Identification of salt-responsive genes and gene ontology enrichment analysis

After filtering transcripts with few reads, salt-responsive genes were identified using the raw count data of fragments as input in the GFOLD software [40]. The GFOLD value is a reliable estimator of the relative gene expression developed for the analysis of pooled experiments [40]. Density plots with log₁₀ normalized expression values were used to compare their distributions. Overlapping expression distributions indicate appropriate homogeneity of the sequencing depth and that count normalization is suitable to compare the expression levels of the different libraries [96]. The sample at 0 min was used as a control for both 8 and 15 min ASE, assuming that few physiological changes occur in this short time under normal conditions. A high absolute GFOLD value indicated greater up- or down-regulation of the genes. Genes with GFOLD values > 1 or < − 1 when c = 0.01 were considered for further analyses as they represent relevant changes in expression levels under stress conditions. The defined c value is a parameter that indicated that in 99% of the cases, the fold change of a gene is above the absolute GFOLD (0.01) value calculated for this gene [40].

The GO enrichment tool from the STEM software was implemented to distinguish the categories of genes over-represented by time point in the contrasting genotypes [97, 98]. Only the gene categories from the transcriptomic background of each genotype were retained in the analysis [99]. A Bonferroni multiple hypothesis correction test was employed, thus GO terms with a corrected p-value < 0.001 were considered as over-represented. The STEM algorithm was used to cluster expression profiles during the osmotic phase, and the over-represented GO categories in the clusters were defined [97, 98]. Key categories over-represented during the osmotic phase were selected to create graphs of the corresponding genes’ time-course expression. A LOESS model was fitted to represent the expression tendency of the clusters of genes. The expression levels of transcripts from the transmembrane transport category related to ion homeostasis were compared in the contrasting genotypes at the ionic phase.