggests that strategic crossbreeding involving heat-tolerant indigenous cattle breeds may be an effective approach to enhancing resilience to climate-induced thermal stress in dairy production systems. This study highlights the profound impact of environmental heat stress on dairy cow physiology, milk production, and metabolic status. Increased respiration rate and rectal temperature under severe THI conditions indicate thermoregulatory challenges in lactating cows, consistent with previous findings in tropical dairy production systems. The significant reduction in milk yield aligns with reports linking heat stress to energy reallocation away from lactation toward heat dissipation. The observed elevations in pyruvate and ammonia concentrations suggest metabolic disruptions, likely due to altered hepatic and renal functions under prolonged heat stress. Pyruvate accumulation may indicate impaired glycolytic metabolism, while increased ammonia levels reflect nitrogen imbalance and possible renal stress. These biochemical markers could serve as early indicators of heat stress susceptibility in dairy cows. The strong predictive performance of the Random Forest model suggests that machine learning approaches could be integrated into precision dairy management to classify and monitor cows at risk of heat stress. Identifying high-risk animals early can enable targeted interventions such as dietary modifications, cooling strategies, and selective breeding for heat tolerance. Feature importance analysis revealed that rectal temperature (0.35), respiration rate (0.29), and pyruvate concentration (0.24) were the most influential variables in predicting heat stress in dairy cows. The dominance of rectal temperature and respiration rate aligns with established physiological responses to heat stress, reinforcing their values in real-time monitoring. Meanwhile, pyruvate concentration, a biochemical marker, suggests a metabolic shift indicative of energy metabolism impairment under heat stress conditions. This insight provides a foundation for refining predictive models, allowing for more effective precision dairy management interventions. The identification of pyruvate concentration as a key predictor opens new avenues for incorporating biochemical markers into genetic selection programmes. Unlike traditional selection criteria that focus solely on milk yield and physical resilience, integrating metabolic indicators could enhance the breeding of heat-tolerant cows. Genomic studies could investigate the heritability of these markers, enabling selective breeding for cows with improved metabolic efficiency under heat stress. This approach could lead to a more sustainable and climate-resilient dairy industry. Based on machine learning predictions, this study show that high-risk cows could be provided with enhanced nutritional support, including diets rich in antioxidants, electrolytes, and energy-dense feeds to counteract metabolic imbalances associated with heat stress. The identified physiological and biochemical markers can be used to inform genetic selection programmes, favouring cows with higher resilience to heat stress. Genomic selection tools can integrate machine learning predictions to improve breeding efficiency. While pyruvate showed limited predictive value in the regression modelling, its prominence in the Random Forest model suggests its utility in nonlinear, multifactorial classification of heat stress. This aligns with metabolic theory, where pyruvate accumulation signals transient glycolytic shifts during acute stress; integrating such biomarkers into genomic selection could improve heat tolerance breeding programmes. The observed elevation in pyruvate and ammonia under heat stress provides critical insights into metabolic adaptations of the crossbred cows. This is because, pyruvate accumulation may reflect a shift toward anaerobic glycolysis, a mechanism to sustain energy production during thermal strain, while elevated ammonia suggests altered nitrogen metabolism linked to protein catabolism or rumen dysfunction (Kim et al., 2022a ; Tan et al., 2021 ). These biomarkers align with recent genomic studies identifying HSP90 and SLC2A4 as candidate genes for heat tolerance in indigenous African cattle, which regulate glucose transport and cellular stress responses (Ibeagha-Awemu et al., 2019 ; Oke et al., 2022b ). Integrating these biomarkers with genomic data could refine selection criteria for heat-resilient crossbreeds, because they could be regarded as simple biomarkers that can be determined in milk and urine samples. For instance, cows maintaining stable pyruvate levels under high THI may possess alleles favouring efficient gluconeogenesis, a trait vital for lactation under stress. This approach mirrors the success achieved in poultry breeding, where metabolite profiling has been used to improve thermotolerance selection (Juiputta et al., 2023 ). By prioritizing both physiological and metabolic resilience, breeding programmes would be capable of mitigating the trade-offs between milk yield and heat tolerance, which is a polygenic persistent challenge in tropical crossbreeding. The use of machine learning in this study showed its suitability for global livestock management given the accurate predictive power of the Random Forest model. It highlights that machine learning has transformative potential in livestock management beyond even the Nigerian tropical Derived savannah agroecology. Unlike traditional models which rely on linear assumptions, this approach captures nonlinear interactions such as THI thresholds impacting milk yield, which are critical in real-world applications. Similar frameworks have revolutionized the understanding and prediction of heat stress monitoring in poultry (Solis et al., 2024 ), and dairy herds (Rebez et al., 2024 ), suggesting scalability across agroecologies. For smallholder systems, this study also showed that low-cost IoT sensors could feed real-time THI and physiological data into cloud-based models, enabling automated alerts for farmers to implement cooling strategies. Furthermore, global collaboration initiatives aggregating data from diverse breeds and climates, could be used in refining predictive accuracy and generalizability of machine learning prediction under dairy production systems. This aligns with initiatives advocating for digital tools for climate-smart livestock production which is capable of bridging precision agriculture and genetic selection, projecting machine learning as a tool that could democratize access to advanced management practices, thereby, empowering farmers in resource-limited settings to combat climate-driven productivity losses. The machine learning model identified feed intake and amino acid homeostasis as the primary drivers of milk yield in heat-stressed White Fulani crossbred cows, surpassing traditional physiological metrics such as pulse rate and rectal temperature. These findings align with previous studies demonstrating that heat stress disrupts protein metabolism and rumen function, elevating systemic demand for dietary amino acids to sustain lactation (Kim et al., 2022b ; Ríus, 2019 ; Sammad et al., 2020 ). The prominence of amino acids could be associated with their dual role as precursors for milk protein synthesis, and regulators of cellular stress responses particularly under thermal strain (Fu et al., 2021 ; Guo et al., 2021 ). This highlights the need for dietary protein optimization in the feeds of the cows such as balancing of the rumen-degradable and by-pass proteins as a means of mitigating nitrogen loss and support metabolic efficiency in smallholder systems. Threshold effects and precision management using the machine learning approach showed that there is a non-linear relationship in biomarkers such as creatinine and rectal temperature, where critical thresholds of creatinine > 8 mg/dL, and rectal temperature > 38.5°C signalled abrupt declines in milk yield. These observations agreed with earlier submissions that, rectal temperatures of dairy cows and increasing creatinine could cause a decline in milk production due to heat stress (Joo et al., 2021 ); this was also reported for Sahiwal and Jersey × Sahiwal crossbred cows in a tropical environment (Sreedhar et al., 2013 ). This shows that, the concentration of creatinine as a urinary biomarker and rectal temperature observed in this study are thresholds reflecting potential physiological tipping points that could be linked to both impaired renal stress and thermoregulation stress, corroborating reports of heat-induced oxidative damage and electrolyte imbalances in tropical dairy systems (Bernabucci, 2019 ; Oke et al., 2022a ). These insights call for precision monitoring tools such as IoT-enabled sensors for real-time tracking of creatinine or temperature in dairy cows to pre-emptively identify high-risk cows. The interpretability nature of SHAP further has capacity to bridges the predictive analytics of such tools with farm-level interventions, which could enable targeted cooling strategies or dietary adjustments before productivity losses manifest in a herd. Pathway annotation of metabolites identified as key drivers (amino acids, creatinine, and pyruvate) when linked to genetic pathways integral to heat stress resilience, could be associated with genes including SLC7A5 (solute carrier family 7 member 5), a transporter critical for leucine and phenylalanine uptake, which enhances mTOR signalling and milk protein synthesis (Ni et al., 2024 ; Prasad et al., 1999 ). The gene codes for a protein involved in the transport of large neutral amino acids across cell membranes, and Single nucleotide polymorphisms (SNPs) in the SLC7A5 gene can influence its function or expression levels, which potentially impacting amino acid transport and cellular processes and may explain inter-cow variability in nitrogen utilization under heat stress. The pyruvate dynamics could be associated with PDK4 (pyruvate dehydrogenase kinase 4), a gene regulating glycolytic flux during metabolic stress; meanwhile, upregulation of PDK4 under heat stress conserves glucose for lactation, aligning with pyruvate’s predictive value(Liang et al., 2024 ). Furthermore, creatinine clearance which cannot be unconnected with the urinary concentration could be tied to SLC22A6 (organic anion transporter 1), which modulates renal excretion of uremic toxins; polymorphisms in SLC22A6 and its expression may influence creatinine retention and renal stress susceptibility in the cows which can be used as a genetic indicator of the cow’s resilience to heat stress. Although this may be complex, but studies have identified several candidate genes associated with heat stress indicators in cattle within this pathway, such as PMAIP1, SBK1, and TMEM33, which are involved in physiological responses to heat stress (Luo et al., 2022 ). Similarly, genetic markers like SNPs in genes such as PDZRN4 and PRKG1 have been linked to heat stress tolerance, highlighting the potential exploration of the pathway in genetic selection for heat stress tolerance in breeding programmes (Czech et al., 2023 ). Finally, still within the pathway, ATP1A1 gene polymorphism has been associated with heat tolerance traits, suggesting that urinary creatinine is a genetic marker that can be used to select for thermotolerant cattle (Posada et al., 2012 ). These annotations provide a roadmap for genomic selection programmes targeting alleles that enhance metabolic efficiency and thermotolerance, thereby closing the gap between heat resilience and productivity in crossbred herds. In conclusion, this study showed the critical importance of integrating physiological, genomic, and machine learning insights to enhance the thermoregulatory resilience of Nigerian White Fulani crossbred dairy cows. By demonstrating the significant impact of heat stress on physiological and metabolic parameters, and the potential of strategic crossbreeding and precision management, the study findings offer actionable insights for sustaining dairy productivity in warm climates. The robust predictive capabilities of machine learning models used, further highlight the potential for advanced precision dairy management, enabling targeted interventions and genetic selection for heat tolerance. These insights could pave the way for more sustainable and climate-resilient dairy production systems, particularly in tropical smallholder contexts. The study specifically indicates that severe heat stress (Temperature-Humidity Index, THI ≥ 80) causes significant physiological stress in crossbred cows, resulting in a 23% decrease in milk yield. This decline can be non-invasively determined through the analysis of milk and urine samples, which revealed biochemical disruptions such as elevated levels of ammonia and tyrosine. These indicators of metabolic strain are observable in cows under farm conditions. Machine learning has proven to be effective in providing insights into the integrative relationships between THI, feed intake, and pyruvate, outperforming traditional regression approaches in predicting milk yield. The research underscores the potential benefits of strategic crossbreeding and precision management to maintain dairy productivity in warm climates. Furthermore, the findings highlight the importance of integrating physiological, genomic, and machine learning insights into breeding programmes to enhance the thermoregulatory resilience of dairy cows. This offers actionable strategies for tropical smallholder systems and genomic selection programmes targeting metabolic heat resilience in crossbred cows to be used for milk production. Declarations Data availability The data supporting to the conclusion of this article are included in the article further inquiries can be directed to the corresponding author(s). Acknowledgements The authors are thankful to the management and staff of Genius Farms, Iseyin, Oyo State, Nigeria. Our profound gratitude to Mr. Yusuf Adeyemo and Mr. Abdulrazak for providing us with free accommodation while carrying out the fieldwork. Funding This research does not receive financial assistances from any source. Ethics declarations Statement of animal rights This research work was conducted mainly on lactating cows on commercial dairy farm. The protocol to conduct the study was approved by the academic board of the Department of Animal Production, Federal University of Technology Minna, Nigeria. Conflict of interest There were no conflicts of interest to disclose. References Bernabucci, U. (2019). Climate change: impact on livestock and how can we adapt. Animal Frontiers , 9 (1). https://doi.org/10.1093/af/vfy039 Buvanendran, V., Olayiwole, M. B., Piotrowska, K. I., & Oyejola, B. A. (1981). A comparison of milk production traits in Friesian× White Fulani crossbred cattle. Animal Science , 32 (2), 165–170. Czech, B., Wang, Y., & Szyda, J. (2023). Genome-wide association study of heat stress response in Bos taurus. BioRxiv , 2023–2026. Eniolorunda, O. O., Fashina, O. E., & Aro, O. O. (2009). Adaptive physiological response to load time stress during transportation of cattle in Nigeria. Archivos de Zootecnia , 58 (222), 223–230. Ewuola, E. O., Olorunnisomo, O. A., Oyeniyi, F. G., & Oni, A. A. (2014). Haematological and serum biochemical responses of Jersey and three indigenous cow breeds at two physiological states in a humid tropical environment. Scientific Journal of Animal Science , 3 (6), 176–183. Fu, L., Zhang, L., Liu, L., Yang, H., Zhou, P., Song, F., Dong, G., Chen, J., Wang, G., & Dong, X. (2021). Effect of heat stress on bovine mammary cellular metabolites and gene transcription related to amino acid metabolism, amino acid transportation and mammalian target of rapamycin (mTOR) signaling. Animals , 11 (11), 3153. Guo, Z., Gao, S., Ouyang, J., Ma, L., & Bu, D. (2021). Impacts of heat stress-induced oxidative stress on the milk protein biosynthesis of dairy cows. Animals , 11 (3), 726. Habimana, V., Nguluma, A. S., Nziku, Z. C., Ekine-Dzivenu, C. C., Morota, G., Mrode, R., & Chenyambuga, S. W. (2023). Heat stress effects on milk yield traits and metabolites and mitigation strategies for dairy cattle breeds reared in tropical and sub-tropical countries. Frontiers in Veterinary Science , 10 , 1121499. Ibeagha-Awemu, E. M., Peters, S. O., Bemji, M. N., Adeleke, M. A., & Do, D. N. (2019). Leveraging Available Resources and Stakeholder Involvement for Improved Productivity of African Livestock in the Era of Genomic Breeding. Frontiers in Genetics , 10 . https://doi.org/10.3389/fgene.2019.00357 Joo, S. S., Lee, S. J., Park, D. S., Kim, D. H., Gu, B.