Estimation of Genetic Parameters and Weighted Single-Step Genome-Wide Association Study for Indicators of Colostrum Quality in Chinese Holstein Cattle

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Respected researchers: This is a work that I found interesting, well thought out and well elaborated. However, I have several doubts, it seems to be a work that replicates what has been done by other researchers in other latitudes, with similar results. I wonder if it would be possible to go a little further from these results: to propose the analysis of those animals that stood out for having excelled in the Brix measurement, with measurements above 28%. In all three groups, their results show animals that are out of the average. Why not check if in these animals there are some genes that could be different and possible candidates to show animals that always produce colostrum of excellent quality. These animals could be reproduced and make an “animal” bank of colostrum of excellent quality. In this way these cows would be the producers of colostrum for all newborns, it would not be necessary to measure in each cow if its colostrum is good or not.

As the discussion is based on the same assertions of other authors. One could speculate a bit about what is going on with colostrum animals with outstanding IgG concentrations. Anyway, it is a matter of finding candidate genes, but the approach could be from another angle.

Comments on the Quality of English Language

I consider the article to be easy to follow and the English used to be university-level English.

Author Response

Comments 1: Respected researchers: This is a work that I found interesting, well thought out and well elaborated. However, I have several doubts, it seems to be a work that replicates what has been done by other researchers in other latitudes, with similar results. I wonder if it would be possible to go a little further from these results: to propose the analysis of those animals that stood out for having excelled in the Brix measurement, with measurements above 28%. In all three groups, their results show animals that are out of the average. Why not check if in these animals there are some genes that could be different and possible candidates to show animals that always produce colostrum of excellent quality. These animals could be reproduced and make an “animal” bank of colostrum of excellent quality. In this way these cows would be the producers of colostrum for all newborns, it would not be necessary to measure in each cow if its colostrum is good or not.

As the discussion is based on the same assertions of other authors. One could speculate a bit about what is going on with colostrum animals with outstanding IgG concentrations. Anyway, it is a matter of finding candidate genes, but the approach could be from another angle.

Response 1: Thank you. We analyzed the data with Brix > 28% as you suggested, but unfortunately, the sample was too small for accurate results. Only 7.62% of the records met this threshold as we added at L172.

The WssGWAS analysis showed large standard errors due to the smaller sample size. For Brix1, Brix2, and Brix3, the estimated heritabilities were 0.24±0.07, 0.18±0.07, and 0.33±0.14, respectively. Although the heritabilities were still similar to those estimated using the entire dataset, the SE was too large. Genetic parameter estimation was not precise enough to draw firm conclusions and conduct subsequent GWAS analyses. More individual genotypes and high Brix phenotypes are needed for further research (L355-356).

Some of our work, such the estimate of genetic parameters, was similar to other findings. However, we also have some special consequences as L364-366 shows that we discovered that the brix from the first parity genetically different from that from the later parities.

Reviewer 2 Report

Comments and Suggestions for Authors

Ma et al. applied weighted single-step genome-wide association studies (WssGWAS) utilizing 58,338 Chinese Holstein cattle colostrum phenotype records to identify cattle genetic factors associated with cow colostrum quality traits. The heritability of the trait was also estimated to be moderate. Finally, two promising candidate genes (CNR1 and ZXDC) were recommended by the authors. This study is of great importance for cattle herd health improvement by providing potential clues to elucidate the complicated colostrum trait.

In general, it’s an interesting study with sound methods and I’m impressed by the complexity and the questions the authors have addressed to investigate the genetic basis of colostrum trait. There are two major points I’d like the authors to address:

1 Since the phenotypes were collected in 37 farms at different time points, the management and weather can affect the phenotype levels, how did you exclude these influences in the GWAS studies?

2 Have you tried to conduct GWAS using colostrum phenotypes collected from same farm, same year or same location with similar weather conditions? Phenotypes in more defined levels could improve GWAS ability to detect genetic factors associated with the traits.

Some minor comments/questions to be discussed with the authors:

L48, “Bielmann et al., 2010” reference was not properly inserted.

L56, [2] should be inserted right after “higher”.

