Data analysis
For each trait and condition, the mean values were obtained from at
least three replicates. Any accessions that did not have data for at
least three replicates for each condition was removed from the analysis.
After filtering the number of accessions used for analysis was 142 for
ERD, 109 for FRD, 127 for FT and 102 for yield. For each accession, the
trait values were estimated as the average of the replicates in each
condition (Tables S2 and S6). Two plasticity measurements were used
namely the fold change (FC) between two conditions and the coefficient
of variation (CV) across conditions. The CV was calculated as the
standard deviation (σ ) divided by the mean (μ ) (CV =σ / μ ) of the averages of each condition. All statistical
analyses were performed in R (https://www.R-project.org/; (Team, 2020)),
using RStudio v. 1.3.1056. The correlations were calculated using thecorr.test function of psych package v. 2.0.9
(https://CRAN.R-project.org/package=psych; (Revelle, 2020)), and
plotted using the corrplot package v. 0.84
(https://github.com/taiyun/corrplot;(Wei, 2017)). The correlations were
calculated using Spearman’s rank coefficient and pairwise comparisons,
and the p-values were corrected by the Benjamini–Hochberg procedure.
The boxplots were generated using the ggplot function ofggplot2 package v. 3.3.2 (http://www.jstatsoft.org/v40/i01/;
(Wickham, 2009)), ddply function of plyr package v. 1.8.6
(Wickham, 2011), and the pivot_longer function of tidyrpackage v. 1.1.2 (https://CRAN.R-project.org/package=tidyr; (Wickham,
2020)). K-means clustering was performed with the kmeans function
of stats package integrated in R (v. 4.0.3), using Hartigan and
Wong algorithm. The optimal number of clusters was determined with the
Silhouette method using the fviz_nbclust function offactoextra package v. 1.0.7
(https://CRAN.R-project.org/package=factoextra;(Kassambara, 2020) ),
using 1000 bootstrap samples. Heatmaps were plotted using theComplexHeatmap package v. 2.6.2 (Gu, Eils, & Schlesner, 2016).
The graphs were further edited using Microsoft Office.