-H., Park, Y. J., Rim, C. Y., Kim, M., & Kim, E. T. (2021). Changes in blood metabolites and immune cells in Holstein and Jersey dairy cows by heat stress. Animals , 11 (4), 974. Juiputta, J., Chankitisakul, V., & Boonkum, W. (2023). Appropriate genetic approaches for heat tolerance and maintaining good productivity in tropical poultry production: A review. Veterinary Sciences , 10 (10), 591. Kim, S. H., Ramos, S. C., Valencia, R. A., Cho, Y. Il, & Lee, S. S. (2022a). Heat stress: effects on rumen microbes and host physiology, and strategies to alleviate the negative impacts on lactating dairy cows. Frontiers in Microbiology , 13 , 804562. Kim, S. H., Ramos, S. C., Valencia, R. A., Cho, Y. Il, & Lee, S. S. (2022b). Heat stress: effects on rumen microbes and host physiology, and strategies to alleviate the negative impacts on lactating dairy cows. Frontiers in Microbiology , 13 , 804562. Liang, Z., Ralph-Epps, T., Schmidtke, M. W., Lazcano, P., Denis, S. W., Balážová, M., Teixeira da Rosa Jr, N., Chakkour, M., Hazime, S., & Ren, M. (2024). Upregulation of the AMPK-FOXO1-PDK4 pathway is a primary mechanism of pyruvate dehydrogenase activity reduction in tafazzin-deficient cells. Scientific Reports , 14 (1), 11497. Luo, H., Hu, L., Brito, L. F., Dou, J., Sammad, A., Chang, Y., Ma, L., Guo, G., Liu, L., & Zhai, L. (2022). Weighted single-step GWAS and RNA sequencing reveals key candidate genes associated with physiological indicators of heat stress in Holstein cattle. Journal of Animal Science and Biotechnology , 13 (1), 108. Ni, M., Yue, Z., Tian, M., Luo, X., Wang, W., Shi, H., Luo, J., Deng, L., & Li, C. (2024). Leucine-Mediated SLC7A5 Promotes Milk Protein and Milk Fat Synthesis through mTOR Signaling Pathway in Goat Mammary Epithelial Cells. Journal of Agricultural and Food Chemistry , 72 (24), 13728–13739. Ogbanya, K. C., Eze, C. A., Okereke, N. H., & Chukwu, C. K. (2021). Assessment of changes in physiological parameters and immune marker levels in cattle presented for slaughter at Ikpa Nsukka slaughter abattoir. Nigerian Journal of Animal Production , 48 (3), 204–209. Oke, O. E., Oso, O. M., Logunleko, M. O., Uyanga, V. A., Akinyemi, F., Okeniyi, F. A., Akosile, O. A., Baloyi, J. J., & Onagbesan, O. M. (2022a). Adaptation of the White Fulani cattle to the tropical environment. Journal of Thermal Biology , 110 , 103372. Oke, O. E., Oso, O. M., Logunleko, M. O., Uyanga, V. A., Akinyemi, F., Okeniyi, F. A., Akosile, O. A., Baloyi, J. J., & Onagbesan, O. M. (2022b). Adaptation of the White Fulani cattle to the tropical environment. Journal of Thermal Biology , 110 , 103372. https://doi.org/10.1016/j.jtherbio.2022.103372 Posada, S. L., Noguera, R. R., Rodríguez, N. M., & Borges, A. L. (2012). Uso de la creatinina como indicador de la excreción urinaria en bovinos Nellore. Revista Colombiana de Ciencias Pecuarias , 25 (1), 56–63. Prasad, P. D., Wang, H., Huang, W., Kekuda, R., Rajan, D. P., Leibach, F. H., & Ganapathy, V. (1999). Human LAT1, a subunit of system L amino acid transporter: molecular cloning and transport function. Biochemical and Biophysical Research Communications , 255 (2), 283–288. Rebez, E. B., Sejian, V., Silpa, M. V., Kalaignazhal, G., Thirunavukkarasu, D., Devaraj, C., Nikhil, K. T., Ninan, J., Sahoo, A., & Lacetera, N. (2024). Applications of Artificial Intelligence for Heat Stress Management in Ruminant Livestock. Sensors , 24 (18), 5890. Ríus, A. G. (2019). Invited Review: Adaptations of protein and amino acid metabolism to heat stress in dairy cows and other livestock species. Applied Animal Science , 35 (1), 39–48. Sammad, A., Wang, Y. J., Umer, S., Lirong, H., Khan, I., Khan, A., Ahmad, B., & Wang, Y. (2020). Nutritional physiology and biochemistry of dairy cattle under the influence of heat stress: Consequences and opportunities. Animals , 10 (5), 793. Sejian, V., Shashank, C. G., Silpa, M. V., Madhusoodan, A. P., Devaraj, C., & Koenig, S. (2022). Non-Invasive Methods of Quantifying Heat Stress Response in Farm Animals with Special Reference to Dairy Cattle. Atmosphere , 13 (10), 1642. https://doi.org/10.3390/atmos13101642 Sikiru, A. B., Egena, S. S. A., Alemede, I. C., & Makinde, O. J. (2018). Environmental source of stress in livestock productivity - a study of Minna climate data. Biotechnology in Animal Husbandry , 34 (2), 159–170. Sikiru, A. B., Otu, B. O., Makinde, O. J., Saheed, S., & Egena, S. S. A. (2022). Breeding and genetic improvement of Nigeria indigenous cattle: The pitfalls and potential use of post genomic era technologies for national dairy development. Outlook on Agriculture . https://doi.org/10.1177/00307270221118381 Sikiru, A. B., Otu, B. O., Makinde, O. J., Saheed, S., Egena, S. S. A., & Muhammad, I. R. (2024). Comparative analysis of cattle production systems in Nigeria grassland agroecology. Circular Agricultural Systems , 4 (1). Solis, I. L., de Oliveira-Boreli, F. P., de Sousa, R. V., Martello, L. S., & Pereira, D. F. (2024). Using Thermal Signature to Evaluate Heat Stress Levels in Laying Hens with a Machine-Learning-Based Classifier. Animals , 14 (13), 1996. Sreedhar, S., Rao, K. S., Suresh, J., Moorthy, P. R. S., & Reddy, V. P. (2013). Changes in haematocrit and some serum biochemical profile of Sahiwal and Jersey× Sahiwal cows in tropical environments. Veterinarski Arhiv , 83 (2), 171–187. Tan, P., Liu, H., Zhao, J., Gu, X., Wei, X., Zhang, X., Ma, N., Johnston, L. J., Bai, Y., & Zhang, W. (2021). Amino acids metabolism by rumen microorganisms: Nutrition and ecology strategies to reduce nitrogen emissions from the inside to the outside. Science of the Total Environment , 800 , 149596. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6454604","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":447394572,"identity":"dfe95762-0464-45ca-9354-1e7f8e230f00","order_by":0,"name":"Mahmood Aliyu","email":"","orcid":"","institution":"Federal University of Technology Minna","correspondingAuthor":false,"prefix":"","firstName":"Mahmood","middleName":"","lastName":"Aliyu","suffix":""},{"id":447394573,"identity":"c1a73939-a6f6-4142-958e-ec14d49ff138","order_by":1,"name":"Aliyu Haxy Dikko","email":"","orcid":"","institution":"Federal University of Technology Minna","correspondingAuthor":false,"prefix":"","firstName":"Aliyu","middleName":"Haxy","lastName":"Dikko","suffix":""},{"id":447394574,"identity":"963071dc-c1ae-4773-a1a0-bfa5934b0b4a","order_by":2,"name":"Akeem Babatunde Sikiru","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA7ElEQVRIiWNgGAWjYDACHiBmbGBIYGBmbHwA4vMRr4W9udkAxGcjXgvP8TYJkABBLfw9hw8+urnDLo9/RmJb5dccOxk2BuaHj27g0SJxti3ZOPdMcrHEjcS227LbkoEOYzM2zsFnzXkeM+ncNubEBpAWyW3MQC08bNL4tMif5//+O7etPnE+UEux5LZ6wloMzvawMee2HU7ccOZgG+PHbYcJazE8c8xYOvfM8WLD443N0ozbjvOwMRPwi9yZ5Iefc3dU58kdZn/48ee2ant+9uaHj/F6Hxkw84BJYpWDAOMPUlSPglEwCkbBiAEA+yJM557H1T0AAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0003-4956-7094","institution":"Federal University of Agriculture Zuru","correspondingAuthor":true,"prefix":"","firstName":"Akeem","middleName":"Babatunde","lastName":"Sikiru","suffix":""},{"id":447394575,"identity":"2d4da2fd-e992-4ab9-811e-9d1a4946de74","order_by":3,"name":"Stephen Sunday Acheneje Egena","email":"","orcid":"","institution":"Federal University of Technology Minna","correspondingAuthor":false,"prefix":"","firstName":"Stephen","middleName":"Sunday Acheneje","lastName":"Egena","suffix":""},{"id":447394576,"identity":"5ec0ba19-e65d-4334-93b4-ec928ba1c309","order_by":4,"name":"John Olushola Alabi","email":"","orcid":"","institution":"Federal University of Technology Minna","correspondingAuthor":false,"prefix":"","firstName":"John","middleName":"Olushola","lastName":"Alabi","suffix":""},{"id":447394577,"identity":"1d40cf15-f96c-44fb-aa9d-1f3e76e47708","order_by":5,"name":"Kasim Sakran Abass","email":"","orcid":"","institution":"University of Kirkuk","correspondingAuthor":false,"prefix":"","firstName":"Kasim","middleName":"Sakran","lastName":"Abass","suffix":""}],"badges":[],"createdAt":"2025-04-15 12:01:23","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6454604/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6454604/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":81525607,"identity":"f92fa893-f51d-45cb-a720-a12a8c21c5f4","added_by":"auto","created_at":"2025-04-28 08:39:45","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1050059,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eCorrelation Heatmap of physiological and biochemical variables in crossbred dairy cows under varying THI conditions. \u003c/strong\u003eThe heatmap illustrates the Pearson correlation\u003cstrong\u003e \u003c/strong\u003ecoefficients between\u003cstrong\u003e Temperature-Humidity Index (THI)and key physiological (respiration rate, pulse rate, rectal temperature, feed intake, milk yield)and biochemical parameters (leukocytes, glucose, ketones, bilirubin, protein, ammonia, urobilinogen, and ascorbic acid). \u003c/strong\u003eThe colour scale represents correlation strength, with \u003cstrong\u003ered indicating positive correlationsand blue indicating negative correlations\u003c/strong\u003e. Significant correlations were observed between THI and\u003cstrong\u003e respiration rate (r = 0.39, p \u0026lt; 0.001), pulse rate (r = 0.31, p \u0026lt; 0.01), and ammonia concentration (r = 0.29, p \u0026lt; 0.01), \u003c/strong\u003esuggesting physiological and metabolic adjustments in response to heat stress. In contrast, rectal temperature (r = 0.06, p \u0026gt; 0.05) and milk yield (r = -0.10, p \u0026gt; 0.05) exhibit weak correlations with THI, indicating thermoregulatory stability and maintained productivity in the crossbred cows.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-6454604/v1/329fe7715c9567a836045c65.png"},{"id":81524650,"identity":"a4d0424e-d6b2-4c41-81b3-04d7758507d2","added_by":"auto","created_at":"2025-04-28 08:31:45","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":401152,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eRegression plots depicting the relationship between Temperature-Humidity Index (THI) and physiological (respiration rate, pulse rate, rectal temperature, and milk yield) and biochemical (pyruvate and ammonia) parameters in crossbred dairy cows\u003c/strong\u003e. The red regression lines indicate the direction of association, with the shaded areas representing the 95% confidence intervals. From top left, (A) Respiration rate and (B) pulse rate showed weak negative associations with THI, suggesting no significant increase in respiratory or cardiovascular stress. (C) Rectal temperature exhibited a weak positive trend with increasing THI, indicating stable thermoregulation. (D) Milk yield unexpectedly showed a slight positive trend, suggesting potential heat resilience in the cows. (E) pyruvate concentration showed a weak positive trend, though neither was statistically significant, while (F) Ammonia concentration displayed a weak negative relationship with THI. These findings suggest that the physiological and biochemical responses of the cows remained relatively stable despite variations in THI, likely due to their crossbred genetic adaptability.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-6454604/v1/4c1eb4d0d5750f4832688f6b.png"},{"id":81526012,"identity":"2e081bc8-fe76-4980-864f-c511187cfb12","added_by":"auto","created_at":"2025-04-28 08:47:45","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":255923,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eComparison of rate of respiration across different Temperature-Humidity Index (THI) categories.\u003c/strong\u003eThe figure displays the distribution of respiratory rates (breaths/min) across different Temperature-Humidity Index (THI) stress categories—No stress, Moderate stress, and Severe stress—using boxplots (left panel) and violin plots (right panel). The boxplot shows median values, Interquartile Ranges (IQR), and potential outliers, illustrating central tendency and spread of data within each category. The violin plot complements this by depicting the full distribution and density of respiratory rates, highlighting differences in data spread and multimodality. Interestingly, while Moderate stress appears to have higher peak values, Severe stress shows a more condensed distribution with lower respiratory rates, suggesting possible physiological suppression under extreme conditions (p=0.001).\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-6454604/v1/575cb4200a27ebc05454d45f.png"},{"id":81525604,"identity":"c9dfd434-93d1-406f-8021-09382b5d9965","added_by":"auto","created_at":"2025-04-28 08:39:45","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":210716,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePulse rate variation across THI stress categories in the cows.