Fig. 4.

Confusion matrices and RNA-seq count heatmaps corresponding to the best gene set of LV 5-Way dataset.

(A) Confusion matrix for classification of LV 5-Way dataset using best gene set identified by filter feature selection. (B) RNA-seq count heatmap of best LV 5-Way gene set averaged per condition. (C) Confusion matrix for classification of AH, AC, and CT samples within independent validation dataset. (D) RNA-seq count heatmap of best gene set within independent validation dataset (AH, AC, and CT) averaged per condition. AC, alcohol-associated cirrhosis; AH, alcohol-associated hepatitis; CT, healthy controls; HP, chronic HCV infection; LV, liver tissue; NF, non-alcohol-associated fatty liver disease; RNA-seq, RNA sequencing.