Prediction of key milk biomarkers in dairy cows through milk mid-infrared spectra and international collaborations

C Grelet; T Larsen; M A Crowe; D C Wathes; C P Ferris; K L Ingvartsen; C Marchitelli; F Becker; A Vanlierde; J Leblois; U Schuler; F J Auer; A Köck; L Dale; J Sölkner; O Christophe; J Hummel; A Mensching; J A Fernández Pierna; H Soyeurt; M Calmels; R Reding; M Gelé; Y Chen; N Gengler; F Dehareng; GplusE consortium,

doi:10.3168/jds.2023-23843

Prediction of key milk biomarkers in dairy cows through milk mid-infrared spectra and international collaborations

C Grelet, T Larsen, M A Crowe, D C Wathes, C P Ferris, K L Ingvartsen, C Marchitelli, F Becker, A Vanlierde, J Leblois, U Schuler, F J Auer, A Köck, L Dale, J Sölkner, O Christophe, J Hummel, A Mensching, J A Fernández Pierna, H SoyeurtM Calmels, R Reding, M Gelé, Y Chen, N Gengler, F Dehareng^*, GplusE consortium,

^*Corresponding author for this work

Research output: Contribution to journal/Conference contribution in journal/Contribution to newspaper › Journal article › Research › peer-review

4 Citations (Scopus)

Abstract

At the individual cow level, sub-optimum fertility, mastitis, negative energy balance and ketosis are major issues in dairy farming. These problems are widespread on dairy farms and have an important economic impact. The objectives of this study were: 1) to assess the potential of milk Mid Infrared (MIR) spectra to predict key biomarkers of energy deficit (citrate, isocitrate, glucose-6P, free glucose), ketosis (BHB and acetone), mastitis (NAGase and LDH), and fertility (progesterone); 2) to test alternative methodologies to partial least square regression (PLS) to better account for the specific asymmetric distribution of the biomarkers; and 3) to create robust models by merging large data sets from 5 international or national projects. Benefiting from this international collaboration, the data set comprised a total of 9,143 milk samples from 3,758 cows located in 589 herds across 10 countries and represented 7 breeds. The samples were analyzed by reference chemistry for biomarker contents while the MIR analyses were performed on 30 instruments from different models and brands, with spectra harmonized into a common format. Four quantitative methodologies were evaluated to address the strongly skewed distribution of some biomarkers. PLS was used as the reference basis, and compared with a random modification of distribution associated with PLS (Random-downsampling-PLS), an optimized modification of distribution associated with PLS (KennardStone-downsampling-PLS) and Support Vector Machine (SVM). When the ability of MIR to predict biomarkers was too low for quantification, different qualitative methodologies were tested to discriminate low vs high values of biomarkers. For each biomarker, 20% of the herds were randomly removed within all countries to be used as the validation data set. The remaining 80% of herds were used as the calibration data set. In calibration, the 3 alternative methodologies outperform the PLS performances for the majority of biomarkers. However, in the external herd validation, PLS provided the best results for isocitrate, glucose-6P, free glucose and LDH (R2v = 0.48, 0.58, 0.28, and 0.24). For other molecules, PLS-Random-downsampling and PLS-KennardStone-downsampling outperformed PLS in the majority of cases, but the best results were provided by SVM for citrate, BHB, acetone, NAGase and progesterone (R2v = 0.94, 0.58, 0.76, 0.68, and 0.15). Hence, PLS and SVM based on the entire data set provided the best results for normal and skewed distributions, respectively. Complementary to the quantitative methods, the qualitative discriminant models enabled the discrimination of high and low values for BHB, acetone, and NAGase with a global accuracy around 90%, and glucose-6P with an accuracy of 83%. In conclusion, MIR spectra of milk can enable quantitative screening of citrate as a biomarker of energy deficit and discrimination of low and high values of BHB, acetone, and NAGase, as biomarkers of ketosis and mastitis. Finally, progesterone could not be predicted with sufficient accuracy from milk MIR spectra to be further considered. Consequently, MIR spectrometry can bring valuable information regarding the occurrence of energy deficit, ketosis and mastitis in dairy cows, which in turn have major influences on their fertility and survival.

