Fig. 7

Prediction of gene expression levels based on intron features. Random Forest Performance and Feature Importance. (a) 10-fold Cross-validated ROC curves for Random Forests trained with the median-split (whole) set and quartile expression set, respectively, (b) MDA feature importance for Random Forest model trained with the lower and upper quartile expression dataset, For feature explanations, see Table 1. (c) SHAP summary plot of Random Forest model trained with the lower and upper quartile expression dataset, (d) SHAP value to feature value plot for distance to TSS, with the respective distance to CDS-start values color-coded. Positive SHAP values indicate an association with the high expression class, negative SHAP, association with the low expression class of genes