Fig. 3

K-means clustering of gene expression data. A k-means clustering was performed on DEG from the time-course and two-bendings experiments. For each cluster, at each time -point, a boxplot represent the dispersion of the log2-converted mean fold-change (bent vs. unbent poplars) for all the genes of the cluster. For graphical clarity, the different clusters were manually re-grouped according to their mean expression profile