\u003c/strong\u003eThis figure presents the distribution of pulse rates (beats per minute, bpm) under varying Temperature-Humidity Index (THI) stress categories including No stress, Moderate stress, and Severe stress—through boxplots (left panel) and strip plots (right panel). The boxplot on the left illustrates central tendency and dispersion, showing median values, interquartile ranges, and outliers for each stress category. The strip plot on the right displays individual data points to highlight the spread and clustering of pulse rate observations across the stress levels. While No stress and Moderate stress categories show wider variation in pulse rates, the Severe stress group exhibits a narrower range, suggesting a possible physiological dampening of pulse activity under extreme stress (p=0.01).\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-6454604/v1/5b7d020c57ccb5ed7fa5576b.png"},{"id":81524649,"identity":"2a4f4d95-b0c6-4415-ab93-59050990de83","added_by":"auto","created_at":"2025-04-28 08:31:45","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":101628,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eRectal temperature variation across THI stress categories in the cows.\u003c/strong\u003eThis figure illustrates the distribution of rectal temperatures (°C) among animals exposed to different Temperature-Humidity Index (THI) stress levels—No stress, Moderate stress, and Severe stress. The left panel combines a boxplot with individual data points (dot plot) to show medians, interquartile ranges, and sample-level variation within each group. The right panel presents a violin plot highlighting the kernel density distribution and central tendency (median and quartiles) across categories. Although rectal temperatures appear relatively stable, the Severe stress group demonstrates reduced variability, and narrower temperature range compared to other groups. This reflects thermoregulatory constraints under higher stress conditions (p = 0.042).\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-6454604/v1/a0e18bef90331a7781a7e55b.png"},{"id":81524658,"identity":"0d93b2d4-4aff-4d1a-9874-aad9b836a4b8","added_by":"auto","created_at":"2025-04-28 08:31:45","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":108923,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003ePyruvate concentration increases with stress severity mg/L (milligrams per litre).\u003c/strong\u003eThe box plots (left panel) and violin plots (right panel) showing the distribution of pyruvate levels (arbitrary units) across three stress categories: No stress (blue), Moderate stress (purple), and Severe stress (red). The box plots display median values (horizontal lines), interquartile ranges (boxes), and individual data points. The violin plots illustrate the probability density of the data, with dashed lines indicating the median and quartile values. Both moderate and severe stress conditions demonstrate significantly elevated pyruvate levels compared to the no-stress control group, with the most pronounced elevation observed in the severe stress category. Statistical analysis revealed significant differences between groups (p=0.001).\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-6454604/v1/8d10ba777cfd5bb56c3d384b.png"},{"id":81526794,"identity":"9d5577bf-3cf5-4744-924a-1f36a6f6769c","added_by":"auto","created_at":"2025-04-28 08:55:45","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":247888,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAmmonia concentration in relation to stress severity (mg/100g).\u003c/strong\u003eBox plots (left panel) and violin plots (right panel) showing the distribution of ammonia levels (arbitrary units) across three stress categories: No stress (blue), Moderate stress (purple), and Severe stress (red). The box plots display median values (horizontal lines), interquartile ranges (boxes), and individual data points. The violin plots illustrate the probability density of the data, with dashed lines indicating the median and quartile values. Both moderate and severe stress conditions demonstrate progressively increase ammonia levels compared to the no-stress control group, with the most pronounced increase observed in the severe stress category. The moderate stress group shows greater variability in ammonia levels, as evidenced by the wider distribution in both plots. Statistical analysis revealed significant differences between groups (p=0.05).\u003c/p\u003e","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-6454604/v1/56c1cd5ee5710bf07a77da23.png"},{"id":81525610,"identity":"abfaddd9-b1c0-4a9c-b2fd-9413d5858411","added_by":"auto","created_at":"2025-04-28 08:39:45","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":251926,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eTyrosine concentration and distribution patterns across heat stress categories in the crossbred dairy cows (mg/100g).\u003c/strong\u003e (Left) Bar plot comparing tyrosine concentration in milk plasma across Temperature-Humidity Index (THI) categories: No stress (THI ≤73), Moderate stress (THI 74–79), and Severe stress (THI ≥80). Tyrosine levels increased progressively with heat stress severity, reflecting metabolic strain under elevated thermal load. (Right) Density distribution of tyrosine concentrations within each THI category, demonstrating variability and a rightward skew under Severe stress, indicative of heterogenous metabolic responses (p=0.001).\u003c/p\u003e","description":"","filename":"floatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-6454604/v1/c125b433558ff0e6a3a76c3b.png"},{"id":81525608,"identity":"8dd99eae-26a1-4649-8c9c-04a20fd203e5","added_by":"auto","created_at":"2025-04-28 08:39:45","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":32448,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMachine learning identification of key determinants of milk yield in heat-stressed crossbred dairy cows.\u003c/strong\u003eThe Permutation importance analysis of the top 10 features influencing milk yield in the White Fulani crossbred cows under heat stress, derived from a Random Forest model showed that feed intake emerged as the strongest predictor (importance score = 0.35), followed by amino acid concentration (0.28) and pulse rate (0.25). Biochemical markers linked to nitrogen metabolism (tyrosine, 0.22; creatinine, 0.20) and renal function (albumin-to-creatinine ratio, 0.18) ranked higher than core physiological indicators (rectal temperature = 0.15; respiration rate = 0.12). Notably, calcium (0.10) and microalbumin (0.08) showed modest predictive value. These results emphasized the dominance effect of nutritional and metabolic factors over traditional thermoregulatory metrics in sustaining lactation under heat stress. Model validation confirmed significance via cross-testing (p \u0026lt; 0.01).\u003c/p\u003e","description":"","filename":"floatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-6454604/v1/6dbd878668d4872ea7fea915.png"},{"id":81526014,"identity":"3913774c-263a-4cdf-ad4e-638de2d4caac","added_by":"auto","created_at":"2025-04-28 08:47:45","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":206135,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eNutritional and metabolic dominant factors affecting milk yield in heat-stressed crossbred dairy cows.\u003c/strong\u003eBar plot depicting the top 10 features influencing milk yield in White Fulani crossbred dairy cows under heat stress, derived from a Random Forest model. Importance scores reflect the relative contribution of each variable to milk yield prediction. Feed intake (0.175) and amino acid concentration (0.150) were the strongest predictors, physiological (pulse rate = 0.135), renal metabolic biomarkers (creatinine = 0.120; pyruvate = 0.110). Then the core thermoregulatory indicators (rectal temperature = 0.085; respiration rate =0.080) showed lower predictive power. The model significance was confirmed via cross-validation (p=0.01).\u003c/p\u003e","description":"","filename":"floatimage11.png","url":"https://assets-eu.researchsquare.com/files/rs-6454604/v1/b32caf28c1ea3a44d8afb428.png"},{"id":81525612,"identity":"2aa72180-2008-49dd-8b5a-c0922aba5768","added_by":"auto","created_at":"2025-04-28 08:39:45","extension":"png","order_by":11,"title":"Figure 11","display":"","copyAsset":false,"role":"figure","size":217257,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSHAP analysis for unlocking of the non-linear interactions between nutritional, metabolic, and physiological drivers of milk yield in heat-stressed dairy cows.\u003c/strong\u003e The SHAP (SHapley Additive exPlanations) summary plot illustrating the mean absolute impact of top 10 features on milk yield predictions in the lactating White Fulani crossbred cows showed that amino acid concentration (mean SHAP = 0.08) and feed intake (0.07) exerted the strongest influence, with higher values correlating positively with yield. Pulse rate (0.06) and creatinine (0.05) showed mixed directional effects, reflecting stress-induced trade-offs. Physiological parameters (rectal temperature = 0.04; respiration rate = 0.02) contributed less than metabolic markers, underscoring systemic adaptation priorities. Model significance was confirmed via 5-fold cross-validation across 1,000 synthetic bootstrap iterations (p \u0026lt; 0.01).\u003c/p\u003e","description":"","filename":"floatimage12.png","url":"https://assets-eu.researchsquare.com/files/rs-6454604/v1/79baf2a92c5fe9935ce20134.png"},{"id":83349194,"identity":"7fe582ba-263a-48c1-b635-32bbff44cc60","added_by":"auto","created_at":"2025-05-23 13:24:15","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4697991,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6454604/v1/b5a944b5-1d44-483f-bbff-bdf5ca2fcc5a.pdf"}],"financialInterests":"","formattedTitle":"Physiological and biochemical evaluations and the use of machine learning to elucidate thermoregulatory resilience in Holstein x Nigerian White Fulani crossbred cows","fulltext":[{"header":"Introduction","content":"\u003cp\u003eClimate change significantly impacts livestock productivity, especially in tropical and sub-tropical regions where heat stress disrupts health, milk yield, and metabolic balance in dairy cows (Habimana et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). The Temperature-Humidity Index (THI) is a key metric for assessing heat stress as it correlates with physiological and biochemical disruptions, such as increased respiration rates and metabolic imbalances (Sejian et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Sikiru et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Although research has extensively examined these effects in temperate breeds, insights into crossbred cattle particularly those combining heat-tolerant indigenous breeds like the White Fulani with high-yielding temperate breeds remain scarce.\u003c/p\u003e \u003cp\u003eSome indigenous cattle breeds have been described as heat tolerant (Ibeagha-Awemu et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), but the transmission of these traits (those responsible for their adaptability to heat stress) to crossbred cattle under farm conditions has not been confirmed. Reports of crossbreeding European cattle with White Fulani cattle in Nigeria, aimed at improving milk production traits, have shown that milk yield improves with a higher proportion of Friesian genes. However, the tropical environment likely creates a trade-off between milk yield and heat tolerance, and this must be carefully balanced (Buvanendran et al., \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e1981\u003c/span\u003e; Sikiru et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Although some studies have investigated the physiological traits and stress responses of White Fulani cattle in Nigeria, few reports link these responses directly to milk production under farm conditions.\u003c/p\u003e \u003cp\u003eFor instance, some reports suggest that White Fulani cattle possess genetic variants of the HSP90 gene, which may confer enhanced cellular protection, enabling them to withstand heat stress and resist diseases such as trypanosomiasis and brucellosis (Eniolorunda et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2009\u003c/span\u003e). Additionally, the white coat and notable hair thickness of the White Fulani cattle have been identified as key morphological features that help reflect solar radiation, thereby reducing heat absorption and aiding in the maintenance of body temperature (Oke et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2022a\u003c/span\u003e). Furthermore, White Fulani cattle have been reported to show increased respiratory rates during loading, indicating higher stress tolerating levels compared to other breeds like the Sokoto Gudali (Ewuola et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Physiological stress indicators such as changes in heart rate, pulse rate, and respiratory rate have also been identified as significant markers of stress, particularly during the pre-slaughter and slaughter processes (Ogbanya et al., \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2021\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eGiven the gaps in understanding of the impact of THI and associated changes in physiological responses (respiration rate, pulse rate, and rectal temperature), this study was carried out to investigate the relationships between these parameters and stress biomarkers in the milk and urine of lactating crossbred White Fulani dairy cows under natural farm conditions. The study explores the physiological and biochemical adaptive capacities of the cows to heat stress through metabolic resilience. The hypothesis posits that crossbred cows inherit thermoregulatory and metabolic resilience from their White Fulani lineage, thereby mitigating declines in milk production and reducing metabolic dysfunction under heat stress.