Supplemental Material for Keep et al., 2020

Figure
S1: Histograms of Minor Allele Frequencies of all SNP markers per population and
averaged over populations.

Figure
S2: Scatter plot of Pearson correlation between allele frequencies of pairs of
SNP markers from a same scaffold and genomic distance between SNP markers in
base pairs.

Figure
S3: Q-Q plots of p-values from GWAS for all available traits accounting for
neither structure nor kinship, only for structure, only for kinship or for both
kinship and structure.

Figure S4:
Pairwise phenotypic Euclidian distances plotted against pairwise Euclidian
genetic distances for all available traits. Genetic Euclidian distances were
calculated based on the 100 most significantly associated SNP markers to the
trait as given by the GWAS.

Table
S1: Genebank accessions from the natural diversity of perennial ryegrass used
in the study.

Table
S2: List of perennial ryegrass cultivars used in the study and their origin.

Table
S3: Hi-Plex genotyping description.

Table
S4: Alternative allele frequencies of 189,781 SNP markers for the natural
populations and cultivars used in this study. SNP markers are referred to as follows: scaffold number (from reference genome of Byrne et al. 2015)_scaffold reference_SNP position in scaffold. The SNP from the Hi-Plex are
indicated by “monsterplex” at the end
of the SNP marker names.

Table
S5: Seasonal climatic conditions at the three experimental gardens over the
duration of the experiments.

Table
S6: Adjusted means per population for all the phenotypic traits used in this
study.

Table
S7: Phenotypic variables description and associated results from GWAS and
genomic prediction analysis.

Table
S8: Pearson correlation coefficients between the phenotypic variables.

Table
S9: SNP markers significantly associated with each trait (q-value<10%). For
each trait, the following information is indicated: P-values, false discovery
rates, SNP marker effects and variance explained by SNP markers from both a
linear model which did not account for kinship (trait-SNP linear regression
model) and a linear model which accounted for kinship (GWAS model).

Table
S10: Number of SNP jointly associated (GWAS with q-value<10%) to phenotypic
trait pairs.

File S1: Detailed description of the phenotypic variables.