Algorithms for Detecting Cattle Diseases at Early Stages and for Making Diagnoses and Related Recommendations
, Alexei S. Dorokhov 1, Fedor E. Vladimirov 1,*, Igor M. Dovlatov 1 and Konstantin S. Lyalin 2Round 1
Reviewer 1 Report
This paper proposes algorithms for cattle diseases detecting at early stages and for diagnosis & recommendations making. The idea in this paper is intersting, but I still have some comments as follows:
- The Simple Summary is generally not needed before abstract.
- The corresponding equations should be added in the paper to help the readers understand the detailed procedure rather than explain with words.
- The legends in the figures are too small to be clearly seen.
- The authors should emphasis where did they obtain the data from, as they are dealing with animal data.
Author Response
Point 1: The Simple Summary is generally not needed before abstract.
Response 1: Good afternoon! We are grateful to you for your detailed review and recommendations. We appreciate it very much as this way, you make a valuable contribution to the science development in general and to our work improvement in this scientific direction in particular.
The section “Simple Summary” was added at the previous editor’s suggestion; however, according to your recommendation, it has been deleted as we agree with your opinion.
Point 2: The corresponding equations should be added in the paper to help the readers understand the detailed procedure rather than explain with words.
Response 2: We have added some formulas to the text so that to identify the correlation dependencies of the cross-correlation function.
Point 3: The legends in the figures are too small to be clearly seen.
Response 3: The small font size on the graphs is due to the fact that the program uploads them in this format; unfortunately, we see no way to change the font, unless by use of a special graphic editor. Anyway, we have increased the font size as much as possible.
Point 4: The authors should emphasis where did they obtain the data from, as they are dealing with animal data.
Response 4: In the text, the names and the locations have been indicated of the farms, where the data array was obtained on the physiological parameters of the cows. In addition, earlier, together with the article, we sent the data archive on each of the cows in form of tables in the exel format.
Author Response File: 
Author Response.docx
Reviewer 2 Report
Figure 1/ are out of place here. in the methods part only the "how was worked" should be described and no results should be presented or discussed. I ask for a clean separation of the topics!
The results section does not contain any relevant results to the question!
The section discussion appears here more than the actual result chapter. Since it goes according to the set up research question around the algorithm and less around the parameters at itself.
The chapter discussion should set the results of the investigation (from the result chapter) with the context of other scientific work and discuss. that is missing here completely!
Author Response
Point 1-3: Figure 1 are out of place here. In the methods part only the "how was worked" should be described and no results should be presented or discussed. I ask for a clean separation of the topics! The results section does not contain any relevant results to the question! The section discussion appears here more than the actual result chapter. Since it goes according to the set up research question around the algorithm and less around the parameters at itself.
Response 1: Good afternoon! We are grateful to you for your detailed review and recommendations. We appreciate it very much as this way, you make a valuable contribution to the science development in general and our work improvement in this scientific direction in particular.
As for the recommendations. As a matter of fact, the previous reviewers asked us to move these graphs, the table and the text related to them to the “Discussion” section. But we agree with your point of view because initially we ourselves adhered to it. So now, we have moved the information back to the relevant sections.
Point 2: The chapter discussion should set the results of the investigation (from the result chapter) with the context of other scientific work and discuss. That is missing here completely!
Response 2: Due to the fact that the article is aimed mainly at the results reporting, we decided to do without the discussion section.
Author Response File: Author Response.docx
Round 2
Reviewer 2 Report
Comments on the second version:
The changes currently made to the structure are understandable and contribute to better readability and understanding. It would have been nice if the changes were made in the change mode as usual.
The omission of the discussion here is not understandable. The discussion is an essential part of every scientific article. Here the own findings should be put in relation to other scientific work and so the results can be evaluated and classified. This is especially important for readers who are not deeply rooted in the subject area. It also serves to answer the research question posed in the introduction or to show the degree of fulfillment of the research thesis.
In favor of an orderly discussion, the conclusion can be shortened considerably. Here, a short and concise conclusion about the presented work should be written in a few sentences.
Author Response
Point 1: It would have been nice if the changes were made in the change mode as usual.
Response 1: Good afternoon! We are grateful to you for your constructive recommendations. We have included the function of changes tracking in the program.
Point 2: The omission of the discussion here is not understandable. The discussion is an essential part of every scientific article. Here the own findings should be put in relation to other scientific work and so the results can be evaluated and classified. This is especially important for readers who are not deeply rooted in the subject area. It also serves to answer the research question posed in the introduction or to show the degree of fulfilment of the research thesis.
