Fig. 5.

Sample dendrogram and trait heatmap generated with WGCNA. The clustering on the top was based on the -log(TPM + 1) values per genes for the 360 samples. The heatmap, with colour intensity, was proportional to Brix, purity, polarity, CCS, fibre levels and respective difference for early, mid- and late season and tons of sugarcane per hectare (TCH). Traits data were normalized using Auto Scaling method. The red and blue colours represented, respectively, positive and negative. Clustering dendrogram of samples was based on Euclidean distance