First Lactation Milk Yield Predicted by the Heifer’s Growth Curve Derivatives
, Mónica Cervantes-Jiménez 1,*, Vicente Lemus-Ramírez 2, José Guadalupe García-Muñiz 3 and Adolfo Kunio Yabuta Osorio 2
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsRecommendations:
Review the manuscript and add missing citations to specific paragraphs in the introduction. Additionally, there are editorial errors that have been highlighted for easy identification within the manuscript, including the following:
Lines 19 - 20: This data also predicts …
Line 21: … weight change rate
Line 25: … modelled …
Line 27: The average deviance of fitted models represented the goodness of fit.
Line 44: … associated whit well-grown …
Line 47: … curve, the heifer’s …, … during the rearing …
Lines 53 – 56: There are missing bibliographic citations related to the statement. “Fourier series consist of sums of trigonometric functions and used to represent oscillatory components in a time series.”
Line 58: … points of inflexion at …
Lines 59 – 60: … second derivative or higher-order derivative.
Line 69: … lower weights had …
Line 78: … lactation could be …
Line 80: … in the modelling, particularly …
Line 95: … were fed whole …
Line 99: … had altered …
Lines 101 – 102: … as a single …
Line 106: … was modelled using …
Line 109: … the modelling instead …
Lines 116 – 118: It is recommended to replace it with “After calving, 72 to 94 daily records were collected using proportional meters from Waikato LTD Co., New Zealand”.
Line 124: … Modelling …
Line 129: … and 2D …
Line 134: … improving modelling compared …
Line 144: Modelling …
Line 146: … the modelling results …
Line 145: … however, no model showed significance …
Line 151: Replace "Many times," with Often. …
Line 155: … considered a better …
Lines 159 – 160: … was also …
Line 164: … importance of variable interpretability
Line 168: … For the heifer’s …
Line 171: … performed to control …, … a level significance of …
Line 184: … lower at six months …
Line 190: … variables than when …
Line 191: Interestingly, …
Line 192: … Figures ….
Line 195: … were unavailable …
Lines 196 – 197: It is recommended to replace it with “The r2 was consistently close to or slightly higher than 0.6 from three to twelve months of age when all variables were included”.
Line 199: Modelling, …
Line 200: Review, I understand that the term "breeding" or "cattle rearing" is more commonly used than just "rearing".
Line 220: … modelling …
Line 228: Generally, …
Line 234: … could offer insight into how …
Line 236: … in an XGBoost …
Line 237: …variables had a negative impact on L … or …variables hurt on L …
Line 252: The importance …, … model is in …
Line 254: … greater magnitude of …
Line 257: … heifer ages, …
Line 263: … at six months …
Lines 264 – 266: It is recommended to replace it with “In a preliminary study, we used multiple regression analysis to identify non-significant relationships between heifer weight and L as the main effect”.
Lines 268 -269: … the database was used limited …
Line 272: …. and parentage, …
Line 278: When relying on this …
Lines 279 – 280: … and the pregnancy is confirmed.
Lines 284 – 285: … r2; contributes to our hypothesis that P, 1D, and 2D can ...
Line 293: … modelling …
Line 296: The modelling …
Line 299: … limit is exceeded …, … the model will fail …
Line 305: Perhaps, …
Line 308: The best …,
Line 308 -309: … 2D, particularly …
Line 312: … yield, especially when the modelling, …
Comments on the Quality of English LanguageIn general, the expressions used are clear. Some meanings need to be corrected and an effort should be made to enhance some phrases where connectors are misused or missing. These have been highlighted, for the most part.
Author Response
Dear reviewer,
We greatly appreciate your comments, please consider a new version incorporating your suggestions. Our responses to each comments are indicated in boldface as follows:
Lines 19 - 20: This data also predicts
This sentence has been replaced
Line 21: … weight change rate
This sentence has been replaced
Line 25: … modelled …
This word and another similar have been replaced.
Line 27: The average deviance of fitted models represented the goodness of fit.
This sentence has been replaced
Line 44: … associated whit well-grown …
This sentence has been replaced
Line 47: … curve, the heifer’s …, … during the rearing …
This sentence has been replaced
Lines 53 – 56: There are missing bibliographic citations related to the statement. “Fourier series consist of sums of trigonometric functions and used to represent oscillatory components in a time series.”
