Prediction of some milk production traits using udder and teat measurements with a spotlight on their genetic background in Friesian cows

The aggregate udder shape (Bowl, Round, Cup), udder measurements (udder width, UW, udder front depth, UFD, udder rear depth, URD, udder levelness, ULV, udder heights, UH) and teat measurements (teat diameters, TD, front teat length, FTL, rear teat length, RTL, distance between front teats, DFT and distance between rear teats, DRT) were measured on 1300 Friesian cows located in a commercial farm under subtropical conditions (Egypt) to appraise udder and teats status and to evaluate the possible relationships with some milk production characteristics in conjunction with udder shape, age at first calving, sire and inbreeding effects on udder morphological traits and milk production ability. For such an available sample size, parity had affected (P < 0.01) UFD, TD, FTL and RTL. In addition, udder shape affected (P < 0.01) UW, UFD, URD, ULV and DFT. None of the other studied factors affected milk production traits. The bowl udder shape (P < 0.01) yielded a high total milk yield (3267.19 kg), adjusted milk yield (2443.01 kg) and lactation length (480.70 d) compared to other udder shapes. The genetic correlations of UW with total milk yield and persistency were strongly positive (0.86 and 0.93, respectively). However, strong negative genetic correlations were found between UW with peak milk yield and lactation length (− 0.92 and − 0.80, respectively), between RTL with peak milk yield (− 0.92) and DRT with persistency (− 0.79). As found from the stepwise multiple regression, UW and URD can be used as good indicators for predicting milk yield and lactation length. Additionally, this study spotlights the genetic background of udder characteristics based on reliable studies and the QTL database for cattle as a first step toward applying this knowledge side by side with phenotypic traits to improve the productivity of the Holstein breed under subtropical conditions.


Traits of concern
The udder and teat measurements (in cm) were taken monthly on a fixed day throughout the lactation period one hour before the evening milking.The udder and teat measurements were udder width (UW); distance between right and left sides of the udder at the widest point, udder front depth (UFD); distance from the merging point of fore udder with the abdomen to a point in front of the fore teats at the level of the udder base, udder rear depth (URD); distance from the bottom of the vulva to the base of the rear udder, udder levelness (ULV); the difference between the rear and front udder heights 39 , udder heights (UH); distance from the ground to udder floor in front of the front or behind rear teat, teat diameters (TD); diameters of front teat or rear teat measured at mid-point of each teat length from teat orifice to base of udder, front teat length (FTL); length of front teat from teat orifice to base of udder, rear teat length (RTL); length of rear teat from teat orifice to base of udder, distance between front teats (DFT); distances between front teats before milking and at mid teat length, distance between rear teats (DRT); distances between rear teats before milking and at mid teat length (Fig. 1).

Genetic background of udder and teat measurements
According to reliable investigations, many genome regions, candidate genes, SNPs and QTLs have been reported to be correlated with several udder characteristics in different species including cattle.In this regard, the genetic background for udder characteristics in the cattle has been investigated utilizing; (1) FAO-Database (https:// www.fao.org), (2) Animal QTL-Database: (https:// www.anima lgeno me.org/ cgi-bin/ QTLdb/ BT/ index) and (3) Genome-Informatics-Resources: (https:// www.anima lgenom e.org/bioinfo) for further and complementary future studies.

Statistical analysis
All sets of data were tested for normality with the Shapiro-Wilk test from the UNIVARIATE procedure SAS 9.0 (2009) 41 , and results indicated that all data were distributed normally (W ≥ 0.90).To study the factors affecting udder or teat characteristics, MIXED procedure of SAS 9.0 (2009) 41 was used according to the following model: where: Y ijklmn = any of the udder or teat measurements, µ = the overall mean, S i = the random effect of ith sire (i = 1:180), P j = the fixed effect of jth parity (j = 1:5), I k = the fixed effect of kth inbreeding (inbred or non-inbred), U l = the fixed effect of lth udder shape (bowel, cup and round), b m = a regression coefficient of each udder and teat measurement on the independent continuous variable of age at first calving (AFC, x)specific to the teat or udder measurement and x = the respective mean, and e ijklmn = the residual error.
Significant differences among means within each factor were tested using Duncan test.The heritability values of each udder or teat measurement was estimated using variance components (Proc VARCOMP, SAS 2009) 41 according to the following equation 42 : where; σ 2 S = the variance of sire and σ 2 e = the variance of error.Simple regressions and correlations between udder or teat measurements and all milk production traits were calculated.Besides, the stepwise multiple regression analysis of milk production traits on udder and teat measurements were calculated using STEPWISE procedure of SAS 41 to determine the best regression model.

