SOM analysis for (A) ABA and (B) MeJA treatments, (C) phosphate deficiency and (D) drought. Genes were selected based on a correlation coefficient of 0.7 or higher in the expression pattern obtained for the two biological replicates. The values of the median intensity ratios for each biological replicate were mean-centered and the average values were used as input for the SOM clustering. The geometry was chosen based on a PCA Analysis. The graphs present the average of the normalized log2 ratio (M) value between the replicates (y axis) plotted against the time course (x axis). The components of the SOM groups obtained are available in their totality as additional files (additional file 4: Table 4, additional file 5: Table 5, additional file 6: Table 6 and additional file 7: Table 7). The number in brackets indicates the number of SAS in each group.