*2.3. Functional Annotation and Localization*

The initial functional annotation of the identified 152 genes is based on the Map-Man Bin categories [29] and was curated manually for gene descriptions and functions of encoded proteins afterwards. To this end each gene was checked for database entries in The Arabidopsis Information Resource (TAIR) (https://www.arabidopsis.org) and The Universal Protein Resourcce (UniProt) (https://www.uniprot.org). Information on potential interaction partners was obtained from the Biological General Repository for Interaction Datasets (BioGRID) (https://thebiogrid.org) and the real-time multiple association network integration algorithm for predicting gene function (GeneMANIA) (https://genemania.org) databases. In addition, information on potential localization of gene products was extracted from the Bio-Analytic Resource (BAR) for Plant Biology and the integrated Arabidopsis Cell eFP Browser (https://bar.utoronto.ca) and aligned with the corresponding information in the Arabidopsis thaliana chloroplast protein database AT\_CHLORO (http://at-chloro.prabi.fr) for sub-plastidial localization [30]. Genes were then classified according to major functional categories (which may differ in some cases from the original MapMan bins) and were given as heat map sorted on the base of the expression values in the LIN data set. All collected information about the gene products are summarized in Supplemental Table S1.
