Genome-Wide Association Study of Parasite Resistance to Gastrointestinal Nematodes in Corriedale Sheep

Round 1

Reviewer 1 Report

Gastrointestinal nematodes(GINs) are one of the most economically important parasites of small ruminants and a major animal health concern in many regions of the world. This manuscript were identified some novel candidate genes of Corriedale sheep by GWAS.

1. The background of the experimental materials is not clear.

I can’t found that how the author select the animals of 454, 702 and 375 population. Are there any overlapping individuals among the three groups?

2. The method is not fully explained. The GO terms analysis didn’t description.

3. The result describe is too simple. And it did not show some traits related FEC records and more informations of 454, 702 and 375 population, respectively.

4. The discussion is definitely not complete, which requires further researches and exploration. 

The discussion is not just a further description of the results. Moreover, it is necessary to analyze why it comes different results with different densities SNP chips. And to investigate the correlation between these results.

Author Response

Response to Reviewer 1 Comments:

Gastrointestinal nematodes (GINs) are one of the most economically important parasites of small ruminants and a major animal health concern in many regions of the world. This manuscript identified some novel candidate genes of Corriedale sheep by GWAS.

1. The background of the experimental materials is not clear.

I can’t found that how the author select the animals of 454, 702 and 375 population. Are there any overlapping individuals among the three groups?

Answer: Animals with FEC data and complete pedigree information were selected for genotyping, mostly lambs and all sires that were available related to those lambs (added in lines 111-113, 116-117). We added a Venn diagram to show the overlapping individuals among the three groups (Figure 1).

2. The method is not fully explained. The GO terms analysis didn’t description.

Answer: GO terms analysis description was added in lines 219-223.

3. The result describe is too simple. And it did not show some traits related FEC records and more informations of 454, 702 and 375 population, respectively.

Answer: No traits other than FEC were included in this study. More information on the 454, 702 and 375 population was added as suggested (lines 119-125).

4. The discussion is definitely not complete, which requires further researches and exploration.

The discussion is not just a further description of the results. Moreover, it is necessary to analyze why it comes different results with different densities SNP chips. And to investigate the correlation between these results.

Answer: a paragraph comparing results from the three different approaches was added as suggested (lines 451-460).

Reviewer 2 Report

The manuscript entitled "Genome wide association study of parasite resistance to gastrointestinal nematodes in Corriedale sheep" by Carracelas and colleagues aimed at detecting genomic regions related to gastro-intestinal nematodes in Corriedale sheep. The authors generated their own SNP (using different SNP panels) and phenotypic data which is valuable. Also, the manuscript is interesting however there are some major and minor concerns which are summarized below:

Major:

-        As the authors mentioned in the manuscript, different SNP panels including 170, 507 and 50K were applied. Also, authors used ssGWAS approach for genomic mapping regions associated with the phenotype. Actually, I am not sure how much a genomic relationship matrix based on 170 and 507 SNPs could be realistic as the authors applied for the estimation of GEBVs in ssGWAS. This low number of SNPs can reduce correlation between elements of pedigree and genomic relationship matrices (diagonal and off-diagonal). This reduction can affect on tuning and blending of relationship matrices, and eventually, lead to getting the non-realistic GEBVs. Therefore, I think that the authors must provide an acceptable reason for the implementation of ssGWAS using low numbers of SNPs.

Minor

-        Line 92: How many blood samples were collected? How did you select individuals for genotyping? Randomly?

-        Please change “Genome wide association” to “Genome-wide association” in title and throughout the manuscript.

-        Line 48-52: Please describe the advantages of ssGWAS and show how this method can be valuable in comparison with multi-step GWAS (in 2-4 lines).

-        Could you please provide how many SNPs are common among different panels (170, 507 and 50K)? A Venn diagram might be helpful to show how many SNPs are common among SNP panels you used.

 

-        Line 142: You must change “Calculate” to “Estimate”. Calculate is not a correct verb for GEBV. Apply it throughout the manuscript.  