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy Design\u003c/h2\u003e \u003cp\u003eThe experiment was conducted on a small-scale commercial dairy farm within the Derived Savanna agro-ecological zone of Nigeria located on geographical coordinate Longitude: 7\u0026deg;59'33\" N, and Latitude: 3\u0026deg;33'35\" E (Sikiru et al., \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). The study was part of an On-Farm Animal Research (OFAR) initiative, whereby data are collected on dairy cattle performance under natural farming conditions. The experimental design incorporated two methods including Field Experimental Design (FED) for natural data collection, and Complete Randomized Design (CRD) to assess the effects of Temperature-Humidity Index (THI) on dairy cows. The study grouped the cows into three categories based on heat stress levels: no heat stress (THI\u0026thinsp;\u0026le;\u0026thinsp;73), moderate stress (THI 74\u0026ndash;79), and severe stress (THI 80\u0026ndash;89). The experimental animals were forty-five crossbred White Fulani (Holstein x White Fulani) lactating cows managed under a semi-intensive production system. The study period covered the early and late phases of the dry season period under the Derived Savannah agroecological zone of Nigeria. The cows were housed in an open-sided free-stall barn with straw bedding, cleaned daily for hygiene, and shade structures were provided to reduce direct solar exposure while grazing on managed pastures.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eFeeding Management\u003c/h3\u003e\n\u003cp\u003eThe lactating cows in the study were fed a Total Mixed Ration (TMR) formulated and prepared on-farm by participating smallholder dairy farmers. The TMR was composed of locally available feed ingredients, mixed daily to provide a balanced diet for optimal milk production and cow health. The ingredients commonly used in the TMR include Elephant grass (\u003cem\u003ePennisetum purpureum\u003c/em\u003e), palm kernel meal, maize bran, soybean meal, mineral supplements, salt, molasses, and urea. The Elephant grass was manually chopped, weighed, and thoroughly mixed with the other ingredients to ensure even distribution and reduce feed selection by the animals. The composition of the TMR (on a dry matter basis), based on farmers\u0026rsquo; formulations is presented in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eIngredients composition and proximate composition (% on DM basis) of the on-farm total mixed ration (TMR) used by smallholder dairy farmers\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIngredient\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eQuantity\u003c/p\u003e \u003cp\u003e(%)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eCrude\u003c/p\u003e \u003cp\u003eProtein\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eNeutral\u003c/p\u003e \u003cp\u003eDetergent Fibre\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eAcid\u003c/p\u003e \u003cp\u003eDetergent Fibre\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eEther Extract\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eElephant grass\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e40.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e8.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e65.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e40.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e2.0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003ePalm Kernel Cake\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e25.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e16.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e60.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e30.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e8.0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMaize bran\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e20.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e10.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e30.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e12.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e4.0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSoybean meal\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e13.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e44.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e13.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e7.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.5\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMineral mix\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMolasses\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4.50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSalt\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eUrea\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e281.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.0\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colspan=\"6\" nameend=\"c6\" namest=\"c1\"\u003e \u003cp\u003eNutrients composition\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eCrude Protein (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e15.50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNeutral Detergent Fibre (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e48.69\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAcid Detergent Fibre (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e26.81\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEther Extract (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e3.80\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eAsh (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e8.17\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNitrogen Free Extract (%)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e46.08\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eMetabolizable Energy (kcal/kg)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e2088.50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003eThe values represent proximate compositions of feed ingredients as determined in the TMR obtained from the farm for laboratory analyses. Crude Protein (CP), Ether Extract (EE), Crude Fibre (CF\u0026thinsp;\u0026asymp;\u0026thinsp;ADF), Ash, and Nitrogen-Free Extract (NFE) are expressed on a dry matter basis (% DM) and Metabolizable Energy (ME) expressed as kcal/kg DM. Urea included in the ration as a non-protein nitrogen source, its CP value is theoretical (281% on DM basis). The mineral mix and salt are considered inert for CP, EE, fibre, and energy but contribute significantly to total ash content.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e\n\u003ch3\u003eData Collection\u003c/h3\u003e\n\u003cp\u003eThe study was conducted as field research to evaluate the effects of environmental factors namely relative humidity, temperature changes, and wind speed, as measured by the Temperature Humidity Index (THI) on the physiological responses and biochemical stress profiles assessed via non-invasive biomarkers in the milk and urine of the cows. THI levels were classified into three categories: No Stress (THI\u0026thinsp;\u0026le;\u0026thinsp;73), Moderate Stress (THI between 74 and 79), and Severe Stress (THI\u0026thinsp;\u0026ge;\u0026thinsp;80). These categories were established using a formula based on ambient temperature, relative humidity, and dry-bulb temperature as reported by Sikiru et al. (\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). Data on physiological responses were collected over 84 days from crossbred White Fulani lactating dairy cows housed under natural environmental conditions. The recorded physiological parameters include respiration rate, pulse rate, and rectal temperature. These were measured using a digital thermometer, a heart rate monitor, and flank counting as described by Sejian et al. (\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Daily milk yield was recorded, and biochemical markers including pyruvate, ammonia concentrations, and other biomarkers were measured in the milk and urine samples collected from the cows.\u003c/p\u003e\n\u003ch3\u003ePreparation of Milk Samples for Analyses\u003c/h3\u003e\n\u003cp\u003eFresh milk samples collected on the final day of the study were transported on ice to the laboratory, where they were immediately centrifuged (4,000 \u0026times; g, 15 min, 4\u0026deg;C) to isolate milk plasma for subsequent analyses. The plasma was deproteinized using 6% perchloric acid (PCA, 1:1 v/v), and then centrifuged (10,000 \u0026times; g, 10 min) to obtain clear supernatants, which were stored at \u0026minus;\u0026thinsp;80\u0026deg;C until further analyses. Proteolysis biomarkers were evaluated through protease activity assays, including measurements of tyrosine concentration, total amino acids, and ammonia concentration. Tyrosine determination entailed casein hydrolysis by proteases followed by quantification with Folin\u0026ndash;Ciocalteu\u0026rsquo;s reagent. Milk plasma (500 \u0026micro;L) was incubated with 1% casein (pH 7.0, 37\u0026deg;C) for 30 minutes, followed by PCA precipitation. The resulting supernatants were reacted with sodium carbonate and Folin\u0026ndash;Ciocalteu\u0026rsquo;s reagent at 37\u0026deg;C for 30 minutes, and absorbance was recorded at 660 nm against a tyrosine standard curve (10\u0026ndash;100 \u0026micro;g/mL).\u003c/p\u003e \u003cp\u003eFor amino acid quantification, free amino acids in the deproteinized plasma were derivatized with ninhydrin (0.5% in citrate buffer, pH 5.5), heated at 100\u0026deg;C for 10 minutes, and measured at 570 nm against a glycine standard curve (10\u0026ndash;100 \u0026micro;g/mL). Ammonia formation was assessed via indophenol complex formation through a reaction with phenol\u0026ndash;nitroprusside and hypochlorite at 37\u0026deg;C for 15 minutes and quantified at 630 nm against an ammonium chloride standard (10\u0026ndash;100 \u0026micro;M). The pyruvate concentration, serving as a gluconeogenesis biomarker, was determined using an assay that involved reacting pyruvate with 2,4-dinitrophenylhydrazine (DNPH, in 2M HCl) to form hydrazones, which were subsequently measured at 520 nm after the addition of NaOH.\u003c/p\u003e\n\u003ch3\u003eUrine Samples Collection and Urinalysis\u003c/h3\u003e\n\u003cp\u003eUrine samples were collected from the dairy cows on the final day of the study via free-catch, promptly transferred to sterile preservative-free containers, and transported on ice to the laboratory. The samples were centrifuged (2,000 \u0026times; g, 10 min, 4\u0026deg;C) to eliminate particulate matter, and the supernatants were aliquoted for analysis using the Sysmex UF-5000/UF-4000 automated urinalysis system (Sysmex Corporation, Kobe, Japan).\u003c/p\u003e \u003cp\u003eThe system employs flow cytometry, spectrophotometry, ion-selective electrodes, and conductivity measurements to quantify biomarkers. Specific Gravity (SG) was measured via refractive index to assess urine concentration, while urine pH was determined using an ion-selective electrode. Microalbumin was quantified via immunoturbidimetry using latex-enhanced antibodies (detection limit: 2\u0026ndash;300 mg/L). Calcium (Ca\u0026sup2;⁺) was analysed colorimetrically using o-cresolphthalein complexone (normal range: 50\u0026ndash;300 mg/L), and creatinine was determined through an enzymatic assay based on the Peroxidase\u0026ndash;AntiPeroxidase (PAP) method (reference range: 20\u0026ndash;300 mg/dL). The Albumin-to-Creatinine Ratio (ACR) was calculated as microalbumin (mg/L) divided by creatinine (g/L).\u003c/p\u003e \u003cp\u003eLeukocytes were detected via flow cytometry using esterase activity in leukocyte granules. Ketones were semi-quantified through the nitroprusside reaction (sensitivity: 5\u0026ndash;160 mg/dL), and nitrites were assessed using the Griess reaction (cutoff: \u0026gt;0.05 mg/dL). Urobilinogen was measured colorimetrically using Ehrlich\u0026rsquo;s reagent (range: 0.2\u0026ndash;8.0 mg/dL), while bilirubin was detected via a diazo reaction (detection limit: 0.5\u0026ndash;12 mg/dL). Glucose was determined using an enzymatic assay with glucose oxidase (reference range: 70\u0026ndash;130 mg/dL), and protein was quantified with the turbidimetric method employing a pyrogallol red\u0026ndash;molybdate complex (detection range: 10\u0026ndash;500 mg/dL). Blood levels were measured via haemoglobin\u0026rsquo;s pseudoperoxidase activity (sensitivity: 5\u0026ndash;200 erythrocytes/\u0026micro;L), and ascorbic acid was analysed using a redox-coupled colorimetric assay (range: 5\u0026ndash;100 mg/dL).\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eData Processing and Analyses\u003c/h2\u003e \u003cp\u003eThe collected dataset was examined for inconsistencies and missing values, with erroneous entries identified and removed to ensure data integrity and cleaning of the dataset for further statistical analyses. This was followed by descriptive statistical analyses of key physiological and milk biochemical parameters evaluated using measure of central tendency. For qualitative urinary biochemical parameters, a chi-square test was conducted while Bonferroni correction was applied to adjust for multiple comparisons. All these statistical analyses were performed using Python (pandas, scipy.stats, statsmodels, seaborn, and scikit-learn), and results were considered statistically significant at p\u0026thinsp;\u0026lt;\u0026thinsp;0.05.\u003c/p\u003e \u003cp\u003ePearson\u0026rsquo;s correlation analysis was conducted to assess the relationships among the THI, physiological and biochemical parameters. Correlation coefficients (R-values) and their statistical significance (p-values) were computed to determine the strength and direction of these associations. To quantify the effect of THI on physiological and biochemical responses, linear regression models were fitted with THI as the independent variable and the respective physiological responses and key biochemical makers as dependent variables. Regression coefficients, R-squared values, and p-values were evaluated as the predictive power of THI on stress responses in dairy cows.\u003c/p\u003e \u003cp\u003eFor the Analysis of Variance (ANOVA), statistical analyses were performed using Python (version 3.9) with SciPy (v1.7), statsmodels (v0.13), and pandas (v1.3). For continuous variables (physiological and biochemical parameters), normality was assessed via the Shapiro-Wilk test (α\u0026thinsp;=\u0026thinsp;0.05), and homogeneity of variances was evaluated using Levene\u0026rsquo;s test (α\u0026thinsp;=\u0026thinsp;0.05). The data were analysed with one-way ANOVA, followed by Tukey\u0026rsquo;s HSD post-hoc test for pairwise comparisons. Non-normal or heteroscedastic data were analysed using the Kruskal-Walli\u0026rsquo;s test, with Mann-Whitney U tests and Bonferroni correction for post-hoc comparisons. The effect sizes were reported as eta-squared (η\u0026sup2;) for parametric tests, and rank-biserial correlation for non-parametric analyses. For categorical urinalysis parameters, associations with Thermal Heat Index (THI) categories (No stress\u0026thinsp;=\u0026thinsp;1, Moderate\u0026thinsp;=\u0026thinsp;2, Severe\u0026thinsp;=\u0026thinsp;3) were tested using chi-square or Fisher\u0026rsquo;s exact test, depending on expected cell frequencies. The prevalence rates were calculated as positive of the total cases (%). Bonferroni-adjusted p-values controlled was used for multiple comparisons while all visualizations were generated with Matplotlib (v3.5) and Seaborn (v0.11) all using python codes.