Annotation and gene ontology (GO) analysis

To obtain a functional annotation for the transcripts, we used blastp (Camacho et al., 2009) to search for the best match among Arabidopsis representative set of proteins (Berardini et al., 2015) downloaded from TAIR FTP server (ftp://ftp.arabidopsis.org/home/tair/Proteins/TAIR10_protein_lists/). The retrieved best hit was assumed to be an Arabidopsis ortholog and the functional annotation and GO category assignment of the ortholog was transferred to the P. lanceolata query transcript. The Arabidopsis GO annotations were downloaded from the TAIR FTP server (ftp://ftp.arabidopsis.org/home/tair/Ontologies/Gene_Ontology/). Gene ontology enrichment analysis was carried out using piano software (Varemo, Nielsen, & Nookaew, 2013) and the log2 fold changes and false discovery rate (FDR) adjusted P-values of differentially expressed genes were imported from DESeq2 results.
The GO hierarchy enrichment tests were carried out using the GOATools toolkit (Klopfenstein et al., 2018). To focus on signaling responses, we selected the responses and signaling sub-branches of the Biological process category. The GO enrichment results were plotted using R with fold change values obtained from the piano package. The GO enrichment analyses of RDA loadings was carried out with piano software (Varemo et al., 2013) using RDA loadings as gene level statistics and GO categories as gene set collection and mean statistics. Again, the results were plotted in R.