2.4. Data analysis

Normalization of the raw data files was carried out with Robust Multi-Array Average (RMA), and annotation of the probesets was performed using HarvEST WheatChip (http://harvest.ucr.edu/) [4]. Identification of probesets that are differentially oxidized between dormant and after-ripened seeds was performed by analysis of variance (ANOVA) using FlexArray software [5]. Probesets were considered to be differentially oxidized if they showed ≥2-fold changes at probability level of ≤0.05. Gene ontological analysis of the probesets oxidized in response to after-ripening, which was performed using the AgriGO analysis toolkit [6], indicated their distribution over different functional categories [1]. Analysis of the Arabidopsis genes corresponding to the wheat probesets representing mRNAs differentially oxidized between dormant and after-ripened seeds using SeedNet (http://vseed.nottingham.ac.uk; [7]) revealed that genes corresponding to the oxidized probesets are overrepresented in a region consisting of gene sets associated with seed dormancy (Fig. 1).

Distribution of genes highly oxidized in dry dormant (a) and after-ripened (b) wheat seeds in SeedNet topology consisting of gene sets associated with dormancy (red) and germination (purple) (http://vseed.nottingham.ac.uk; [7]).