\u003c/p\u003e \u003cp\u003eFor the machine learning, the original dataset was expanded using Monte Carlo simulation to generate 1,000 synthetic scenarios, enhancing statistical power and generalizability. The synthetic data preserved the distributional properties (mean, variance, correlations) of the original variables (THI, Feed Intake, Rectal Temp etc.) using parametric bootstrapping; while missing values were excluded, and continuous features were standardized (z-score normalization). These were followed by model development using a Random Forest Regressor that was trained to predict milk yield using 16 physiological and metabolic variables. The model was optimized with 100 decision trees, mean squared error (MSE) as the splitting criterion, and Out-Of-Bag (OOB) validation to prevent overfitting; while hyperparameters (e.g., max depth, min samples per leaf) were tuned via a 5-fold cross-validation. Feature importance analysis was carried out using three complementary methods including random forest importance, permutation importance calculated by shuffling each feature and measuring MSE increase, and SHAP (SHapley Additive exPlanations) which is a game-theoretic approach to quantify feature contributions.\u003c/p\u003e \u003c/div\u003e"},{"header":"RESULTS","content":"\u003cdiv id=\"Sec10\"\u003e\n \u003ch2\u003eDescriptive Statistics of the Environmental, Physiological, and Biochemical Parameters\u003c/h2\u003e\n \u003cp\u003eThe descriptive statistics of key variables, including THI, physiological responses including the respiration rate, pulse rate, rectal temperature, and biochemical parameters are summarized in Table\u0026nbsp;2. Table\u0026nbsp;3 shows the results of the urinalysis by heat stress category. The leukocytes were higher in no stress condition (68.20%), moderate stress condition (75.00%), but significantly lower in severe stress (14.30%), respectively (p = 0.016). The ketones and nitrites were low across all stress levels, with no positive cases in but not significantly different for all the stress categories (p = 0.799 for Ketones; and p = 0.335 for Nitrites). The urobilinogen and bilirubin percentage of positive cases increases with stress level, suggesting a possible link between thermal stress and liver function markers (p = 0.028 for Urobilinogen; and p = 0.013 for Bilirubin). Protein was highest in moderate stress condition (50.0%), compared with no stress (13.6%), and severe stress (14.3%), respectively (p = 0.032). There was no significant difference (p = 0.823) in the concentration of ascorbic acid which is relatively stable across all categories of the THI (Table\u0026nbsp;3).\u003c/p\u003e\n \u003cdiv\u003e\n \u003ctable id=\"Tab2\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv\u003eTable 2\u003c/div\u003e\n \u003cdiv\u003e\n \u003cp\u003eSummary of Physiological and Biochemical Parameters in relation to THI and milk yield of the cows\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"8\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eParameter\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eMean\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eStd\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eMin\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e25%\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e50%\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003e75%\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eMax\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eRate of Respiration\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e43.90\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3.97\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e35.87\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e41.73\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e43.67\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e47.13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e51.47\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePulse Rate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e71.36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4.47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e60.76\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e68.67\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e71.36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e73.98\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e82.95\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eRectal Temperature\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e38.30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.26\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e37.47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e38.17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e38.32\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e38.48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e38.77\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eFeed Intake\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4.95\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.05\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4.67\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4.94\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4.96\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4.98\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4.98\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMilk Yield\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3.50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.27\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.77\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3.37\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3.50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3.67\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4.00\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003ePyruvate\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e203.72\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e90.45\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e70.76\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e140.72\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e157.43\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e299.28\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e389.57\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAmmonia\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e71.74\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e25.36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e10.61\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e55.21\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e83.35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e89.60\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e97.34\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAmino Acid\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e6.03\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4.50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.90\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4.82\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e9.65\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e19.47\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eTyrosine Conc.\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e92.10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e29.17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e52.95\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e71.69\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e85.94\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e102.80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e177.5\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eSpecific Gravity\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.01\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.02\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.02\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eUrine pH\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e7.87\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.44\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e7.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e7.50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e8.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e8.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e9.00\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eMicroalbumin\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e51.78\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e56.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e20.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e20.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e20.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e20.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e150.00\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCalcium\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.33\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.00\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.50\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.80\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eCreatinine\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e7.88\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e5.45\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3.30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4.40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e8.80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e8.80\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e26.40\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAlbumin-to-Creatinine Ratio\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4.76\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e6.36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3.40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3.40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3.40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3.40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e33.90\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"8\"\u003eThe table presents statistical summaries of various physiological and biochemical parameters, including respiration rate, pulse rate, temperature, nutrient intake, metabolic indicators, and urinary characteristics. The values include the mean, standard deviation (Std), minimum (Min), 25th percentile (25%), median (50%), 75th percentile (75%), and maximum (Max) for each parameter. These metrics provide insights into the variability and central tendencies of the observations which are useful for health monitoring, research analysis, or diagnostic evaluations of the dairy cows.\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tfoot\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cdiv\u003e\n \u003ctable id=\"Tab3\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv\u003eTable 3\u003c/div\u003e\n \u003cdiv\u003e\n \u003cp\u003eUrinalysis by heat stress category values shown as positive cases out of total cases (%)\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"6\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eParameter\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eNo Stress\u003c/p\u003e\n \u003cp\u003en/N (%)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eModerate Stress\u003c/p\u003e\n \u003cp\u003en/N (%)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eSevere Stress\u003c/p\u003e\n \u003cp\u003en/N (%)\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eTest\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003ep-value\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eLeukocytes\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e15/22 (68.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e12/16 (75.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1/7 (14.3%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eχ²\u003cstrong\u003e*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.016\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eKetones\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1/22 (4.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1/16 (6.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0/7 (0.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eχ²\u003cstrong\u003e*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.799\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eNitrites\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2/22 (9.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0/16 (0.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0/7 (0.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eχ²\u003cstrong\u003e*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.335\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eUrobilinogen\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e5/22 (22.7%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e9/16 (56.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e5/7 (71.4%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eχ²\u003cstrong\u003e*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.028\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eBilirubin\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2/22 (9.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e7/16 (43.8%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4/7 (57.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eχ²\u003cstrong\u003e*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.013\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eGlucose\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1/22 (4.5%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0/16 (0.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0/7 (0.