Fig. 1
From: Deep learning decodes species-specific codon usage signatures in Brassica from coding sequences
Three dimensionality reduction techniques applied to genomic data: (a) t-SNE, (b) UMAP, and (c) PCA. Each panel visualizes distinct aspects of genetic diversity, transcriptional patterns, and evolutionary relationships across Brassica species (napus, juncea, oleracea, and rapa). The t-SNE and UMAP plots (a, b) employ a purple-to-yellow gradient (100–102) to highlight expression density, revealing local and global structures in gene clusters. Meanwhile, the PCA scatter plot (c) elucidates phylogenetic relationships through RSCU patterns, emphasizing key variances and evolutionary trends. Together, these methods provide complementary insights into complex genomic datasets.