Ethical approval
The study was performed in accordance with the ethical standards as laid down in the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards.Also, this experimental protocol was approved by the Ethics Committee of Damanhur University under approval number DFU-2023-2.

Effect of sire, inbreeding, parity, udder shape and age at first calving on udder and teat measurements
Means, standard deviations and coefficients of variation of udder and teat characteristics in different udder shape groups are shown in Table 1.All characteristics were varied from shape to other, bowl shape had highest udder measurements (UW, UFD, URD and ULV) expect (UH).On the contrary cup shape had lowest udder measurements (UW, UFD, URD and ULV) expect (UH) which were high.The variation in udder measurements (UW, UFD and URD) between udder shapes ranged from 9.62 to 21.27% expect ULV (cm) which had the highest variation (C.V. = − 76.70).Bowl shaped udder had highest teat diameters and distance between front teats (2.79 and 15.25 cm, respectively) while the round shape had highest front teat length and rear teat length (7.03 and 6.78 cm, respectively).On the other hand, the cup shaped udder had highest distance between rear teats (8.63 cm).
Least squares means and standard errors for UW, UFD, URD, ULV and UH as affected by sire, inbreeding status, parity, udder shape and age at first calving are presented in Table 2. Sire, inbreeding status and age at first calving had no effects (P > 0.01) on UW, UFD, URD, ULV or UH, but parity had an effect (P < 0.01) on UFD which was the highest (22.58 cm) in parity 3. Udder shape had effects (P < 0.01) on UW, UFD, URD and ULV with highest values of UW, UFD and URD (38.47, 23.65 and 28.91, respectively) obtained for bowl shaped udders but the highest values of ULV was obtained in cases of round (− 3.48 cm) and cup shaped (− 3.03 cm).
Least squares means and standard errors for TD, FTL, RTL, DFT and DRT of sire, inbreeding status, parity, udder shape and age at first calving are presented in Table 3. Sire, inbreeding status and age at first calving had no effect (P > 0.01) on TD, FTL, RTL, DFT and DRT.Parity had effects (P < 0.01) on TD, FTL and RTL with the highest values obtained in parities 3, 4 and 5. Udder shape had no effects on teat measurements except DFT which was the highest (P < 0.01) for cup shaped udder (16.41 cm).

Relationships between udder or teat measurements and some milk production traits
Estimates of heritability and genetic correlation for udder and teat measurements with milk production traits are summarized in Table 5.The estimated Heritability for the reproduction traits were close to zero.Heritability estimates ranged between 0.464 (for TD) and 0.103 (for UH).The genetic correlations between TMY, 305d-MY, Peak, LL and PR with udder and teat measurements are presented in Table 5. UW had strong positive genetic correlation with 305d-MY and PR (0.86 and 0.93).Also, UH has similar strong positive correlation with PR (0.87).On the other hand, strong negative genetic correlations were found between UW with both PMY and LL (− 0.92 and − 0.80, respectively) and between UFD with PR (− 0.72), URD with PMY (− 0.84), RTL with PMY (− 0.92) and DRT with PR (− 0.79).This reflects a negative relationship between these traits.Simple correlations and regressions of some milk production traits on the examined udder or teat characteristics (cm) are presented in Table 6.UW had the highest (P < 0.05) positive correlations with milk production traits and ranged from 0.23 to 0.87.Also, UFD and URD had high (P < 0.05) positive correlations with milk production traits except that between URD and PR which were moderate positive but insignificant.The correlations between ULV, UH with milk traits were negative except those with PR which were nearly equal zero (0.03).On the other hand, the correlations between teat measurements and PR were moderate negative (P < 0.05).Also, the correlations between teat measurements with TMY, 305d-MY and LL approached zero (P > 0.05).Beside, TD, DFT Stepwise multiple regression analysis was utilized to predict milk traits from udder and teats measurements as shown in Table 7. Indeed, UW still the best predictor for most milk production traits followed by udder rear depth for predicting TMY, 305d-MY and LL (R 2 = 74.02,77.43 and 42.26, respectively) and then no other variables met the level of significant.In contrast, teat measurements were entered in equation for predicting PMY and PR but with low change in R 2 .DRT was the first measurement entered for predicting persistency (R 2 = 6.67) followed by UW with change in R 2 = 4.97.