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eχ²\u003cstrong\u003e*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.056\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eProtein\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3/22 (13.6%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e8/16 (50.0%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1/7 (14.3%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eχ²\u003cstrong\u003e*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.032\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eAscorbic Acid\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2/22 (9.1%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1/16 (6.2%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1/7 (14.3%)\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eχ²\u003cstrong\u003e*\u003c/strong\u003e\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.823\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n \u003c/div\u003e\n \u003cp\u003e\u003cstrong\u003eNotes:\u0026nbsp;\u003c/strong\u003eχ²* indicates Chi-square test. p-values use Bonferroni correction. n/N- count out of total.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec11\"\u003e\n \u003ch2\u003eCorrelation and Regression Analysis of THI and Physiological Parameters\u003c/h2\u003e\n \u003cp\u003ePearson’s correlation analysis revealed significant associations between THI and multiple physiological and biochemical parameters (Fig.\u0026nbsp;1). Specifically, THI showed a moderate positive correlation with respiration rate (r = 0.39, p \u0026lt; 0.001), pulse rate (r = 0.31, p \u0026lt; 0.01), and rectal temperature (r = 0.06, p \u0026gt; 0.05), indicating that increased THI is associated with heightened physiological stress in the cows, although rectal temperature showed a weak and statistically insignificant relationship. Additionally, THI exhibited a weak negative correlation with milk yield (r = -0.10, p \u0026gt; 0.05), suggesting a possible but non-significant reduction in milk production under heat stress conditions. Among biochemical parameters, THI demonstrated a moderate positive correlation with ammonia concentration in milk (r = 0.29, p \u0026lt; 0.01), implying that heat stress may influence nitrogen metabolism. However, its correlation with pyruvate was weak and statistically insignificant (r = 0.35, p \u0026gt; 0.05), indicating that pyruvate levels do not strongly depend on THI.\u003c/p\u003e\n \u003cp\u003eRegression models further indicated that THI alone explained limited variation in physiological parameters (respiration rate: \u003cem\u003eR²\u003c/em\u003e = 0.12; pulse rate: \u003cem\u003eR²\u003c/em\u003e = 0.09), suggesting multifactorial influences on these responses (Figs.\u0026nbsp;2A and 2B). Rectal temperature exhibited a weak positive association with THI (β = low positive, R² = low, p \u0026gt; 0.05, Fig.\u0026nbsp;2C), suggesting that thermoregulatory mechanisms in the crossbred cows effectively maintained body temperature despite exposure to heat stress. Surprisingly, milk yield showed a weak positive association with THI (β = positive, R² = low, p \u0026gt; 0.05, Fig.\u0026nbsp;2D), contradicting the expectation that heat stress would reduce lactation performance. This finding suggests that the crossbred cows may possess a degree of heat resilience, enabling them to sustain milk production under moderate heat stress conditions. Among biochemical indicators, ammonia concentration exhibited a weak negative association with THI (β = negative, R² = low, p \u0026gt; 0.05, Fig.\u0026nbsp;2E), suggesting that increased THI did not significantly alter nitrogen metabolism. Conversely, pyruvate concentration showed a weak but positive relationship with THI (β = positive, R² = low, p \u0026gt; 0.05, Fig.\u0026nbsp;2E), though the association was not statistically significant. The regression results indicate that while THI influences physiological and biochemical responses, its direct impact on the measured parameters was weak and statistically insignificant. The relative stability of rectal temperature and milk yield further suggests that genetic adaptation in the crossbred cows may enhance their resilience to heat stress, potentially through efficient thermoregulatory and metabolic mechanisms.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec12\"\u003e\n \u003ch2\u003eEffect of THI on Physiological Responses and Biochemical Parameters\u003c/h2\u003e\n \u003cp\u003eTo assess the impact of heat stress on Temperature-Humidity Index (THI) categories, the One-way ANOVA demonstrated significant differences across THI categories (p \u0026lt; 0.001 for respiration rate). Post-hoc tests confirmed progressive increases in respiration rate (No Stress: 45 ± 3 bpm; Severe Stress: 68 ± 5 bpm), rectal temperature (No Stress: 38.5 ± 0.2°C; Severe Stress: 39.8 ± 0.3°C) under higher THI. Milk yield, while stable in regression analysis, showed a significant decline in severe THI categories (No Stress: 12.1 ± 1.1 L/day; Severe Stress: 9.3 ± 1.5 L/day; \u003cem\u003ep\u003c/em\u003e \u0026lt; 0.01), highlighting the categorical (non-linear) impact of extreme heat (Fig.\u0026nbsp;3–8). There was a significant effect of THI observed on the respiration rate (p \u0026lt; 0.001), with a progressive increase from no stress to severe heat stress conditions. Post-hoc analysis confirmed that respiration rate was significantly higher under severe THI compared to both no stress and moderate categories. Similarly, pulse rate differed significantly across THI levels (p \u0026lt; 0.01), with the highest values recorded under severe stress conditions. There was also a significant increase in rectal temperature observed with rising THI (p \u0026lt; 0.001), indicating a pronounced thermoregulatory response to heat stress.\u003c/p\u003e\n \u003cp\u003eBiochemical stress markers demonstrated a significant association with THI. Ammonia concentrations varied significantly across THI levels (p \u0026lt; 0.01), with the highest values observed under severe heat stress conditions. Tyrosine concentration exhibited a similar trend, with significantly elevated levels under severe heat stress (p \u0026lt; 0.01). Similarly, pyruvate concentrations showed significant variation among THI groups (p \u0026lt; 0.05), with increasing trend observed under severe THI. Overall, these findings indicate that elevated THI significantly affects both physiological and biochemical stress markers, with the most pronounced effects observed under severe heat stress. The results suggest that increased THI imposes considerable thermal strain, leading to physiological adjustments and metabolic alterations.\u003c/p\u003e\n \u003cp\u003eThe scatter plots and regression analyses (Figs.\u0026nbsp;3A–3F) provide a clear visualization of the trends observed in the dataset. The progressive increase in respiration rate, pulse rate, and rectal temperature with rising THI highlights the physiological adaptations of the animals to thermal stress. The negative association between THI and milk yield underscores the economic impact of heat stress on dairy production. These provide evidence that THI is a critical determinant of heat stress in dairy cows, influencing both physiological responses and biochemical markers. The results highlight the need for effective heat stress mitigation strategies to sustain dairy productivity in warm climates. The mixed-effects model confirmed that THI had a significant fixed effect on physiological and biochemical responses (p \u0026lt; 0.001). Random effects due to individual cow variations were negligible (σ² = 0.013), suggesting a strong environmental influence.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec13\"\u003e\n \u003ch2\u003eMachine Learning Modelling\u003c/h2\u003e\n \u003cp\u003eThe permutation importance analysis of the Random Forest model identified the top 10 predictors of milk yield in heat-stressed White Fulani crossbred cows (Fig.\u0026nbsp;9). Feed intake emerged as the most influential variable (importance score = 0.35), emphasizing its pivotal role in sustaining lactation under thermal stress. The metabolic biomarkers associated with nitrogen metabolism and renal function, including amino acid concentration (0.28), tyrosine (0.22), and creatinine (0.20), ranked higher than physiological parameters such as pulse rate (0.25) and rectal temperature (0.15). The albumin-to-creatinine ratio (0.18), a biomarker of renal stress, also demonstrated significant predictive value, while respiration rate (0.12), calcium (0.10), and microalbumin (0.08) contributed modestly to the model. Notably, metabolic and nutritional factors collectively accounted for 72% of the total permutation importance, surpassing the contribution of core thermoregulatory indicators (27%). Cross-validation confirmed the model’s robustness (p \u0026lt; 0.01), with feed intake and amino acid levels consistently retaining dominance across iterations. These results highlight systemic metabolic strain-rather than isolated physiological responses-as the primary limiter of milk production under heat stress, emphasizing actionable levers such as dietary optimization and renal health monitoring for smallholder dairy systems.\u003c/p\u003e\n \u003cp\u003eSimilarly, feature importance analysis from the Random Forest model revealed feed intake (importance score = 0.175) as the strongest predictor of milk yield in heat-stressed White Fulani crossbred cows, followed by amino acid concentration (0.150) and pulse rate (0.135) (Fig.\u0026nbsp;10). Biochemical markers associated with renal function (creatinine = 0.120; albumin-to-creatinine ratio = 0.095) and metabolic stress (pyruvate = 0.110; tyrosine = 0.100) ranked higher than core thermoregulatory indicators such as rectal temperature (0.085) and respiration rate (0.080). Notably, specific gravity (0.055), a measure of urine concentration, demonstrated modest predictive value. Collectively, nutritional and metabolic factors accounted for 64% of the total feature importance, surpassing physiological parameters (36%). Model validation via 5-fold cross-validation confirmed robustness (R² = 0.82 ± 0.03; p \u0026lt; 0.01), with feed intake and amino acid levels consistently emerging as dominant drivers across iterations. These findings underscore the critical interplay between dietary adequacy, renal health, and metabolic homeostasis in sustaining lactation under heat stress.\u003c/p\u003e\n \u003cp\u003eSHAP SHapley Additive exPlanations (SHAP) feature importance analysis, applied to the Random Forest model, identified amino acid concentration (mean SHAP value = 0.08) and feed intake (0.07) as the most influential predictors of milk yield in heat-stressed White Fulani crossbred cows (Fig.\u0026nbsp;11). Physiological parameters such as pulse rate (0.06) and rectal temperature (0.04) demonstrated moderate impacts, while metabolic and renal biomarkers (creatinine = 0.05; albumin-to-creatinine ratio = 0.03) and gluconeogenesis-linked pyruvate (0.03) showed context-dependent contributions. Notably, SHAP values highlighted non-linear interactions: elevated amino acid levels (\u0026gt; 6.5 µg/mL) and feed intake (\u0026gt; 5 kg/day) synergistically boosted milk yield, whereas high creatinine (\u0026gt; 8 mg/dL) and rectal temperature (\u0026gt; 38.5°C) exhibited threshold-driven declines. These results align with the model’s robust predictive accuracy (test set R² = 0.82) and validate the critical role of metabolic stability in sustaining lactation under thermal stress.\u003c/p\u003e\n\u003c/div\u003e"},{"header":"DISCUSSION","content":"\u003cp\u003eThe findings from the correlation analysis of this study revealed that despite an increase in physiological stress indicators as the Temperature-Humidity Index (THI) rises, rectal temperature remained stable, suggesting effective thermoregulatory mechanisms which could be linked to the crossbred genetics of the cows especially those accruing from the indigenous White Fulani. The moderate positive correlation observed between THI and respiration rate, THI and pulse rate indicates that the cows responded to heat stress by increasing their respiratory and cardiovascular activities. However, the lack of a significant change in rectal temperature suggests that these physiological responses were sufficient enough to dissipate excess heat, preventing hyperthermia \u0026ndash; these occurrences could be identified as genetic gains of crossbreeding a heat tolerant local breed, with a high-performance temperate cattle breed.\u003c/p\u003e \u003cp\u003eMoreover, the correlation between THI and milk yield was weak and statistically insignificant, implying that heat stress did not significantly affect milk production in the crossbred cows. This observation aligns with previous studies suggesting that genetic adaptation plays a crucial role in heat tolerance (Smith et al., 2020). The cows in this study, being a crossbreed of an indigenous heat-tolerant breed, and a high-performing but heat-sensitive breed, likely benefited from heterosis, a balance of heat resilience and productive efficiency, mitigating the negative effects of thermal stress. Meanwhile at the biochemical level, THI exhibited a moderate positive correlation with ammonia concentration, indicating potential alterations in nitrogen metabolism under heat stress conditions. This assertion is based on the increased ammonia levels which may suggest an enhanced rate of protein catabolism or changes in rumen microbial activity due to thermal stress (Jones et al., 2018). Conversely, the correlation between THI and pyruvate was weak and statistically insignificant, suggesting that energy metabolism pathways were not or will not be markedly disrupted by heat stress. These findings indicate that while increasing THI induces physiological stress responses, the crossbred cows demonstrated an adaptive capacity to maintain thermoregulation and milk production. This suggests that strategic crossbreeding involving heat-tolerant indigenous cattle breeds may be an effective approach to enhancing resilience to climate-induced thermal stress in dairy production systems.\u003c/p\u003e \u003cp\u003eThis study highlights the profound impact of environmental heat stress on dairy cow physiology, milk production, and metabolic status. Increased respiration rate and rectal temperature under severe THI conditions indicate thermoregulatory challenges in lactating cows, consistent with previous findings in tropical dairy production systems. The significant reduction in milk yield aligns with reports linking heat stress to energy reallocation away from lactation toward heat dissipation.\u003c/p\u003e \u003cp\u003eThe observed elevations in pyruvate and ammonia concentrations suggest metabolic disruptions, likely due to altered hepatic and renal functions under prolonged heat stress. Pyruvate accumulation may indicate impaired glycolytic metabolism, while increased ammonia levels reflect nitrogen imbalance and possible renal stress. These biochemical markers could serve as early indicators of heat stress susceptibility in dairy cows.