Original language	English
Journal	Journal of Dairy Science
Volume	107
Issue	3
Pages (from-to)	1669-1684
Number of pages	16
ISSN	0022-0302
DOIs	https://doi.org/10.3168/jds.2023-23843
Publication status	Published - Mar 2024

Keywords

Fourier transform mid-infrared spectrometry
fertility
ketosis
mastitis
negative energy balance
Isocitrates
Acetylglucosaminidase
Citrates
Glucose
3-Hydroxybutyric Acid
Cattle Diseases
L-Lactate Dehydrogenase
Animals
Cattle
Biomarkers
Mastitis/veterinary
Female
Ketosis/diagnosis
Progesterone
Citric Acid
Acetone
Milk

Access to Document

10.3168/jds.2023-23843Licence: CC BY

Cite this

Grelet, C., Larsen, T., Crowe, M. A., Wathes, D. C., Ferris, C. P., Ingvartsen, K. L., Marchitelli, C., Becker, F., Vanlierde, A., Leblois, J., Schuler, U., Auer, F. J., Köck, A., Dale, L., Sölkner, J., Christophe, O., Hummel, J., Mensching, A., Pierna, J. A. F., ... GplusE consortium, (2024). Prediction of key milk biomarkers in dairy cows through milk mid-infrared spectra and international collaborations. Journal of Dairy Science, 107(3), 1669-1684. https://doi.org/10.3168/jds.2023-23843

@article{ff114b22f82348e09bd4629ac0518c0c,

title = "Prediction of key milk biomarkers in dairy cows through milk mid-infrared spectra and international collaborations",

abstract = "At the individual cow level, sub-optimum fertility, mastitis, negative energy balance and ketosis are major issues in dairy farming. These problems are widespread on dairy farms and have an important economic impact. The objectives of this study were: 1) to assess the potential of milk Mid Infrared (MIR) spectra to predict key biomarkers of energy deficit (citrate, isocitrate, glucose-6P, free glucose), ketosis (BHB and acetone), mastitis (NAGase and LDH), and fertility (progesterone); 2) to test alternative methodologies to partial least square regression (PLS) to better account for the specific asymmetric distribution of the biomarkers; and 3) to create robust models by merging large data sets from 5 international or national projects. Benefiting from this international collaboration, the data set comprised a total of 9,143 milk samples from 3,758 cows located in 589 herds across 10 countries and represented 7 breeds. The samples were analyzed by reference chemistry for biomarker contents while the MIR analyses were performed on 30 instruments from different models and brands, with spectra harmonized into a common format. Four quantitative methodologies were evaluated to address the strongly skewed distribution of some biomarkers. PLS was used as the reference basis, and compared with a random modification of distribution associated with PLS (Random-downsampling-PLS), an optimized modification of distribution associated with PLS (KennardStone-downsampling-PLS) and Support Vector Machine (SVM). When the ability of MIR to predict biomarkers was too low for quantification, different qualitative methodologies were tested to discriminate low vs high values of biomarkers. For each biomarker, 20% of the herds were randomly removed within all countries to be used as the validation data set. The remaining 80% of herds were used as the calibration data set. In calibration, the 3 alternative methodologies outperform the PLS performances for the majority of biomarkers. However, in the external herd validation, PLS provided the best results for isocitrate, glucose-6P, free glucose and LDH (R2v = 0.48, 0.58, 0.28, and 0.24). For other molecules, PLS-Random-downsampling and PLS-KennardStone-downsampling outperformed PLS in the majority of cases, but the best results were provided by SVM for citrate, BHB, acetone, NAGase and progesterone (R2v = 0.94, 0.58, 0.76, 0.68, and 0.15). Hence, PLS and SVM based on the entire data set provided the best results for normal and skewed distributions, respectively. Complementary to the quantitative methods, the qualitative discriminant models enabled the discrimination of high and low values for BHB, acetone, and NAGase with a global accuracy around 90%, and glucose-6P with an accuracy of 83%. In conclusion, MIR spectra of milk can enable quantitative screening of citrate as a biomarker of energy deficit and discrimination of low and high values of BHB, acetone, and NAGase, as biomarkers of ketosis and mastitis. Finally, progesterone could not be predicted with sufficient accuracy from milk MIR spectra to be further considered. Consequently, MIR spectrometry can bring valuable information regarding the occurrence of energy deficit, ketosis and mastitis in dairy cows, which in turn have major influences on their fertility and survival.",

keywords = "Fourier transform mid-infrared spectrometry, fertility, ketosis, mastitis, negative energy balance, Isocitrates, Acetylglucosaminidase, Citrates, Glucose, 3-Hydroxybutyric Acid, Cattle Diseases, L-Lactate Dehydrogenase, Animals, Cattle, Biomarkers, Mastitis/veterinary, Female, Ketosis/diagnosis, Progesterone, Citric Acid, Acetone, Milk",

author = "C Grelet and T Larsen and Crowe, {M A} and Wathes, {D C} and Ferris, {C P} and Ingvartsen, {K L} and C Marchitelli and F Becker and A Vanlierde and J Leblois and U Schuler and Auer, {F J} and A K{\"o}ck and L Dale and J S{\"o}lkner and O Christophe and J Hummel and A Mensching and Pierna, {J A Fern{\'a}ndez} and H Soyeurt and M Calmels and R Reding and M Gel{\'e} and Y Chen and N Gengler and F Dehareng and {GplusE consortium,}",