Genetic background of udder and teat measurements
By investigating the QTL database for the cattle, the results revealed that there are 193,641 QTLs uncovered by 1122 different studies which were associated with 686 different economic traits.4040 QTLs out of 193,641 were associated with udder traits (Table 8).These QTLs represent a huge raw material for future further studies as complementary to the current study on the same breed and individuals.Worth mentioning, Fig. 2 shows the whole genome analysis for 4,040 QTLs which were associated with 19 udder traits spread on 31 chromosomes in the cattle, while Fig. 3 shows locations where the udder traits are mapped by QTLs or candidate SNPs associations on different chromosomes in the cattle.

Discussion
The calculated coefficients of variation revealed low discrepancies among cows for udder characteristics which confirm a possible genetic control for udder conformation.Every udder shape has its own specific characteristics which impose an expressive effect on udder and teat parameters.The present results on udder circumference were similar to those reported by Mona and Semaida 43 Sid Ahmed and El Barbary 44 on cup shaped udder of Friesian cows, but the fore udder depths was lower than those of cup (29.2 cm), round (27.4 cm) and goaty (25.2 cm) udder shapes.Also, it was higher than 17.2 cm found in Romanian Black Spotted cows 45 .The hind udder depth results were much lower than 33.7 cm and 31.2 cm that reported for cup and round udder shapes of Friesian dairy cows, respectively 44 .The present result of udder rear depths for bowl udder shape (28.3 cm) was similar to that reported by Avarvarei 45 in Romanian Black Spotted cows.The present results of teat measurements were similar to those of Deng et al. 46 .Milne 47 revealed that morphology of the teat is recognized as part of the passive defence mechanism against intra-mammary infection, so short teats are more favourable for high milk producing cows than long teats.
The effects of parity and udder shape on udder and teat measurements in the current study were similar to those reported by Tilki et al. 28 who reported that the udder measurements were affected by lactations number because udder tissue might be continuous to develop up to parity six then after starts to regress thereafter.This was confirmed by Singh et al. 48who reported that all udder and teat measurements had showed increasing trends up to the 5th parity.Consistently, Bhuiyan et al. 14 reported that udder length increased gradually up to lactation six.Modh et al. 49 reported an increasing trend at the rate of 24.3 and 9.7% in udder length and width, respectively of Gir cows at the first two parity, but afterwards the udder length became static while udder width became static   up to parity four then exhibited 15.16% increased between parity four and five, whereas udder depth was at par.However, the mean udder depth increased from 19.4 to 28.5 cm, respectively in Vrindavani cattle 48 similar to the present results.Modh et al. 49 reported gradual increase in length, width and depth of the udder as the number of parity increases; multiparous cows had larger volume of udder than primiparous cows.The differences observed in the present study in udder length, width and depth in different parities were statistically significant (P < 0.05).Similar findings were reported by Patel 50 .
The current FTL and RTL were smaller than those reported by Modh et al. 49 who obtained FTL and RTL of Gir cows in different parities ranging from 8.74 to 9.82 and 7.97 to 8.88 cm, respectively with gradual nonsignificant increase in length of fore (7.56%) and rear (8.28%) teat between parities one and two.Moreover, Patel et al. 51 found significant differences observed between FTL and RTL in different parities of crossbred cows ranging from 5.48 to 6.52 cm and 4.92 to 5.94 cm, respectively with gradual increase in length by advancement of parity.Similar results were reported on Holstein, Vrindavani cattle, Hariana cows, Tharparkar cows, Kankrej cows and Gir cows by many workers 48,52 .With respect to teat diameters, Modh et al. 49 found that the differences observed between fore and rear teat diameters in different parities were at par and no definite trend was detected with advancement of parity.Similar results were recorded by Sharma et al. 16 on Hariana and Tharparkar cows.For teats length, Antalík and Strapák 53 observed a gradual increase with advancement parity order.Similar to the current results, Tilki et al. 28 and Antalík and Strapák 53 reported that stage of lactation did not exert any effect on the udder or teat measurements.
The present parity effect on milk production traits was not significant.These results were in agreement with Atakan 54 who found that parity had no significant effect on lactation length and 305-day milk yield of Friesian cows.On the contrary, Mellado et al. 55 found that parity had significant influence on milk yield.Also, Lee and Kim 56 reported that the differences in total milk yield among parities were significant.The contradicting results may be attributed to difference in analytical models, in herd size and the age of animals.However, the relationship between udder shapes and milk yield of cattle are well established.The present results are in agreement with those of Ghosh and Prasad 57 and Bhuiyan et al. 14 who obtained high milk yield from cows having bowl shaped udder compared to other udder shapes in cows.Also, Prasad et al. 58 found that Murrah buffaloes bowl shaped udders produced the highest average daily milk yield followed by pendulous, globular and then goaty udders.Thus raising cows with bowel shaped udder may improve the efficiency of milk production.