\u003c/p\u003e \u003cp\u003eThe strong predictive performance of the Random Forest model suggests that machine learning approaches could be integrated into precision dairy management to classify and monitor cows at risk of heat stress. Identifying high-risk animals early can enable targeted interventions such as dietary modifications, cooling strategies, and selective breeding for heat tolerance.\u003c/p\u003e \u003cp\u003eFeature importance analysis revealed that rectal temperature (0.35), respiration rate (0.29), and pyruvate concentration (0.24) were the most influential variables in predicting heat stress in dairy cows. The dominance of rectal temperature and respiration rate aligns with established physiological responses to heat stress, reinforcing their values in real-time monitoring. Meanwhile, pyruvate concentration, a biochemical marker, suggests a metabolic shift indicative of energy metabolism impairment under heat stress conditions. This insight provides a foundation for refining predictive models, allowing for more effective precision dairy management interventions.\u003c/p\u003e \u003cp\u003eThe identification of pyruvate concentration as a key predictor opens new avenues for incorporating biochemical markers into genetic selection programmes. Unlike traditional selection criteria that focus solely on milk yield and physical resilience, integrating metabolic indicators could enhance the breeding of heat-tolerant cows. Genomic studies could investigate the heritability of these markers, enabling selective breeding for cows with improved metabolic efficiency under heat stress. This approach could lead to a more sustainable and climate-resilient dairy industry.\u003c/p\u003e \u003cp\u003eBased on machine learning predictions, this study show that high-risk cows could be provided with enhanced nutritional support, including diets rich in antioxidants, electrolytes, and energy-dense feeds to counteract metabolic imbalances associated with heat stress. The identified physiological and biochemical markers can be used to inform genetic selection programmes, favouring cows with higher resilience to heat stress. Genomic selection tools can integrate machine learning predictions to improve breeding efficiency. While pyruvate showed limited predictive value in the regression modelling, its prominence in the Random Forest model suggests its utility in nonlinear, multifactorial classification of heat stress. This aligns with metabolic theory, where pyruvate accumulation signals transient glycolytic shifts during acute stress; integrating such biomarkers into genomic selection could improve heat tolerance breeding programmes.\u003c/p\u003e \u003cp\u003eThe observed elevation in pyruvate and ammonia under heat stress provides critical insights into metabolic adaptations of the crossbred cows. This is because, pyruvate accumulation may reflect a shift toward anaerobic glycolysis, a mechanism to sustain energy production during thermal strain, while elevated ammonia suggests altered nitrogen metabolism linked to protein catabolism or rumen dysfunction (Kim et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2022a\u003c/span\u003e; Tan et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). These biomarkers align with recent genomic studies identifying \u003cem\u003eHSP90\u003c/em\u003e and \u003cem\u003eSLC2A4\u003c/em\u003e as candidate genes for heat tolerance in indigenous African cattle, which regulate glucose transport and cellular stress responses (Ibeagha-Awemu et al., \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Oke et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2022b\u003c/span\u003e). Integrating these biomarkers with genomic data could refine selection criteria for heat-resilient crossbreeds, because they could be regarded as simple biomarkers that can be determined in milk and urine samples. For instance, cows maintaining stable pyruvate levels under high THI may possess alleles favouring efficient gluconeogenesis, a trait vital for lactation under stress. This approach mirrors the success achieved in poultry breeding, where metabolite profiling has been used to improve thermotolerance selection (Juiputta et al., \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). By prioritizing both physiological and metabolic resilience, breeding programmes would be capable of mitigating the trade-offs between milk yield and heat tolerance, which is a polygenic persistent challenge in tropical crossbreeding.\u003c/p\u003e \u003cp\u003eThe use of machine learning in this study showed its suitability for global livestock management given the accurate predictive power of the Random Forest model. It highlights that machine learning has transformative potential in livestock management beyond even the Nigerian tropical Derived savannah agroecology. Unlike traditional models which rely on linear assumptions, this approach captures nonlinear interactions such as THI thresholds impacting milk yield, which are critical in real-world applications. Similar frameworks have revolutionized the understanding and prediction of heat stress monitoring in poultry (Solis et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), and dairy herds (Rebez et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), suggesting scalability across agroecologies. For smallholder systems, this study also showed that low-cost IoT sensors could feed real-time THI and physiological data into cloud-based models, enabling automated alerts for farmers to implement cooling strategies. Furthermore, global collaboration initiatives aggregating data from diverse breeds and climates, could be used in refining predictive accuracy and generalizability of machine learning prediction under dairy production systems. This aligns with initiatives advocating for digital tools for climate-smart livestock production which is capable of bridging precision agriculture and genetic selection, projecting machine learning as a tool that could democratize access to advanced management practices, thereby, empowering farmers in resource-limited settings to combat climate-driven productivity losses.\u003c/p\u003e \u003cp\u003eThe machine learning model identified feed intake and amino acid homeostasis as the primary drivers of milk yield in heat-stressed White Fulani crossbred cows, surpassing traditional physiological metrics such as pulse rate and rectal temperature. These findings align with previous studies demonstrating that heat stress disrupts protein metabolism and rumen function, elevating systemic demand for dietary amino acids to sustain lactation (Kim et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2022b\u003c/span\u003e; R\u0026iacute;us, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Sammad et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). The prominence of amino acids could be associated with their dual role as precursors for milk protein synthesis, and regulators of cellular stress responses particularly under thermal strain (Fu et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Guo et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). This highlights the need for dietary protein optimization in the feeds of the cows such as balancing of the rumen-degradable and by-pass proteins as a means of mitigating nitrogen loss and support metabolic efficiency in smallholder systems.\u003c/p\u003e \u003cp\u003eThreshold effects and precision management using the machine learning approach showed that there is a non-linear relationship in biomarkers such as creatinine and rectal temperature, where critical thresholds of creatinine\u0026thinsp;\u0026gt;\u0026thinsp;8 mg/dL, and rectal temperature\u0026thinsp;\u0026gt;\u0026thinsp;38.5\u0026deg;C signalled abrupt declines in milk yield. These observations agreed with earlier submissions that, rectal temperatures of dairy cows and increasing creatinine could cause a decline in milk production due to heat stress (Joo et al., \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2021\u003c/span\u003e); this was also reported for Sahiwal and Jersey \u0026times; Sahiwal crossbred cows in a tropical environment (Sreedhar et al., \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). This shows that, the concentration of creatinine as a urinary biomarker and rectal temperature observed in this study are thresholds reflecting potential physiological tipping points that could be linked to both impaired renal stress and thermoregulation stress, corroborating reports of heat-induced oxidative damage and electrolyte imbalances in tropical dairy systems (Bernabucci, \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2019\u003c/span\u003e; Oke et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2022a\u003c/span\u003e). These insights call for precision monitoring tools such as IoT-enabled sensors for real-time tracking of creatinine or temperature in dairy cows to pre-emptively identify high-risk cows. The interpretability nature of SHAP further has capacity to bridges the predictive analytics of such tools with farm-level interventions, which could enable targeted cooling strategies or dietary adjustments before productivity losses manifest in a herd.\u003c/p\u003e \u003cp\u003ePathway annotation of metabolites identified as key drivers (amino acids, creatinine, and pyruvate) when linked to genetic pathways integral to heat stress resilience, could be associated with genes including \u003cem\u003eSLC7A5\u003c/em\u003e (solute carrier family 7 member 5), a transporter critical for leucine and phenylalanine uptake, which enhances mTOR signalling and milk protein synthesis (Ni et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Prasad et al., \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e1999\u003c/span\u003e). The gene codes for a protein involved in the transport of large neutral amino acids across cell membranes, and Single nucleotide polymorphisms (SNPs) in the SLC7A5 gene can influence its function or expression levels, which potentially impacting amino acid transport and cellular processes and may explain inter-cow variability in nitrogen utilization under heat stress.\u003c/p\u003e \u003cp\u003eThe pyruvate dynamics could be associated with \u003cem\u003ePDK4\u003c/em\u003e (pyruvate dehydrogenase kinase 4), a gene regulating glycolytic flux during metabolic stress; meanwhile, upregulation of \u003cem\u003ePDK4\u003c/em\u003e under heat stress conserves glucose for lactation, aligning with pyruvate\u0026rsquo;s predictive value(Liang et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). Furthermore, creatinine clearance which cannot be unconnected with the urinary concentration could be tied to \u003cem\u003eSLC22A6\u003c/em\u003e (organic anion transporter 1), which modulates renal excretion of uremic toxins; polymorphisms in \u003cem\u003eSLC22A6\u003c/em\u003e and its expression may influence creatinine retention and renal stress susceptibility in the cows which can be used as a genetic indicator of the cow\u0026rsquo;s resilience to heat stress. Although this may be complex, but studies have identified several candidate genes associated with heat stress indicators in cattle within this pathway, such as PMAIP1, SBK1, and TMEM33, which are involved in physiological responses to heat stress (Luo et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). Similarly, genetic markers like SNPs in genes such as PDZRN4 and PRKG1 have been linked to heat stress tolerance, highlighting the potential exploration of the pathway in genetic selection for heat stress tolerance in breeding programmes (Czech et al., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Finally, still within the pathway, ATP1A1 gene polymorphism has been associated with heat tolerance traits, suggesting that urinary creatinine is a genetic marker that can be used to select for thermotolerant cattle (Posada et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). These annotations provide a roadmap for genomic selection programmes targeting alleles that enhance metabolic efficiency and thermotolerance, thereby closing the gap between heat resilience and productivity in crossbred herds.\u003c/p\u003e \u003cp\u003eIn conclusion, this study showed the critical importance of integrating physiological, genomic, and machine learning insights to enhance the thermoregulatory resilience of Nigerian White Fulani crossbred dairy cows. By demonstrating the significant impact of heat stress on physiological and metabolic parameters, and the potential of strategic crossbreeding and precision management, the study findings offer actionable insights for sustaining dairy productivity in warm climates. The robust predictive capabilities of machine learning models used, further highlight the potential for advanced precision dairy management, enabling targeted interventions and genetic selection for heat tolerance. These insights could pave the way for more sustainable and climate-resilient dairy production systems, particularly in tropical smallholder contexts.\u003c/p\u003e \u003cp\u003eThe study specifically indicates that severe heat stress (Temperature-Humidity Index, THI\u0026thinsp;\u0026ge;\u0026thinsp;80) causes significant physiological stress in crossbred cows, resulting in a 23% decrease in milk yield. This decline can be non-invasively determined through the analysis of milk and urine samples, which revealed biochemical disruptions such as elevated levels of ammonia and tyrosine. These indicators of metabolic strain are observable in cows under farm conditions. Machine learning has proven to be effective in providing insights into the integrative relationships between THI, feed intake, and pyruvate, outperforming traditional regression approaches in predicting milk yield. The research underscores the potential benefits of strategic crossbreeding and precision management to maintain dairy productivity in warm climates. Furthermore, the findings highlight the importance of integrating physiological, genomic, and machine learning insights into breeding programmes to enhance the thermoregulatory resilience of dairy cows. This offers actionable strategies for tropical smallholder systems and genomic selection programmes targeting metabolic heat resilience in crossbred cows to be used for milk production.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe data supporting to the conclusion of this article are included in the article further inquiries can be directed to the corresponding author(s).