We add some citations to this paragraph to support the ideas as follows “Fourier series are functions that consist of sums of trigonometric functions and are used to represent oscillatory components in a time series [6]. Transient or instantaneous growth rates may reveal growth spurts due to environmental factors [7]. As such, the instantaneous ADG is the first derivative of a growth curve at a given time [8]. The first derivative of nonlinear functions can have maximum or minimum values representing points of inflexion at specific times. Because a derivative is a function, it can be derived to obtain a second derivative or higher-order derivatives [9].
Line 58: … points of inflexion at …
This sentence has been replaced.
Lines 59 – 60: … second derivative or higher-order derivative.
This sentence has been replaced
Line 69: … lower weights had …
This sentence has been replaced
Line 78: … lactation could be …
This sentence has been replaced
Line 80: … in the modelling, particularly …
This sentence has been replaced
Line 95: … were fed whole …
This sentence has been replaced
Line 99: … had altered …
This sentence has been replaced
Lines 101 – 102: … as a single …
This sentence has been replaced
Line 106: … was modelled using …
This sentence has been replaced
Line 109: … the modelling instead …
This sentence has been replaced
Lines 116 – 118: It is recommended to replace it with “After calving, 72 to 94 daily records were collected using proportional meters from Waikato LTD Co., New Zealand”.
This sentence was replaced by “After calving, between 72 to 94 daily records were collected using proportional milk meters from Waikato LTD Co., New Zealand”
Line 124: … Modelling …
This word has been replaced.
Line 129: … and 2D …
This sentence has been replaced
Line 134: … improving modelling compared …
This sentence has been replaced
Line 144: Modelling …
This word has been replaced.
Line 146: … the modelling results …
This sentence has been replaced
Line 145: … however, no model showed significance …
This sentence has been replaced
Line 151: Replace "Many times," with Often. …
This sentence has been replaced
Line 155: … considered a better …
This sentence has been replaced
Lines 159 – 160: … was also …
This sentence has been replaced
Line 164: … importance of variable interpretability
We replace the subtitle by “Importance of variables and model interpretability”, because the interpretability refers to the models not to variables importance.
Line 168: … For the heifer’s …
This sentence has been replaced
Line 171: … performed to control …, … a level significance of …
This sentence has been replaced
Line 184: … lower at six months …
This sentence has been replaced
Line 190: … variables than when …
This sentence has been replaced
Line 191: Interestingly, …
This word has been replaced
Line 192: … Figures ….
This word has been replaced
Line 195: … were unavailable …
This sentence has been replaced
Lines 196 – 197: It is recommended to replace it with “The r2 was consistently close to or slightly higher than 0.6 from three to twelve months of age when all variables were included”.
This sentence has been replaced
Line 199: Modelling, …
This word has been replaced
Line 200: Review, I understand that the term "breeding" or "cattle rearing" is more commonly used than just "rearing".
We appreciate this precision; the term "heifers rearing" was changed to refer specifically to the moment of rearing, excluding artificial insemination in breeding.
Line 220: … modelling …
This word has been replaced
Line 228: Generally, …
This word has been replaced
Line 234: … could offer insight into how …
This sentence has been replaced
Line 236: … in an XGBoost …
This sentence has been replaced
Line 237: …variables had a negative impact on L … or …variables hurt on L …
These sentences have been replaced
Line 252: The importance …, … model is in …
This sentence has been replaced
Line 254: … greater magnitude of …
This sentence has been replaced
Line 257: … heifer ages, …
This sentence has been replaced
Line 263: … at six months …
This sentence has been replaced
Lines 264 – 266: It is recommended to replace it with “In a preliminary study, we used multiple regression analysis to identify non-significant relationships between heifer weight and L as the main effect”.
This paragraph has been replaced
Lines 268 -269: … the database was used limited …
This sentence has been replaced
Line 272: …. and parentage, …
This sentence has been replaced
Line 278: When relying on this …
This sentence has been replaced
Lines 279 – 280: … and the pregnancy is confirmed.
This sentence has been replaced
Lines 284 – 285: … r2; contributes to our hypothesis that P, 1D, and 2D can ...