Author Response

Response to Reviewer 2 Comments:

The manuscript entitled "Genome wide association study of parasite resistance to gastrointestinal nematodes in Corriedale sheep" by Carracelas and colleagues aimed at detecting genomic regions related to gastro-intestinal nematodes in Corriedale sheep. The authors generated their own SNP (using different SNP panels) and phenotypic data which is valuable. Also, the manuscript is interesting however there are some major and minor concerns which are summarized below:

Major:

-        As the authors mentioned in the manuscript, different SNP panels including 170, 507 and 50K were applied. Also, authors used ssGWAS approach for genomic mapping regions associated with the phenotype. Actually, I am not sure how much a genomic relationship matrix based on 170 and 507 SNPs could be realistic as the authors applied for the estimation of GEBVs in ssGWAS. This low number of SNPs can reduce correlation between elements of pedigree and genomic relationship matrices (diagonal and off-diagonal). This reduction can affect on tuning and blending of relationship matrices, and eventually, lead to getting the non-realistic GEBVs. Therefore, I think that the authors must provide an acceptable reason for the implementation of ssGWAS using low numbers of SNPs.

Answer: The correlation between off-diagonal elements of A and G matrices for 170 SNPs was 0.505 and for 507 SNPs was 0.58, so both correlations are within optimal range (between 0.5 and 0.9 according to Lourenco et al. (2020). This shows that these relationships were well captured in the pedigree so the correlation between off-diagonal elements of A and G matrices is ok. However, the correlation between diagonal elements of A and G matrices, which correspond to inbreeding coefficients, is lower than expected (0.2 – 0.4), with 0.141 for 170 SNPs and 0.052 for 507 SNPs. Even though we used a complete pedigree, it was only 3 generations deep, probably with many 0 values for A in contrast with small inbreeding coefficients for G, this would explain the low correlation between diagonal elements of A and G matrices. We know this methodology is suitable for at least 50K SNPs but we wanted to try it with the small arrays since they were previously selected for their association to gastrointestinal resistance, so they are not random markers. In a ssGBLUP study performed by Vallejo et al. (2018) they found accurate genomic predictions for bacterial cold water disease resistance in rainbow trout using a 500 SNP panel and also a 70 SNP panel that included SNPs that were previously identified through GWAS. In another study, specifically in FEC trait using ssGWAS and a 12K SNP panel, several candidate genes related to immune system activation and inflammatory response were found (Berton et al., 2017).

Lourenco, D.; Legarra, A.; Tsuruta, S.; Masuda, Y.; Aguilar, I.; Misztal, I. Single-Step Genomic Evaluations from Theory to Practice: Using SNP Chips and Sequence Data in BLUPF90. Genes 2020, 11, 790, doi:10.3390/genes11070790.

Vallejo, R.L.; Silva, R.M.O.; Evenhuis, J.P.; Gao, G.; Liu, S.; Parsons, J.E.; Martin, K.E.; Wiens, G.D.; Lourenco, D.A.L.; Leeds, T.D.; Palti, Y. Accurate genomic predictions for BCWD resistance in rainbow trout are achieved using low-density SNP panels: Evidence that long-range LD is a major contributing factor. Journal of Animal Breeding and Genetics 2018, 135, 263-274, doi: 10.1111/jbg.12335.

Berton, M.P.; de Oliveira Silva, R.M.; Peripolli, E.; Stafuzza, N.B.; Martin, J.F.; Álvarez, M.S.; Gavinã, B.V.; Toro, M.A.; Banchero, G.; Oliveira, P.S.; et al. Genomic Regions and Pathways Associated with Gastrointestinal Parasites Resistance in Santa Inês Breed Adapted to Tropical Climate. Journal of Animal Science and Biotechnology 2017, 8, 73, doi:10.1186/s40104-017-0190-4.

Minor

- Line 92: How many blood samples were collected? How did you select individuals for genotyping? Randomly?

Answer: 1548 blood samples (added in line 127). Individuals with FEC data and complete pedigree information were selected for genotyping, mostly lambs and all sires that were available related to those lambs  (added in lines 111-113 and 116-117).

- Please change “Genome wide association” to “Genome-wide association” in title and throughout the manuscript.

