Fig. 2
From: Decoding cell-type contributions to the cfRNA transcriptomic landscape of liver cancer
Analysis of hepatocyte proportions in cfRNA-seq datasets. A differences in hepatocyte proportions were analyzed with a one-sided Wilcoxon sum-rank test. P values were adjusted with the Benjamini–Hochberg procedure where appropriate. ns not significant, ***P < 10–4, ****P < 10–5. HD, healthy donor; LC, liver cancer; CRC, colorectal cancer; ESCA, esophageal cancer; LUAD, lung cancer; STAD, stomach cancer. B Hepatocyte proportion cutoff-based classification of HD and LC samples for proportions in the range of 0.2—0.4. Samples with a hepatocyte proportion higher than the tested proportion were classified as LC and HD if otherwise. The dotted vertical line indicates the most accurate hepatocyte proportion cutoff (0.27). C Confusion matrix of the results achieved after using the optimal hepatocyte proportion cutoff of 0.27. The left-hand top (71.7%) and right-hand bottom (74.1%) numbers represent the positive and negative predictive values, respectively. The numbers on the top (74.3%) and bottom (71.4%) represent the success rate of classifications