“…For example, we compared the mlDNApredicted salt-related and cold-related genes with the genes associated with a strong phenotypic response in their T-DNA mutation lines documented in a recent large-scale phenotype screening work (Luhua et al, 2013) and found 16 salt-related genes (AT5G41080, AT2G32210, AT1G02660, AT1G76600, AT1G55040, AT2G47710, AT3G14060, AT3G26470,AT3G60520, AT2G15560, AT3G29370, AT5G06130, AT2G34600, AT5G57340, AT3G46960, and AT1G18900) and 13 cold-related genes (AT2G32210, AT5G51570, AT2G40000, AT1G79660, AT5G62920, AT4G31730, AT1G78070, AT2G25250, AT1G16730, AT1G18850, AT4G34630, AT1G30200, and AT1G56230) that did not pass the statistical cutoff by either the t test, Limma, or SAM. In addition, there were 83 stress-related genes experimentally supported by the large-scale phenotypic screening work (Luhua et al, 2013), of which 19 were known stress-related genes in the positive sample set, and 64 were newly predicted by mlDNA (Supplemental Data Set 4). For all six types of stress, we found that the majority of stress-related genes were tissue specific, with 75 to 85% of the genes usually specific to either roots or shoots ( Figure 5C).…”