year = "2024",

month = mar,

doi = "10.3168/jds.2023-23843",

language = "English",

volume = "107",

pages = "1669--1684",

journal = "Journal of Dairy Science",

issn = "0022-0302",

publisher = "Elsevier Inc.",

number = "3",

}

Grelet, C, Larsen, T, Crowe, MA, Wathes, DC, Ferris, CP, Ingvartsen, KL, Marchitelli, C, Becker, F, Vanlierde, A, Leblois, J, Schuler, U, Auer, FJ, Köck, A, Dale, L, Sölkner, J, Christophe, O, Hummel, J, Mensching, A, Pierna, JAF, Soyeurt, H, Calmels, M, Reding, R, Gelé, M, Chen, Y, Gengler, N, Dehareng, F & GplusE consortium, 2024, 'Prediction of key milk biomarkers in dairy cows through milk mid-infrared spectra and international collaborations', Journal of Dairy Science, vol. 107, no. 3, pp. 1669-1684. https://doi.org/10.3168/jds.2023-23843

Prediction of key milk biomarkers in dairy cows through milk mid-infrared spectra and international collaborations. / Grelet, C; Larsen, T; Crowe, M A et al.
In: Journal of Dairy Science, Vol. 107, No. 3, 03.2024, p. 1669-1684.

Research output: Contribution to journal/Conference contribution in journal/Contribution to newspaper › Journal article › Research › peer-review

TY - JOUR

T1 - Prediction of key milk biomarkers in dairy cows through milk mid-infrared spectra and international collaborations

AU - Grelet, C

AU - Larsen, T

AU - Crowe, M A

AU - Wathes, D C

AU - Ferris, C P

AU - Ingvartsen, K L

AU - Marchitelli, C

AU - Becker, F

AU - Vanlierde, A

AU - Leblois, J

AU - Schuler, U

AU - Auer, F J

AU - Köck, A

AU - Dale, L

AU - Sölkner, J

AU - Christophe, O

AU - Hummel, J

AU - Mensching, A

AU - Pierna, J A Fernández

AU - Soyeurt, H

AU - Calmels, M

AU - Reding, R

AU - Gelé, M

AU - Chen, Y

AU - Gengler, N

AU - Dehareng, F

AU - GplusE consortium,

PY - 2024/3

Y1 - 2024/3

N2 - At the individual cow level, sub-optimum fertility, mastitis, negative energy balance and ketosis are major issues in dairy farming. These problems are widespread on dairy farms and have an important economic impact. The objectives of this study were: 1) to assess the potential of milk Mid Infrared (MIR) spectra to predict key biomarkers of energy deficit (citrate, isocitrate, glucose-6P, free glucose), ketosis (BHB and acetone), mastitis (NAGase and LDH), and fertility (progesterone); 2) to test alternative methodologies to partial least square regression (PLS) to better account for the specific asymmetric distribution of the biomarkers; and 3) to create robust models by merging large data sets from 5 international or national projects. Benefiting from this international collaboration, the data set comprised a total of 9,143 milk samples from 3,758 cows located in 589 herds across 10 countries and represented 7 breeds. The samples were analyzed by reference chemistry for biomarker contents while the MIR analyses were performed on 30 instruments from different models and brands, with spectra harmonized into a common format. Four quantitative methodologies were evaluated to address the strongly skewed distribution of some biomarkers. PLS was used as the reference basis, and compared with a random modification of distribution associated with PLS (Random-downsampling-PLS), an optimized modification of distribution associated with PLS (KennardStone-downsampling-PLS) and Support Vector Machine (SVM). When the ability of MIR to predict biomarkers was too low for quantification, different qualitative methodologies were tested to discriminate low vs high values of biomarkers. For each biomarker, 20% of the herds were randomly removed within all countries to be used as the validation data set. The remaining 80% of herds were used as the calibration data set. In calibration, the 3 alternative methodologies outperform the PLS performances for the majority of biomarkers. However, in the external herd validation, PLS provided the best results for isocitrate, glucose-6P, free glucose and LDH (R2v = 0.48, 0.58, 0.28, and 0.24). For other molecules, PLS-Random-downsampling and PLS-KennardStone-downsampling outperformed PLS in the majority of cases, but the best results were provided by SVM for citrate, BHB, acetone, NAGase and progesterone (R2v = 0.94, 0.58, 0.76, 0.68, and 0.15). Hence, PLS and SVM based on the entire data set provided the best results for normal and skewed distributions, respectively. Complementary to the quantitative methods, the qualitative discriminant models enabled the discrimination of high and low values for BHB, acetone, and NAGase with a global accuracy around 90%, and glucose-6P with an accuracy of 83%. In conclusion, MIR spectra of milk can enable quantitative screening of citrate as a biomarker of energy deficit and discrimination of low and high values of BHB, acetone, and NAGase, as biomarkers of ketosis and mastitis. Finally, progesterone could not be predicted with sufficient accuracy from milk MIR spectra to be further considered. Consequently, MIR spectrometry can bring valuable information regarding the occurrence of energy deficit, ketosis and mastitis in dairy cows, which in turn have major influences on their fertility and survival.

