This manuscript by Barres et al investigates one of the key questions in the disease ecology, how diversity is maintained in host populations. They test Janzen-Connell hypothesis using 15 oak families. They investigated the effect of powdery mildew disease on oak population structure in a multiyear experiment. The authors followed powdery mildew epidemics, host mortality and population diversity (half sib family diversity) across years in controlled field study. They also SNP genotyped the experimental plants to assess changes in heterozygosity and to perform GWAS on infection traits, height, survival, and acorn wight. They present enormous dataset and all experiments seem adequately designed and analysed. 

They find that against expectations, there are no strong effect of powdery mildew infection on population structure of oak. The same fittest families survive in diseased and protected treatments. 

In GWAS, the authors are able to identify multiple loci that contribute to acorn weight and some that contribute either to survival, infection or height. 

This manuscript has both applied importance as it describes potential loci of interest for breeding and provides understanding on the lack of growth defense trade-off in a forest tree. 

I have no major criticism, only few comments how to clarify the paper. 

l. 80 it would be good to describe the causal agents of powdery mildew in the introduction somewhere more specifically. In the first mention I recommend stating the genus or species scientific name.

l. 127 to me the maternal conditions high vs low disease pressure come here as a surprise. I would recommend using a small paragraph earlier in introduction for transgenerational effects of disease. There is also a recent publication on maternal effects in powdery mildew resistance in Plantago: Höckerstedt et al. J Ecol. 2021; 109: 1439–1451. https://doi.org/10.1111/1365-2745.13568.

l. 141-142 Please, be more specific.

l. 144 under which stress?

l. 149 A brief description of the genetics and biology of  Q. robur is needed here. E.g. is it diploid or what? What powdery mildew species might infect it? What other stresses are important in its ecology?

l. 151 Please, describe here how infected the mother trees were and for how long was that assessed? How far the mother trees were from each other? How is this species pollinated? How it is known that the seeds are half sibs?

l. 154 when and where the trees were sown? It would be good to state here the duration of the study.

l. 168 please, provide details of the annual disease scoring for each year. Was the scoring performed every year? How many times?

l.270 How the 7 % minor allele frequency threshold was chosen?

l. 671 it would be good to explain the underlying genetics of powdery mildew-oak system somewhere. 

Minor

l.23 it would be good to mention also scientific name for the pathogens causing powdery mildew in the abstract.

l. 89-90 specialized feeding structures that penetrate in living cells?

l. 94 replace little with rarely

l. what kind of importance? Economic, cultural,  ecological or what?

l. 97-100 this sentence is unclear and needs clarification.

l. 105-108, this sentence could be simplified.

l.113-115 could the authors place the citation to refer the study in question.

l. 339 something weird in this sentence

Fig. 5. This is very informative way of presenting the results! There seems to be error in the direction of  ] in the infection rate legend. 

l. 591 see also Susi et al. Journal of Ecology 105, no. 5 (2017): 1399-1412.

The manuscript submitted by Barrès and colleagues describes the effect of several years of exposure to powdery mildew on the demographics and genomic diversity of a young population of Quercus robur. The latter consisted of 1,733 individuals recovered from 15 half-sib families from a single geographic area (Cestas), and exposed to natural infection by agents responsible for powdery mildew. The study is based on the same experimental set-up as the one used in previously published research on the host phenotypic response to infection (Desprez-Lousteau et al. 2014. Annals of Forest Science 71:563–573). In the new study, infection was monitored from 2009 to 2013, and survival was evaluated until the end of the 2017 growing season. In addition, an exhaustive combination of genetic /genomic analyses was carried out on close to 1,200 individuals. The authors then used phenotypic and genomic data in statistical models to estimate the role of abiotic and biotic factors in survival and identify genomic regions that might contribute to disease resistance. The authors also concluded that there was no evidence of a trade-off between host growth and defense in their pathosystem.

Globally, I am impressed by the experimental approach, diversity of analyses, and overall quality of the work. I agree with the authors that, in the population of Q. robur they studied, « Powdery mildew infection thus did not suppress the competitive advantage of the dominant, fast-growing families, and no evidence of a growth-defense trade-off was obtained.» While it may be argued that the nine years of growth considered in the analysis represent only a small fraction of the lifespan of a long-lived tree species, the authors have generated a large quantity of useful data on a phase of tree growth and establishment that is critical, especially with regard to exposure to powdery mildew.

I have identified a few relatively minor points which I would like the authors to address.

1. In the Discussion, the authors state that the study of a single population sampled locally may not have allowed for the detection of trade-offs between traits, and suggest that sampling across a wider spatial range might have yielded different results (L613-617). Yet, in the Introduction, the authors state that high genetic diversity is present within oak populations (L97-98; L115-116). So, if the level of genetic diversity found within the Cestas population is comparable to the average level found in populations of Q. robur throughout its range, the problem associated with sampling only a local population may not be so important after all. On the other hand, if the Cestas population has an unusually low level of genetic diversity, the possibility that methodological constraints prevented the detection of trade-offs between traits becomes higher. How does the level of genetic/genomic diversity of the Cescas population compare with that of other Q. robur populations?
2. In their description of the experimental design, the authors state that there were two levels of exposure to powdery mildew: either Natural or Protected, with protection provided by the application of the fungicide Myclobutanil in the latter case. In their 2014 Annals of Forest Science paper based on the same experimental design, Desprez-Lousteau et al. indicate there were three levels of exposure (low, medium, and high) to powdery mildew. The low level corresponded to repeated applications of Myclobutanil. The medium level corresponded to natural inoculation, whereas the high level was obtained by blowing spores from highly sporulating leaves above the seedlings, using a portable air compression device. My understanding is that the new study included only plots with medium (Natural) and low (Protected) exposures. If this is indeed the case, I suggest mentioning this in the Materials and Methods section.
3. The identity of the causal agents of powdery mildew is not mentioned in the text, except for the mention of Erysiphe quercicola and E. alphitoides in the Keywords. I suggest including at least one sentence on the identity and specificity of the pathogens responsible for powdery mildew on Q. robur, as was done in the 2014 Annals of Forest Science paper.
4. The caption of Figure 3 should include information on the meaning of the small vertical green and beige/light brown bars at the top and bottom of the graph, respectively.
5. The caption of Figure 10 should include information on the acronyms for the genotypic classes C/C, T/C, and T/T. I have indicated a few typos on the annotated PDF copy of the manuscript.