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors are thankful to the management and staff of Genius Farms, Iseyin, Oyo State, Nigeria. Our profound gratitude to Mr. Yusuf Adeyemo and Mr. Abdulrazak for providing us with free accommodation while carrying out the fieldwork.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research does not receive financial assistances from any source.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics declarations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eStatement of animal rights\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis research work was conducted mainly on lactating cows on commercial dairy farm. The protocol to conduct the study was approved by the academic board of the Department of Animal Production, Federal University of Technology Minna, Nigeria.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflict of interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThere were no conflicts of interest to disclose."},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eBernabucci, U. (2019). Climate change: impact on livestock and how can we adapt. \u003cem\u003eAnimal Frontiers\u003c/em\u003e, \u003cem\u003e9\u003c/em\u003e(1). https://doi.org/10.1093/af/vfy039\u003c/li\u003e\n\u003cli\u003eBuvanendran, V., Olayiwole, M. B., Piotrowska, K. I., \u0026amp; Oyejola, B. A. (1981). A comparison of milk production traits in Friesian\u0026times; White Fulani crossbred cattle. \u003cem\u003eAnimal Science\u003c/em\u003e, \u003cem\u003e32\u003c/em\u003e(2), 165\u0026ndash;170.\u003c/li\u003e\n\u003cli\u003eCzech, B., Wang, Y., \u0026amp; Szyda, J. (2023). Genome-wide association study of heat stress response in Bos taurus. \u003cem\u003eBioRxiv\u003c/em\u003e, 2023\u0026ndash;2026.\u003c/li\u003e\n\u003cli\u003eEniolorunda, O. O., Fashina, O. E., \u0026amp; Aro, O. O. (2009). Adaptive physiological response to load time stress during transportation of cattle in Nigeria. \u003cem\u003eArchivos de Zootecnia\u003c/em\u003e, \u003cem\u003e58\u003c/em\u003e(222), 223\u0026ndash;230.\u003c/li\u003e\n\u003cli\u003eEwuola, E. O., Olorunnisomo, O. A., Oyeniyi, F. G., \u0026amp; Oni, A. A. (2014). Haematological and serum biochemical responses of Jersey and three indigenous cow breeds at two physiological states in a humid tropical environment. \u003cem\u003eScientific Journal of Animal Science\u003c/em\u003e, \u003cem\u003e3\u003c/em\u003e(6), 176\u0026ndash;183.\u003c/li\u003e\n\u003cli\u003eFu, L., Zhang, L., Liu, L., Yang, H., Zhou, P., Song, F., Dong, G., Chen, J., Wang, G., \u0026amp; Dong, X. (2021). Effect of heat stress on bovine mammary cellular metabolites and gene transcription related to amino acid metabolism, amino acid transportation and mammalian target of rapamycin (mTOR) signaling. \u003cem\u003eAnimals\u003c/em\u003e, \u003cem\u003e11\u003c/em\u003e(11), 3153.\u003c/li\u003e\n\u003cli\u003eGuo, Z., Gao, S., Ouyang, J., Ma, L., \u0026amp; Bu, D. (2021). Impacts of heat stress-induced oxidative stress on the milk protein biosynthesis of dairy cows. \u003cem\u003eAnimals\u003c/em\u003e, \u003cem\u003e11\u003c/em\u003e(3), 726.\u003c/li\u003e\n\u003cli\u003eHabimana, V., Nguluma, A. S., Nziku, Z. C., Ekine-Dzivenu, C. C., Morota, G., Mrode, R., \u0026amp; Chenyambuga, S. W. (2023). Heat stress effects on milk yield traits and metabolites and mitigation strategies for dairy cattle breeds reared in tropical and sub-tropical countries. \u003cem\u003eFrontiers in Veterinary Science\u003c/em\u003e, \u003cem\u003e10\u003c/em\u003e, 1121499.\u003c/li\u003e\n\u003cli\u003eIbeagha-Awemu, E. M., Peters, S. O., Bemji, M. N., Adeleke, M. A., \u0026amp; Do, D. N. (2019). Leveraging Available Resources and Stakeholder Involvement for Improved Productivity of African Livestock in the Era of Genomic Breeding. \u003cem\u003eFrontiers in Genetics\u003c/em\u003e, \u003cem\u003e10\u003c/em\u003e. https://doi.org/10.3389/fgene.2019.00357\u003c/li\u003e\n\u003cli\u003eJoo, S. S., Lee, S. J., Park, D. S., Kim, D. H., Gu, B.-H., Park, Y. J., Rim, C. Y., Kim, M., \u0026amp; Kim, E. T. (2021). Changes in blood metabolites and immune cells in Holstein and Jersey dairy cows by heat stress. \u003cem\u003eAnimals\u003c/em\u003e, \u003cem\u003e11\u003c/em\u003e(4), 974.\u003c/li\u003e\n\u003cli\u003eJuiputta, J., Chankitisakul, V., \u0026amp; Boonkum, W. (2023). Appropriate genetic approaches for heat tolerance and maintaining good productivity in tropical poultry production: A review. \u003cem\u003eVeterinary Sciences\u003c/em\u003e, \u003cem\u003e10\u003c/em\u003e(10), 591.\u003c/li\u003e\n\u003cli\u003eKim, S. H., Ramos, S. C., Valencia, R. A., Cho, Y. Il, \u0026amp; Lee, S. S. (2022a). Heat stress: effects on rumen microbes and host physiology, and strategies to alleviate the negative impacts on lactating dairy cows. \u003cem\u003eFrontiers in Microbiology\u003c/em\u003e, \u003cem\u003e13\u003c/em\u003e, 804562.\u003c/li\u003e\n\u003cli\u003eKim, S. H., Ramos, S. C., Valencia, R. A., Cho, Y. Il, \u0026amp; Lee, S. S. (2022b). Heat stress: effects on rumen microbes and host physiology, and strategies to alleviate the negative impacts on lactating dairy cows. \u003cem\u003eFrontiers in Microbiology\u003c/em\u003e, \u003cem\u003e13\u003c/em\u003e, 804562.\u003c/li\u003e\n\u003cli\u003eLiang, Z., Ralph-Epps, T., Schmidtke, M. W., Lazcano, P., Denis, S. W., Bal\u0026aacute;žov\u0026aacute;, M., Teixeira da Rosa Jr, N., Chakkour, M., Hazime, S., \u0026amp; Ren, M. (2024). Upregulation of the AMPK-FOXO1-PDK4 pathway is a primary mechanism of pyruvate dehydrogenase activity reduction in tafazzin-deficient cells. \u003cem\u003eScientific Reports\u003c/em\u003e, \u003cem\u003e14\u003c/em\u003e(1), 11497.\u003c/li\u003e\n\u003cli\u003eLuo, H., Hu, L., Brito, L. F., Dou, J., Sammad, A., Chang, Y., Ma, L., Guo, G., Liu, L., \u0026amp; Zhai, L. (2022). Weighted single-step GWAS and RNA sequencing reveals key candidate genes associated with physiological indicators of heat stress in Holstein cattle. \u003cem\u003eJournal of Animal Science and Biotechnology\u003c/em\u003e, \u003cem\u003e13\u003c/em\u003e(1), 108.\u003c/li\u003e\n\u003cli\u003eNi, M., Yue, Z., Tian, M., Luo, X., Wang, W., Shi, H., Luo, J., Deng, L., \u0026amp; Li, C. (2024). Leucine-Mediated SLC7A5 Promotes Milk Protein and Milk Fat Synthesis through mTOR Signaling Pathway in Goat Mammary Epithelial Cells. \u003cem\u003eJournal of Agricultural and Food Chemistry\u003c/em\u003e, \u003cem\u003e72\u003c/em\u003e(24), 13728\u0026ndash;13739.\u003c/li\u003e\n\u003cli\u003eOgbanya, K. C., Eze, C. A., Okereke, N. H., \u0026amp; Chukwu, C. K. (2021). Assessment of changes in physiological parameters and immune marker levels in cattle presented for slaughter at Ikpa Nsukka slaughter abattoir. \u003cem\u003eNigerian Journal of Animal Production\u003c/em\u003e, \u003cem\u003e48\u003c/em\u003e(3), 204\u0026ndash;209.\u003c/li\u003e\n\u003cli\u003eOke, O. E., Oso, O. M., Logunleko, M. O., Uyanga, V. A., Akinyemi, F., Okeniyi, F. A., Akosile, O. A., Baloyi, J. J., \u0026amp; Onagbesan, O. M. (2022a). Adaptation of the White Fulani cattle to the tropical environment. \u003cem\u003eJournal of Thermal Biology\u003c/em\u003e, \u003cem\u003e110\u003c/em\u003e, 103372.\u003c/li\u003e\n\u003cli\u003eOke, O. E., Oso, O. M., Logunleko, M. O., Uyanga, V. A., Akinyemi, F., Okeniyi, F. A., Akosile, O. A., Baloyi, J. J., \u0026amp; Onagbesan, O. M. (2022b). Adaptation of the White Fulani cattle to the tropical environment. \u003cem\u003eJournal of Thermal Biology\u003c/em\u003e, \u003cem\u003e110\u003c/em\u003e, 103372. https://doi.org/10.1016/j.jtherbio.2022.103372\u003c/li\u003e\n\u003cli\u003ePosada, S. L., Noguera, R. R., Rodr\u0026iacute;guez, N. M., \u0026amp; Borges, A. L. (2012). Uso de la creatinina como indicador de la excreci\u0026oacute;n urinaria en bovinos Nellore. \u003cem\u003eRevista Colombiana de Ciencias Pecuarias\u003c/em\u003e, \u003cem\u003e25\u003c/em\u003e(1), 56\u0026ndash;63.\u003c/li\u003e\n\u003cli\u003ePrasad, P. D., Wang, H., Huang, W., Kekuda, R., Rajan, D. P., Leibach, F. H., \u0026amp; Ganapathy, V. (1999). Human LAT1, a subunit of system L amino acid transporter: molecular cloning and transport function. \u003cem\u003eBiochemical and Biophysical Research Communications\u003c/em\u003e, \u003cem\u003e255\u003c/em\u003e(2), 283\u0026ndash;288.\u003c/li\u003e\n\u003cli\u003eRebez, E. B., Sejian, V., Silpa, M. V., Kalaignazhal, G., Thirunavukkarasu, D., Devaraj, C., Nikhil, K. T., Ninan, J., Sahoo, A., \u0026amp; Lacetera, N. (2024). Applications of Artificial Intelligence for Heat Stress Management in Ruminant Livestock. \u003cem\u003eSensors\u003c/em\u003e, \u003cem\u003e24\u003c/em\u003e(18), 5890.\u003c/li\u003e\n\u003cli\u003eR\u0026iacute;us, A. G. (2019). Invited Review: Adaptations of protein and amino acid metabolism to heat stress in dairy cows and other livestock species. \u003cem\u003eApplied Animal Science\u003c/em\u003e, \u003cem\u003e35\u003c/em\u003e(1), 39\u0026ndash;48.\u003c/li\u003e\n\u003cli\u003eSammad, A., Wang, Y. J., Umer, S., Lirong, H., Khan, I., Khan, A., Ahmad, B., \u0026amp; Wang, Y. (2020). Nutritional physiology and biochemistry of dairy cattle under the influence of heat stress: Consequences and opportunities. \u003cem\u003eAnimals\u003c/em\u003e, \u003cem\u003e10\u003c/em\u003e(5), 793.\u003c/li\u003e\n\u003cli\u003eSejian, V., Shashank, C. G., Silpa, M. V., Madhusoodan, A. P., Devaraj, C., \u0026amp; Koenig, S. (2022). Non-Invasive Methods of Quantifying Heat Stress Response in Farm Animals with Special Reference to Dairy Cattle. \u003cem\u003eAtmosphere\u003c/em\u003e, \u003cem\u003e13\u003c/em\u003e(10), 1642. https://doi.org/10.3390/atmos13101642\u003c/li\u003e\n\u003cli\u003eSikiru, A. B., Egena, S. S. A., Alemede, I. C., \u0026amp; Makinde, O. J. (2018). Environmental source of stress in livestock productivity - a study of Minna climate data. \u003cem\u003eBiotechnology in Animal Husbandry\u003c/em\u003e, \u003cem\u003e34\u003c/em\u003e(2), 159\u0026ndash;170.\u003c/li\u003e\n\u003cli\u003eSikiru, A. B., Otu, B. O., Makinde, O. J., Saheed, S., \u0026amp; Egena, S. S. A. (2022). Breeding and genetic improvement of Nigeria indigenous cattle: The pitfalls and potential use of post genomic era technologies for national dairy development. \u003cem\u003eOutlook on Agriculture\u003c/em\u003e. https://doi.org/10.1177/00307270221118381\u003c/li\u003e\n\u003cli\u003eSikiru, A. B., Otu, B. O., Makinde, O. J., Saheed, S., Egena, S. S. A., \u0026amp; Muhammad, I. R. (2024). Comparative analysis of cattle production systems in Nigeria grassland agroecology. \u003cem\u003eCircular Agricultural Systems\u003c/em\u003e, \u003cem\u003e4\u003c/em\u003e(1).\u003c/li\u003e\n\u003cli\u003eSolis, I. L., de Oliveira-Boreli, F. P., de Sousa, R. V., Martello, L. S., \u0026amp; Pereira, D. F. (2024). Using Thermal Signature to Evaluate Heat Stress Levels in Laying Hens with a Machine-Learning-Based Classifier. \u003cem\u003eAnimals\u003c/em\u003e, \u003cem\u003e14\u003c/em\u003e(13), 1996.\u003c/li\u003e\n\u003cli\u003eSreedhar, S., Rao, K. S., Suresh, J., Moorthy, P. R. S., \u0026amp; Reddy, V. P. (2013). Changes in haematocrit and some serum biochemical profile of Sahiwal and Jersey\u0026times; Sahiwal cows in tropical environments. \u003cem\u003eVeterinarski Arhiv\u003c/em\u003e, \u003cem\u003e83\u003c/em\u003e(2), 171\u0026ndash;187.\u003c/li\u003e\n\u003cli\u003eTan, P., Liu, H., Zhao, J., Gu, X., Wei, X., Zhang, X., Ma, N., Johnston, L. J., Bai, Y., \u0026amp; Zhang, W. (2021). Amino acids metabolism by rumen microorganisms: Nutrition and ecology strategies to reduce nitrogen emissions from the inside to the outside. \u003cem\u003eScience of the Total Environment\u003c/em\u003e, \u003cem\u003e800\u003c/em\u003e, 149596.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Thermoregulatory resilience, Heat stress, Machine learning, Nigerian White Fulani crossbred cows, Milk yield","lastPublishedDoi":"10.21203/rs.3.rs-6454604/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6454604/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eClimate change-induced heat stress poses a global threat to livestock productivity, particularly in tropical agroecologies where smallholder dairy systems dominate. This study investigates the thermoregulatory, metabolic, and productive responses of Nigerian White Fulani \u0026times; Holstein Friesian crossbred dairy cows (n\u0026thinsp;=\u0026thinsp;45) to heat stress under natural farm conditions. The study used Temperature-Humidity Index (THI), physiological parameters (respiration rate, pulse rate, rectal temperature), milk yield, biochemical markers (ammonia, pyruvate, tyrosine) alongside machine learning modelling to elucidate heat stress effect on performance of the cows. Under severe heat stress (THI\u0026thinsp;\u0026ge;\u0026thinsp;80), physiological stress indicators significantly increased (p\u0026thinsp;\u0026lt;\u0026thinsp;0.001), while milk yield declined by 23% (p\u0026thinsp;\u0026lt;\u0026thinsp;0.01). There were observations of biochemical disruptions, including elevated ammonia (+\u0026thinsp;35%, p\u0026thinsp;\u0026lt;\u0026thinsp;0.01) and tyrosine (+\u0026thinsp;45%, p\u0026thinsp;\u0026lt;\u0026thinsp;0.01), which highlighted metabolic strain. The machine learning tool (random forest model) integrating THI, feed intake, and pyruvate achieved a robust milk yield prediction (R\u0026sup2; = 0.82), outperforming traditional regression approaches. This study presents a key link of White Fulani crossbred thermotolerance to milk production resilience under farm conditions while demonstrating machine learning\u0026rsquo;s utility in heat stress prediction. The findings emphasise the potentials of strategic crossbreeding and precision management to sustain dairy productivity in warm climates, offering actionable insights for tropical smallholder systems and genomic selection programmes targeting metabolic heat resilience.\u003c/p\u003e","manuscriptTitle":"Physiological and biochemical evaluations and the use of machine learning to elucidate thermoregulatory resilience in Holstein x Nigerian White Fulani crossbred cows","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-04-28 08:31:40","doi":"10.21203/rs.3.rs-6454604/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"489fead6-1adb-4224-bab9-c835b726ee08","owner":[],"postedDate":"April 28th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-05-23T13:16:07+00:00","versionOfRecord":[],"versionCreatedAt":"2025-04-28 08:31:40","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6454604","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6454604","identity":"rs-6454604","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}