Response 2: We have moved some of the conclusions into the “Discussion” section. Also we have added some sources that confirm our results somehow:
Discussion
In the existing algorithms for deviations detecting in the indicators of the cattle‘s physiological state, no checking-up is foreseen for locations of the animals under study-ing. That is why we conducted the comparative analysis of the motor activity index of both sick and healthy dairy cows. The cows of the 2nd group had deviations in frequency and amplitude of their legs’ oscillatory (pendulum) movements when walking (the sinus-oid was ahead of the normal sinusoid or lagged behind it). Meanwhile at the cows from the 1st group, everything was within the normal range.
The animals’ activity index in the 1st group was on average not less than 10 units, while in the 2nd group below 5 units. The motor activity lowest duration was observed at the sick animals from the 2nd group: their rest time made 801 minutes/day, which was longer on average by 47 minutes (by 6%) than at the normal cows. The differences between the groups are statistically significant (p<0,05).
Davison et al. (2020) came to the conclusion that the due consideration of the loca-tions was necessary for the deviations identification in milk cows’ physiological state. It is important because, for example, the heat stress affects significantly the body temperature and the locomotive activity of animals; so not always their deviations would be associated with diseases [28]. Ramón-Moragues et al. (2021) revealed that under a heat stress, cows had a noticeable decrease in locomotive activity and body temperature; besides, it was a violation of the animals’ daily routine [29]. Rutherford et al. (2019) came to the conclusion that the rumen temperature varied greatly depending on the animals’ location. For exam-ple, when being in a pasture, due to the predominance of succulent feeds in the diet, the pH level was below normal, while the rumen temperature wass higher than in norm [30]. Exactly that is why, when creating algorithms, it is necessary to take into account the an-imals’ location. .
As for the correlations identification between the parameters of the cattle’s physiolog-ical state, we confirmed successfully the results obtained in the previous studies.
Additionally, there were obtained the following data: the average value of the linear correlation coefficients for the rumen temperature and the locomotive activity made 0.199 (=0.088), and between the level of locomotive activity and the rumen pH -0.222 (= 0.117). Between the pH level and the level of locomotive activity of the animals, the linear correla-tion coefficient reached -0.519, which suggests the presence of a stable connection. The correlation between the rumen temperature and the locomotive activity level reached -0.424, and between the rumen temperature and the rumen pH level up to -0.470, which can suggest a stable connection. The analysis of the cross-correlation function showed a noticeable connection between the level of locomotive activity and the air temperature (0.622); moderate correlations were observed between the rumen temperature and the lo-comotive activity (0.33), between the rumen pH level and the air temperature (-0.36) and between the level of locomotive activity and the relative humidity (-0.347).
The similar results speaking for the high correlation between the studied cattle’s physiological parameters were obtained by Alzahal et al (2008) and then again by Anta-naitis et al (2016). The research results by Dorokhov et al (2021) showed the negative sta-tistically significant correlation (r = -0,932, P <0,01) between the rumen temperature and its pH level [28, -33].
Soriani et al. revealed a significant correlation between the rectal temperature and the temperature-humidity index (THI) (r = -0,41; P < 0,01). In this regard, it is very important to take the factors of the microclimate into account, when developing algorithms for diseases detection [34].
Based on the scientific literature analysis, 3 main groups of cattle diseases were iden-tified: infectious, invasive (parasitic) and nervous (the nervous system diseases). (Bessa-rabov et al. 2007; Isimbekov et al. 2016). The authors described the main criteria (Fig. 4) as a basis, how they can be detected at the initial stages of their development before the ap-pearance of the main clinical signs; also recommendations were given for their treatment [35, 36].
As for the existing monitoring systems, they do not allow detecting diseases at their early stages due to the outdated algorithms used. Wang et al. (2018) highlighted, for ex-ample, that a level of feeding is difficult for assessment if to use an accelerometer only; but its combining with data from other sensors would improve performance by as much as 20%. The criteria resulting from this result can be applied to the rest of the measured pa-rameters, too [37].
Point 3: In favor of an orderly discussion, the conclusion can be shortened considerably. Here, a short and concise conclusion about the presented work should be written in a few sentences.
Response 3: According to your recommendation, we cut down the conclusion. We really hope that we did everything correctly in line with the requirements.
Conclusions
Algorithms have been developed able to detect signs of nervous, parasitic and infec-tious diseases at cattle and to issue recommendations based on changes in the cows’ physiological parameters.The developed algorithms are based on the conducted both an-alytical research and field tests. The algorithms were obtained based on the comparison of cows having similar reproductive status and physiological parameters, being kept in the same conditions and divided into 2 groups: sick and healthy ones.
Currently, the developed algorithms are used in the software called “Systems of in-ternal monitoring of cattle’s physiological state” («Federal Scientific Agroengineering Centre VIM», Moscow, Russian Federation).
Author Response File: Author Response.docx
