Skip to main content
. 2021 Feb 3;12:765. doi: 10.1038/s41467-021-20988-w

Fig. 5. Identification of predictive genotypes for pathogenicity in avian E. coli using random forest (RF) models.

Fig. 5

a The importance of the predictors derived from the four GWAS using the primary classifier model (trained using data from the four lineages A, B1, B2, ST-117); b Receiver operating characteristic (ROC) curve showing the overall performance of the primary classifier model; c Importance of the top 10 predictors in the primary classifier model; d ROC curve showing the overall performance of the follow-up classifier model (trained using data from the four phylogroups A, B1, B2 and predicting in ST-117); e The importance of the top 10 predictors in the follow-up classifier model. Data are presented as mean values ± SD from n = 10 repeated analyses. Source data are provided as a Source Data file.