L110, Genotype information needs to be classified in detail like how the 1,855 cows were genotyped (54K, 770K or whole genome sequence?). Why minor allele frequency threshold was set as 0.01 not 0.05?

L159, “ggplot2” reference needs to be cited. L163, reference of ARS-UCD1.2. needs to be cited.

L243, “table 5” needs to change to Table 5. Same for L266 “figure 3”.

Comments on the Quality of English Language

Sentence like L274 "Many enriched genes related to cell cycle regulation, intracellular protein transport, cytoskeletal, and transcriptional control were enriched." is  a little bit confused, it is better to change the last word "enriched" to "revealed".

Author Response

Comments 1: Since the phenotypes were collected in 37 farms at different time points, the management and weather can affect the phenotype levels, how did you exclude these influences in the GWAS studies?

Response 1: Thank you. Farm-area-scale was included into the WssGWAS model to explain farm-related influences. Please see L304-312 for our explanation.

 

Comments 2: Have you tried to conduct GWAS using colostrum phenotypes collected from same farm, same year or same location with similar weather conditions? Phenotypes in more defined levels could improve GWAS ability to detect genetic factors associated with the traits.

Response 2: Thank you for your suggestion. To lessen environmental variation, we attempted to concentrate on only one area (Beijing). The heritability estimates for the phenotypes were reasonably consistent throughout the various datasets, according to our analysis of the data. In Beijing dataset, Brix1 is 0.32 ± 0.06, Brix2 is 0.23 ± 0.30, and Brix3 is 0.23 ± 0.04, which is very close to the results using the entire data set. However, we ran into some difficulties with genetic correlations. Genetic correlation between Brix1 and Brix2 was 0.32 ± 0.15, and the other two didn’t converge properly. We also conducted WssGWAS. There are many overlapping genomic regions (18, 20, and 27) between Beijing dataset and original dataset, and the regions in Beijing dataset included the candidate genes CNR1 and ZXDC, too. Briefly, the results didn’t differ much from those using the original dataset. More importantly, limiting the analysis to just one region meant we would have lost a lot of data, which would have significantly reduced the power of the study. So, while we did explore this approach, we felt the broader dataset provided more robust and reliable results for the current research.

 

Comments 3: L48, “Bielmann et al., 2010” reference was not properly inserted.

Response 3: Done.

 

Comments 4: L56, [2] should be inserted right after “higher”.

Response 4: Done.

 

Comments 5: L110, Genotype information needs to be classified in detail like how the 1,855 cows were genotyped (54K, 770K or whole genome sequence?). Why minor allele frequency threshold was set as 0.01 not 0.05?

Response 5: Thank you. the chip is 150K. please see L108-109 for our revision.

Regarding the MAF threshold, we apologize for the earlier confusion. The threshold used in our analysis was 0.1, not 0.01. To get more robust results, we only kept common SNPs (MAF > 0.1) in this study.

 

Comments 6: L159, “ggplot2” reference needs to be cited. L163, reference of ARS-UCD1.2. needs to be cited.

Response 6: Done.

 

Comments 7: L243, “table 5” needs to change to Table 5. Same for L266 “figure 3”.

Response 7: Done.

 

Response to Comments on the Quality of English Language

 

Point 1: Sentence like L274 "Many enriched genes related to cell cycle regulation, intracellular protein transport, cytoskeletal, and transcriptional control were enriched." is a little bit confused, it is better to change the last word "enriched" to "revealed".

Response 1: Thank you. We revised the whole sentence to make it clearer. Please see L280-282.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Thank you for reviewing my suggestion. It is clear that this research must be carried out as it was in ancient genetics, selecting specimens with the best characteristics. This is how changes were achieved in breeds such as Holstein, Hereford, etc.

Author Response

Comments1: Thank you for reviewing my suggestion. It is clear that this research must be carried out as it was in ancient genetics, selecting specimens with the best characteristics. This is how changes were achieved in breeds such as Holstein, Hereford, etc.

Response 1: Thank you for your suggestion. We will continue to accumulate data in our later research, hoping to select a suitable dataset.