Enrichment analysis of the genes correlated with geoclimatic features. (a) Projection of all the geoclimatic features into the first two canonical factors. A subset of the dots represents labels as examples. (b) Projection of the 5,000 genes in the transcriptome data together with two geoclimatic features: “Net Radiation in Summer” and “Precipitation in the Driest Month”, marked by the arrows. The 100 genes closest to the two geoclimatic features are shown in red and yellow, respectively. (c) The improved enrichment analysis of the 100 closest genes as more levels of HCCA is considered. The baseline results are obtained by directly calculating the correlation between gene expressions and the two geoclimatic features without any projection or coprojection.