Answer: it has been changed, as suggested.

- Line 48-52: Please describe the advantages of ssGWAS and show how this method can be valuable in comparison with multi-step GWAS (in 2-4 lines).

Answer: the advantages of ssGWAS and the comparison with multi-step GWAS were addressed, as suggested (lines 56-59).

- Could you please provide how many SNPs are common among different panels (170, 507 and 50K)? A Venn diagram might be helpful to show how many SNPs are common among SNP panels you used.

Answer: Only the 507 SNP chip has 91 SNPs in common with the 50K SNP chip, with no SNPs overlapping among the other ones (lines 145-146).

- Line 142: You must change “Calculate” to “Estimate”. Calculate is not a correct verb for GEBV. Apply it throughout the manuscript.  

Answer: it has been changed, as suggested.

Reviewer 3 Report

Authors conducted a GWAS of parasite resistance on a local sheep breed. The general topic sounds interesting and generally is important for sheep production. However there are some crucial points (especially in Material and Results)  that are not so well clarified in the text and need further attention. Please find bellow  my comments.

l. 36-39. Could you justify under what criteria?

l. 50 Omit "in this article"

l. 62 omit "in 2014"

Material and methods

line 70-81: Please specify the period that sampling was conducted  (winter, spring?) using fecal? How fecals was sampled? From animal's  intestine? In addition, please specify if any anti-parasitic treatment was implemented in animals, and if yes how long before sampling. Sampling was conducted only in females? Please specify how many females and males were used in the study. Moreover, do authors repeated the same sampling from the animals as a control?

-What was the  productive system? Extensive, intensive?

-Could you please add more information about the breed?

Statistical analysis

As the three bead chips were conducted in different number of animals, did the statistical analysis referred to the same animals and how many animals were finally tested? please specify.

line 80-81. Performance data what did included?

Please also clarify the assumptions about gene ontology that were used.

Results

It is not shown clearly if genome wide associations using different approaches are conducted with the same or different number of animals. Please clarify.

Discussion

298-300 please clarify why

line 300-301 statement it is not so clear how is related to the current  study.

-Authors should also connected their findings with the number of analyzed animals used in each bead chip (if they were not included the same animals)

-The manuscript would gain floor if a comparison approach on explaining why the differences between different bead chip approach were noted.

 

Author Response

Response to Reviewer 3 Comments:

Authors conducted a GWAS of parasite resistance on a local sheep breed. The general topic sounds interesting and generally is important for sheep production. However, there are some crucial points (especially in Material and Results) that are not so well clarified in the text and need further attention.

Please find bellow my comments.

l. 36-39. Could you justify under what criteria?

Answer: The utilization and relevance of FEC as selection criterion is now described in lines 39-41.

l. 50 Omit "in this article"

Answer: it was omitted, as suggested.   

l. 62 omit "in 2014"

Answer: it was omitted, as suggested. 

Material and methods

line 70-81: Please specify the period that sampling was conducted (winter, spring?) using fecal? How fecals was sampled? From animal's intestine? In addition, please specify if any anti-parasitic treatment was implemented in animals, and if yes how long before sampling. Sampling was conducted only in females? Please specify how many females and males were used in the study. Moreover, do authors repeated the same sampling from the animals as a control?

Answer: sampling was conducted in autumn (line 77). FEC sampling description, including quantity of males and females was added, as suggested (lines 83-89 and line 78). We did not repeat the same sampling as control.

- What was the productive system? Extensive, intensive?

Answer: Sheep production systems in Uruguay are extensive and pasture-based. This was added, as suggested (line 78).

- Could you please add more information about the breed?

Answer: Additional breed information was added, as suggested (lines 79-82).

Statistical analysis

As the three bead chips were conducted in different number of animals, did the statistical analysis referred to the same animals and how many animals were finally tested? please specify.

Answer: Available genomic information of the three chips was used in this study. The number of animals were 454, 702 and 375 for 170 SNP, 507 SNP and 50K SNP, respectively. There was a small proportion of animal genotyped with more than one chip. This information is now  illustrated in Figure 1.

line 80-81. Performance data what did included?