AB - At the individual cow level, sub-optimum fertility, mastitis, negative energy balance and ketosis are major issues in dairy farming. These problems are widespread on dairy farms and have an important economic impact. The objectives of this study were: 1) to assess the potential of milk Mid Infrared (MIR) spectra to predict key biomarkers of energy deficit (citrate, isocitrate, glucose-6P, free glucose), ketosis (BHB and acetone), mastitis (NAGase and LDH), and fertility (progesterone); 2) to test alternative methodologies to partial least square regression (PLS) to better account for the specific asymmetric distribution of the biomarkers; and 3) to create robust models by merging large data sets from 5 international or national projects. Benefiting from this international collaboration, the data set comprised a total of 9,143 milk samples from 3,758 cows located in 589 herds across 10 countries and represented 7 breeds. The samples were analyzed by reference chemistry for biomarker contents while the MIR analyses were performed on 30 instruments from different models and brands, with spectra harmonized into a common format. Four quantitative methodologies were evaluated to address the strongly skewed distribution of some biomarkers. PLS was used as the reference basis, and compared with a random modification of distribution associated with PLS (Random-downsampling-PLS), an optimized modification of distribution associated with PLS (KennardStone-downsampling-PLS) and Support Vector Machine (SVM). When the ability of MIR to predict biomarkers was too low for quantification, different qualitative methodologies were tested to discriminate low vs high values of biomarkers. For each biomarker, 20% of the herds were randomly removed within all countries to be used as the validation data set. The remaining 80% of herds were used as the calibration data set. In calibration, the 3 alternative methodologies outperform the PLS performances for the majority of biomarkers. However, in the external herd validation, PLS provided the best results for isocitrate, glucose-6P, free glucose and LDH (R2v = 0.48, 0.58, 0.28, and 0.24). For other molecules, PLS-Random-downsampling and PLS-KennardStone-downsampling outperformed PLS in the majority of cases, but the best results were provided by SVM for citrate, BHB, acetone, NAGase and progesterone (R2v = 0.94, 0.58, 0.76, 0.68, and 0.15). Hence, PLS and SVM based on the entire data set provided the best results for normal and skewed distributions, respectively. Complementary to the quantitative methods, the qualitative discriminant models enabled the discrimination of high and low values for BHB, acetone, and NAGase with a global accuracy around 90%, and glucose-6P with an accuracy of 83%. In conclusion, MIR spectra of milk can enable quantitative screening of citrate as a biomarker of energy deficit and discrimination of low and high values of BHB, acetone, and NAGase, as biomarkers of ketosis and mastitis. Finally, progesterone could not be predicted with sufficient accuracy from milk MIR spectra to be further considered. Consequently, MIR spectrometry can bring valuable information regarding the occurrence of energy deficit, ketosis and mastitis in dairy cows, which in turn have major influences on their fertility and survival.

KW - Fourier transform mid-infrared spectrometry

KW - fertility

KW - ketosis

KW - mastitis

KW - negative energy balance

KW - Isocitrates

KW - Acetylglucosaminidase

KW - Citrates

KW - Glucose

KW - 3-Hydroxybutyric Acid

KW - Cattle Diseases

KW - L-Lactate Dehydrogenase

KW - Animals

KW - Cattle

KW - Biomarkers

KW - Mastitis/veterinary

KW - Female

KW - Ketosis/diagnosis

KW - Progesterone

KW - Citric Acid

KW - Acetone

KW - Milk

UR - http://www.scopus.com/inward/record.url?scp=85186745391&partnerID=8YFLogxK

U2 - 10.3168/jds.2023-23843

DO - 10.3168/jds.2023-23843

M3 - Journal article

C2 - 37863287

SN - 0022-0302

VL - 107

SP - 1669

EP - 1684

JO - Journal of Dairy Science

JF - Journal of Dairy Science

IS - 3

ER -

Prediction of key milk biomarkers in dairy cows through milk mid-infrared spectra and international collaborations

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this