The high positive genetic correlation of UW, UFD and URD with milk production activities were in agreement with these evaluated by various authors.Such high genetic correlations indicate that genetic selection for high www.nature.com/scientificreports/milk yield will be associated with wider udders.In the current study and in previous literature for Holsteins 59 udder depth was the most unfavourably correlated with milk yield but the rear udder height and width were the most favourably related to milk yield.Also, Berry et al. 60 reported mild positive genetic correlation (0.36) between udder support and milk yield.DeGroot et al. 61 reported negative genetic correlations (− 0.45) between fore-udder attachment and milk yield.On the contrary, Samoré et al. 62 found favourable correlation between strong fore-udder attachments and high yield of Brown Swiss data.High genetic correlations between the teat diameter and milk production indicate that genetic selection for milk yield will establish cows with substantial teat size.However, weak correlations were estimated between front teat placement, teat length, rear teat position and central ligament with milk yield 63 .Mavrogenis et al. 64 , reported that the relationships between teat measures and milk production were generally low.Also, Otwinowska-Mindur et al. 65 found that the genetic correlations of persistency measures with rear teat placement were rather low negative.this should be regarded as favourable, because selection for better persistency would decrease the scores for rear teat placement towards the optimum of this trait.
In this study, heritability estimates for udder and teat measurements were low to moderate.Low estimates of h 2 for udder and teat measurements with milk production traits concluded that environmental variation contributed the major part of the total variation for the milk production traits, thus management may be an effective factor in improving such traits.In addition, the heritability estimates for udder and teat measurements can change without any genetic change occurring when the environment starts contributing to more variation.In the previous studies, heritability recorded for rear udder height was − 0.77 59 and 0.31 61 .Špehar et al. 66 obtained a heritability value of 0.14 for fore udder attachment.The heritability of the teat diameter in this study was similar to that reported by Seykora and McDanel 17 which was 0.44 and that of teat length was equal to Tapki and Guzey 59 estimate.
In this study, UH and RTL having low heritability declared the importance of environmental and non-additive genetic effects on alteration of these traits.However, the traits moderate to high heritability estimates may show response for direct selection and reasonable correlation response for other traits.The magnitudes of these values indicate that a considerable proportion of the phenotypic variation occurs due to differences in genes with additive effects and that genetic gain in response to selection procedures might occur.Heritability values of udder characteristics in various studies show great variability depending on breed and housing and scoring system 67 .
The current UW having the highest correlation coefficients and R 2 was the most suitable indicator for predicting milk traits follow by udder front and rear depth.None of the teat measurements could be used efficiently for predicting milk yield in this dataset.Deng et al. 46 obtained relatively high positive correlation coefficient (0.64) between milk yield and udder length and concluded that the latter should be an important criterion for selecting dairy cows and can be used with a satisfactory precision for predicting milk yield because the regression of milk yield on udder length had the highest R 2 and each cm increase in udder length represent 0.22 kg extra yield of milk.Sinha et al. 68 reported 0.51, 0.51 and 0.55 coefficients of determinations for the regressions of total milk yield, 305 days milk yield and peak yield, respectively on udder width and Singh et al. 48eported higher accuracy of prediction for the regression of milk yield on udder width measurement.
Concerning the genetic background of udder characteristics, the heritability of udder and teat measurements/ characteristics is moderate (0.23-0.45), which in turn facilitates the genetic improvement programmes 69 .Genomic regions related to the udder and teat traits were reported in several dairy cattle 38 .Numerous reliable studies uncovered many QTLs and their association with udder and teat structure, as economic and significant phenotypes (Table 8, Figs. 2 and 3).In this aspect, 15 SNP loci on BTA-5 were related to udder support scores in the cattle.One out of these 15 was associated with average teat diameter 70 .Additionally, Flury 71 confirmed that seven udder conformation traits were correlated with five genomic regions on BTA-3, -5, -6, -17, and -25 in Brown Swiss cattle.Where, several significant SNPs on BTA-6 were associated with teats diameter and fore udder length.While height-significant SNPs in the coding-region of SNX-29 gene were related to trait rear udder.Moreover, many significant SNPs on BTA-17 (62 Mb) were related to front teat placement, rear teat placement and rear udder width.Also, Marete et al. 72 reported that 10 candidate genes were uncovered for their associations with udder traits, including RREB1, FGF2, FGFR-2, ESR-1, IQGAP-3, GLI-2, PGR, BTRC , TGFBR-2 and PRLR, in French dairy cattle.Furthermore, 18 candidate SNPs within STXBP-6 and SLF-1 genes were related to five udder traits in the Chinese Holstein breed 38 .Meanwhile, several SNPs within LGALS-2, GCLC and ADGRB-3 genes were related to udder depth trait in Holstein cattle 38,73 .
On the other side, udder traits are related to sustainable milk production 74 , mastitis resistance 75 and longevity 76 .Thus, there is a great benefit to considering and investigating these traits side by side with their genetic background in future studies.