Answer: only FEC data was included in this study.

Please also clarify the assumptions about gene ontology that were used.

Answer: more information about gene ontology was added in lines 219-223.

Results

It is not shown clearly if genome wide associations using different approaches are conducted with the same or different number of animals. Please clarify.

Answer:  genome wide associations using three different SNP densities were conducted with different number of animals (added in each Manhattan plot legend). There is some overlapping among individuals (explained in Figure 1).

Discussion

298-300 please clarify why

Answer: We modified the sentence with the aim of clarifying this point (lines 346-353).

line 300-301 statement it is not so clear how is related to the current study.

Answer:  The 170 SNP set was preselected by Periasamy et al. (2014) based on international information using a candidate gene approach and was focused on genetic resistance to gastrointestinal nematodes. The potential significance of these SNPs was investigated in our study which validates seven SNPs in the Uruguayan Corriedale breed.

-Authors should also connected their findings with the number of analyzed animals used in each bead chip (if they were not included the same animals)

Answer: we found no connection between the significant SNPs and the number of analyzed animals.

-The manuscript would gain floor if a comparison approach on explaining why the differences between different bead chip approach were noted.

Answer: a comparison between the three different approaches was added as suggested (lines 451-460).

Round 2

Reviewer 1 Report

After the revision of the manuscript, it has been greatly improved.

At present, different experimental samples, different chips, different results. Overall, it looks chaotic.

I noticed that there are still some overlaps among the three experimental samples. What about the results of these overlaps samples using the same chip? Has the author tried?

For example, there are 307 overlapping samples in the 507 SNP and 170 SNP experimental samples. What about the results of these two chips with 307 samples? Whether there are common significant loci?

 

 

 

Author Response

Response to Reviewer 1 Comments:

After the revision of the manuscript, it has been greatly improved.

At present, different experimental samples, different chips, different results. Overall, it looks chaotic.

I noticed that there are still some overlaps among the three experimental samples. What about the results of these overlaps samples using the same chip? Has the author tried?

Answer: No, we haven´t tried since there are only 44 animals genotyped with the three chips.

For example, there are 307 overlapping samples in the 507 SNP and 170 SNP experimental samples. What about the results of these two chips with 307 samples? Whether there are common significant loci?

Answer: When we did the analysis just with the 307 overlapping samples, there were no common significant loci found.

Reviewer 2 Report

The authors applied most of comments. Correlations between elements of A22 and G matrices seem to be suitable. I have another concern about the accuracy of GEBVs which needs to be applied. As GWAS analysis is based on GEBVs coming from ssGBLUP method in this study, it is important to check and report the accuracy of GEBVs. Authors can add accuracy of GEBVs as a supplementary table. If the accuracies of GEBVs are not accurate enough, I suggest using a simple GWAS method. 

Author Response

Response to Reviewer 2 Comments:

The authors applied most of comments. Correlations between elements of A22 and G matrices seem to be suitable. I have another concern about the accuracy of GEBVs, which needs to be applied. As GWAS analysis is based on GEBVs coming from ssGBLUP method in this study, it is important to check and report the accuracy of GEBVs. Authors can add accuracy of GEBVs as a supplementary table. If the accuracies of GEBVs are not accurate enough, I suggest using a simple GWAS method.

Answer: a table containing GEBV mean accuracies, standard deviations, minimum and maximum for the three analyses (170, 507 and 50K SNPs) was added as supplementary material, as suggested (Table S2). Tables with individual accuracies can be supplied if requested.

GEBV mean accuracies for the three analyses were as follows: 0.58 for 170 SNPs, 0.57 for 507 SNPs and 0.57 for 50K SNPs.

Reviewer 3 Report

Authors incorporated in their revised version the suggested comments and improved the quality of the manuscript. Therefore, I suggest the publication of the certain manuscript in its current form.

Author Response

Response to Reviewer 3 Comments:

Authors incorporated in their revised version the suggested comments and improved the quality of the manuscript. Therefore, I suggest the publication of the certain manuscript in its current form.

Answer: thank you for your comments.