Marker-Trait Associations for Tolerance to Ash Dieback in Common Ash (Fraxinus excelsior L.)

Round 1

Reviewer 1 Report

The manuscript describes an original research. The molecular methods and tools used help researchers better study and identify the genetic differences responsible for decay

The manuscript describes an original research. The molecular methods and tools the researchers used to better study and identify the genetic differences responsible for the tree's decay. In particular, the experimental research was properly designed in order to obtain statistically adequate and significant correlations. In fact - and in detail, (for example) the linear model used in the research to test the association between genotypic variation and disease severity showed that the two SNPs have significant associations: one is significantly associated with disease severity (in a gene that encodes for an S8 / S53 peptidase related to the subtilisin domain), while the other is only marginally so. The results obtained in this research, together with those of other researches, are extremely useful: for understanding the degree of disease-coding genes association, for genetic improvement programs, analysis of forest diseases and the environment.....

Author Response

We would like to thank the reviewer for his comments.

Reviewer 2 Report

The ability to find SNP related to tolerance to any disease is like trying to find ‘a needle in a haystack’. Therefore any attempts to find such associations are much needed. The whole-genome sequencing of large numbers of individuals seems to be prohibitively expencive yet.

The approach used in this study was not very successful. Starting with 1000 amplicons, after filtering od data, only 40 contigs remained for the final association study. 40 SNPs and 249 individuals is not an impressive dataset for association analyses. I would like to see the explanations: why starting with more than 150 k SNPs on 655 contigs, the dataset was reduced (filtered) down to 40 contigs only. Was missing data a problem?

Finally, only one SNP appeared to be significantly associated with the tolerance to HF, with a low level of explained variance. Given the way of generating genomic data (amplicon selection, filtering etc.) the ability to find the association with the used approach seems rather the pure (by chance) coincidence, and the overal approach is not the one that I would recommend for future research.

Summing up, despite that I see no flaws in the paper, it presents the approach which turned out to be not very efficient in finding genetic association to trait such as pathogen tolerance. Therefore, the paper will have a minor effect on the progress in the studied area.

Minor points:

How many SNPs in total were identified within the contig 5992? If there were additional SNPs within this contig, what was their location, and what was their effect on the predicted protein?

Line 88-90; the presentation of the main finding of the paper in the last line of the introduction is unusual. This information is in the abstract already. The introduction should end with hypotheses and expectations of the results, not the results by themselves.

193, why minimum coverage is indicated in %, % of what?

205, 10,0000 is unysual notation, should it be 10,000 or 100,000  ?

219, word genotypes is repeated, consider re-wording.

249, does this mean that only SNPs that were present in 70% of samples were retained, please clarify the sentence.

251, what criteria were used to select the one SNP per contig, was it random or not?

254, from 503 kb to 1.35 kb – is this correct statement? According to the published Fraxinus genome, the largest contig in their assembly is 208 kb.

Author Response

Author Response File: Author Response.docx