Exploring the relationship between fractal features and bacterial essential genes
Yu Yong-Ming1, Yang Li-Cai1, †, , Zhou Qian2, 3, Zhao Lu-Lu1, Liu Zhi-Ping1
       

Mean values and SDs of AUC scores of genes essentiality prediction in 27 bacteria based on Random Forest classifier trained by different integrations of fractal features. The average AUC scores of 27 bacteria are above 0.64. There are 20 (75.07%) bacteria obtaining better predictive performance with fractal features of gene sequences than that of protein sequences and the better predictive performance is obtained by using all fractal features.