Fig. 1

Graphical representations of machine learning model against the dataset. a Force-plot of most important features as calculated by Recursive Feature Elimination by Cross-Validation with XGBoost. Features highlighted in red are responsible for driving the final prediction of a sample into the positive category (A probable xylose transporter) while features in blue drive the prediction into the negative category (A non-xylose transporter). The base value represents the average prediction for the samples, while the size of the feature represents its impact (higher or lower importance). b Common metrics used to evaluate a model, the grey values correspond to the base threshold model and blue to the altered threshold. c Confusion matrix showing the results of predictions against the test data