Polygenic architecture of flowering time and its relationship with local environments in the grass Brachypodium distachyon

Abstract Synchronizing the timing of reproduction with the environment is crucial in the wild. Among the multiple mechanisms, annual plants evolved to sense their environment, the requirement of cold-mediated vernalization is a major process that prevents individuals from flowering during winter. In many annual plants including crops, both a long and short vernalization requirement can be observed within species, resulting in so-called early-(spring) and late-(winter) flowering genotypes. Here, using the grass model Brachypodium distachyon, we explored the link between flowering-time-related traits (vernalization requirement and flowering time), environmental variation, and diversity at flowering-time genes by combining measurements under greenhouse and outdoor conditions. These experiments confirmed that B. distachyon natural accessions display large differences regarding vernalization requirements and ultimately flowering time. We underline significant, albeit quantitative effects of current environmental conditions on flowering-time-related traits. While disentangling the confounding effects of population structure on flowering-time-related traits remains challenging, population genomics analyses indicate that well-characterized flowering-time genes may contribute significantly to flowering-time variation and display signs of polygenic selection. Flowering-time genes, however, do not colocalize with genome-wide association peaks obtained with outdoor measurements, suggesting that additional genetic factors contribute to flowering-time variation in the wild. Altogether, our study fosters our understanding of the polygenic architecture of flowering time in a natural grass system and opens new avenues of research to investigate the gene-by-environment interaction at play for this trait.

Dear Dr. Roulin: Two experts in the field have reviewed your manuscript, and I have read it as well, finding much interest in it.While your manuscript is not currently acceptable for publication in GENETICS, we would welcome a substantially revised manuscript.Both reviewers have comments and concerns to be addressed in a revised manuscript.You can read their reviews at the end of this email.
We look forward to receiving your revised manuscript.Please let the editorial office know approximately how long you expect to need for revisions.
Upon resubmission, please include: 1.A clean version of your manuscript; 2. A marked version of your manuscript in which you highlight significant revisions carried out in response to the major points raised by the editor/reviewers (track changes is acceptable if preferred); 3. A detailed response to the editor's/reviewers' feedback and to the concerns listed above.Please reference line numbers in this response to aid the editor and reviewers.4. Please identify a public repository from which all salient data are available.
Your paper will likely be sent back out for review.
Additionally, please ensure that your resubmission is formatted for GENETICS https://academic.oup.com/genetics/pages/general-instructionsFollow this link to submit the revised manuscript: Link Not Available Sincerely,
A few questions and concerns come to mind as I read the manuscript: 1. Line 193: what is the population structure of accessions used in this study?Is this population structure correlated with their geographical origins? 2. Line 244-250: How many accessions with outdoor flowering time measurements from Stritt et al. 2022 are used for this study?Given that there were 105 accessions measured in 2017, it would be helpful to clarify whether the accessions used in this study are considered as good representatives for all available lines.Additionally, the authors should present and interpret the results of flowering time in the accessions grown in greenhouse and outdoor in a more comparative way.
3. Line 251-258: After the filtering process, how many SNPs in total were retained for the association study?What is the general marker density across the genome?My concern with this GWAS section is its statistical power to identify significant peaks.One year of phenotypic data for association mapping, especially for a polygenic trait, might be insufficient.Both the acquisition and utilization of the phenotypic (flowering time under outdoor conditions) and genetic data (SNPs) should be better described in the current study.This clarification is necessary because the lack of significant associations overlapping with known flowering-time genes (Figure 5D) cannot be solely attributed to environmental or gene-environment interactions.The authors should provide more evidence to support their claim that these flowering-time genes may not largely contribute to the flowering time variation in Brachypodium distachyon in the wild (line29).Multiple genes with relatively small to moderate effect sizes (undetected in this study) could collectively have more substantial additive effects on a polygenic trait.
Other minor comments: Line 76: please provide more context regarding the term "near base-perfect reference genome".Fig. 1B, 1C: please add the unit of precipitation.Line 339: Table 1 only presents environmental variables, not genetic clades.Line 353: spell out the full name of AFT-genes.Line 604-629: consider a more concise presentation for this section, as the discussion of flowering-gene functions and regulatory mechanisms may not be directly relevant to the main topic of this study.Line 733-734: delete one of the duplicate "therefore".Line 743-744: please restructure this sentence.
Reviewer #2 (Comments for the Authors (Required)): The manuscript by Minadakis et al. explores the evolutionary genetics of natural variation in flowering time and its response to vernalization across natural accessions of the genetic model plant Brachypodium distachyion.The work is a helpful addition to a small but growing and important literature implicating selection as an evolutionary force capable of maintaining long-range linkage disequilibrium.The authors thoughtfully analyze their data in light of the population structure inherent in their sampling, and they provide complementary analyses with appropriate genome-wide controls that evaluate alternate hypotheses for their observations.In addition, the manuscript also builds on a series of results over the last two decades or so finding that the loci detected as underlying genetic variation in a trait when observed in controlled or greenhouse conditions are not the same as those detected in field conditions relevant to natural populations.In general, I thought the work was well designed, described, and notable, and I have just a handful of major comments that I would ask the authors to consider as they revise their manuscript.
Quantifying flowering time parameters: Two measures of flowering time are extracted from the observed vernalization reaction norms.Minimum DTF with and minimum DTF without including the vernalization period.At least three other notable parameters could be extracted from the dataset using actual or modeled (Donohue et al. 2015 TREE) values: the minimum threshold duration of vernalization necessary to be permissive for flowering, the DTF when that minimum vernalization period is received, and the saturating threshold duration of vernalization above which no further reduction in flowering time is gained.Depending on how correlated these parameters with the flowering parameters already assessed, they may track different environmental variables, informing additional hypotheses about how local conditions act as agents of selection on phenology, and may help parse what GWAS signals relate to what aspects of the regulatory control of the floral transition.I encourage the authors to consider how best to infer additional parameters like these for parallel use in all their downstream analyses, and a sentence or two describing the developmental and/or ecological rationale for choosing the measures of DTF that they have chosen would improve interpretation by the reader as well.
Reporting of climate parameters: How spatially correlated are the environmental factors with each other?Could solar rad and precipitation be yielding mostly the same information?Please add a supplemental figure showing the pairwise correlations among variables and the climate PCA, and a supplemental table with the climate PC1 and PC2 loadings.The climate parameter data should also be addressed in the data availability statement.
Genotype-environment association analysis: A recurring theme is that population structure and flowering time variation covary, potentially leading to weaker magnitude or false negative results when population structure is controlled for.For the geneenvironment association analysis, the extent of this problem could be quantified using RDA, which provides estimates of how much of the total genetic variation is explained by structure, environment, or both.
flowering time and its relationship with local environments in the grass Brachypodium distachyon.