Conclusion
Result of the present study shows that parity had significant effect on udder and teat characteristics.Udder shape has a considerable relationship and affects milk yield and lactation length.The bowel udders shape produces high milk yield and associate with long lactation length compared to other udder shapes, therefore, should be considered while selecting Friesian cows.The genetic correlation of udder, teat measurements with milk yield indicated a correlated response; therefore, selection to improve teat diameter and udder width should increase milk yield.All udder and teat characteristics could be included in a selection index criterion (except UH and RTL) to improve the total merit of lactating cows directly or through correlated response with other traits.Also, udder width and udder rear depth can be used as a good indicator for predicting milk yield and lactation length.Finally, the detected QTLs, candidate genes and significant SNPs are potential tools to improve milk productivity side by side with the udder phenotypic investigations.

Figure 2 .
Figure 2. Whole genome analysis for QTL/association enrichment in cattle, including 4,040 QTLs associated with 19 udder traits spread on 31 chromosomes, utilizing QTL Data-Base.*FDR: is short for "false discovery rate", representing the expected proportion of type I errors.A type I error is where you incorrectly reject the null hypothesis, i.e. you get a false positive.It's statistical definition is FDR = E (V/R|R > 0) P(R > 0), where V = Number of Type I errors (false positives); R = Number of rejected hypotheses.Benjamini-Hochberg procedure is a practical way to estimate FDR.

Figure 3 .
Figure 3. QTL/associations for udder characteristics in the Cattle Genome (It shows genome locations where an udder trait is mapped by QTL or SNP associations).

Table 1 .
Means, standard deviation (SD) and coefficient of variation (CV) for udder and teat measurements in different udder shapes of Friesian cows.a UW Udder width, UFD Udder front depth, URD Udder rear depth, ULV Udder levelness, UH Udder height, TD Teat diameter, FTL Front teat length, RTL Rear teat length, DFT Distance between front teats, DRT Distance between rear teats.

Table 2 .
Effect of sire, inbreeding, parity, udder shape and age at first calving on udder measurements (cm) for Friesian cows (LSM ± SE).UW Udder width, UFD Udder front depth, URD Udder rear depth, ULV udder levelness, UH Udder height, ns Not significant. 1Number of sire. 2 Number between practices: simple regression (b ± SE). a -c Least Squares Means with different letters in the same column are significantly different.**Significant at P < 0.01.

Table 5 .
Genetic parameters estimates for udder and teat measurements with some milk production traits.a UW Udder width, UFD Udder front depth, URD Udder rear depth, ULV Udder levelness, UH Udder height, TD Teat diameter, FTL Front teat length, RTL Rear teat length, DFT Distance between front teats, DRT Distance between rear teats, TMY Total milk yield, 305d-MY Adjusted milk yield, PMY Peak milk yield, LL Lactation length, PR Persistency.

Table 6 .
Simple DRT teats were moderated correlated (P < 0.05) with PMY.The same trend was found for regression; udder width had the highest R 2 values for all milk traits except persistency and ranged from 31.61 to 75.22 followed by URD and UFD.The regression coefficients indicated that, for each 1 cm increase in UW an increase of 110.8 kg (R 2 = 70.41) in TMY and of 10.79 day (R 2 = 40.26)increase in LL occurred.Low R 2 values were found for ULV and UH (ranged from 0.09 to 11.95) and all teat measurements (ranged from 0.06 to 6.95) with milk production traits.

Table 7 .
The best stepwise multiple regression equations of milk production traits on udder and teat measurements.UW Udder width, URD Udder rear depth, TD Teat diameter, FTL Front teat length, RTL Rear teat length, DFT Distance between front teats, DRT Distance between rear teats, TMY Total milk yield, 305d-MY Adjusted milk yield, PMY Peak milk yield, LL Lactation length, PR Persistency, a Intercept, b Regression coefficient, R 2 Coefficient of determination.

Table 8 .
Number of QTLs for many species based on animal QTL Database updated to 2023.