This sentence has been replaced
Line 293: … modelling …
This word has been replaced
Line 296: The modelling …
This sentence has been replaced
Line 299: … limit is exceeded …, … the model will fail …
This sentence has been replaced
Line 305: Perhaps, …
This word has been replaced
Line 308: The best …
This word has been replaced
Line 308 -309: … 2D, particularly …
This word has been replaced
Line 312: … yield, especially when the modelling, …
This word sentence been replaced
Best Regards
Mónica Cervantes-Jiménez
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsManuscript ID dairy-2946514
The paper entitled "First lactation milk yield predicted by the heifer's growth curve derivatives" investigates the use of some indicators to predict the milk yield at first lactation. The mathematical function of the 'first derivative' of the growth curve represents the rate of change in weight at a given time, while the 'higher order derivatives' of a biological process, such as growth, can be studied to predict the outcome of another process, such as lactation. Seventy-eight heifers (Holstein-Friesian, Jersey and their crosses) reared on irrigated pasture with rotational grazing were monitored every three months (from 3 to 21 months of age) for body weight (P), the first (1D) and second (2D) derivatives of an individually modelled Fourier function. The other two variables were age at actual insemination (AI) and season of the year in which the heifer was born (E). Models were tested using 85% of the records and fit was assessed using the remaining 15% of data. The best models predicted milk yield during entire lactation (L) using heifer data at six months of age (r2 = 0.62) and the importance of the variables in the model was 35, 28, 21 and 16% for 1D, AI, 2D and P respectively. The authors conclude that using this type of models, it would be possible to select and eliminate surplus heifers at an early stage, thus optimising breeding costs.
This paper is interesting, well written and in line with the aim of the journal, but it needs a few minor additions, which I report below, line by line:
Line 87: check thi sentence “was in a year-round milk sys-tem of Holstein-Friesian,”
Lines 117-118: “obtained using proportional meters (Waikato LTD Co., New Zealand).” Maybe “obtained using proportional milk meters (Waikato LTD Co., New Zealand).
Lines 220-227: table 1 and table 2. Check the superscript number 1, 2, the number 0..08 and 0.0151
I think the mean and the standard error of the mean need to be reorganised better, perhaps where the age is given.
Lines 240-251: Figure 3 should be enlargement.
Line 296-306: The whole of the paragraph should be re-written in a more appropriate form.
Lines 307-316: 5. Conclusions. The practical and economic benefits to farmers that emerged from this study must be added.
Comments on the Quality of English Language
Minor change of English language is required.
Author Response
Dear reviewer,
We greatly appreciate your comments. Please consider a new version that incorporates your suggestions. Our responses to each comment are indicated in boldface as follows:
Line 87: check this sentence “was in a year-round milk sys-tem of Holstein-Friesian,”
This sentence has been replaced
Lines 117-118: “obtained using proportional meters (Waikato LTD Co., New Zealand).” Maybe “obtained using proportional milk meters (Waikato LTD Co., New Zealand).
This sentence was replaced by “After calving, between 72 to 94 daily records were collected using proportional milk meters from Waikato LTD Co., New Zealand”
Lines 220-227: table 1 and table 2. Check the superscript number 1, 2, the number 0..08 and 0.0151.I think the mean and the standard error of the mean need to be reorganised better, perhaps where the age is given.
We changed the title of Tables 1 and 2 to specify standard and mean values.
Lines 240-251: Figure 3 should be enlargement.
Figure 3 was enlarged and enhanced in quality. The original file is attached separately from the manuscript.
Line 296-306: The whole of the paragraph should be re-written in a more appropriate form.
The paragraph was reworded as follows:
“The modeling could be improved by increasing the execution time and reducing the number of algorithms included in the construction phase. It is possible that the solution space was not thoroughly explored due to the limited execution time of 200 seconds for completing a model. If the time limit was exceeded, the model would fail and be discarded.
Deep learning was the algorithm that provided the best solution in most runs. However, there were two main variants depending on how the model's hyperparameters were optimized. The traditional method involved using a Cartesian grid, while the alternative method involved using a random search grid [21]. The average deviance for the Cartesian grid was 377,733, while for the random search grid it was 571,712. For XGBoost, something similar happened with deviance values of 491,663 and 748,663. Possibly, the random search for hyperparameters was not efficient within the allocated time and the optimization of hyperparameters should be restricted to Cartesian grid search.”
Lines 307-316: 5. Conclusions. The practical and economic benefits to farmers that emerged from this study must be added.
Conclusion has been modifying as follows:
“The best prediction of first lactation yield resulted from models using 1D and 2D, particularly when using data from six to twelve months of the heifer’s growth period. For these models, 1D was the most important variable at six and nine months, while 2D was the most important at three and twelve months of age. The derivatives of the heifer growth curve were informative in predicting milk yield, especially when the modelling did not incorporate the age at effective insemination. Models incorporating the derivatives of the growth curve would help producers make management decisions for heifers at early ages, optimizing resource investment.”
Best Regards
Mónica Cervantes-Jiménez
Author Response File: Author Response.pdf
