Figure 3.

Feature importance analysis of random forests sub-Golgi classifier, rfGPT _4: (A) importance of feature types (B) the ranking orders of 93 features for rfGPT_4 and their integrated importance (red line), and (C) the importance of the top 25 features, which accounted for 50% of the integrated importance (blue line). The A₁A₂.gap2 means the composition of dipeptide A1A2. A1 or A2 is one of the 20 amino acid residues. Nterminal_D means the composition of amino acid residues D (aspartate) in NH₂-terminal of protein sequence. InterTier_K, interTier_W, and interTier_F mean K(lysine), W(tryptophan), and F(phenylalanine) amino acid residues composition of the inter-tier between NH₂-terminal and COOH-terminal of protein sequence.