Table 1 Results from binary classification analysis of 11,574 samples from the TCGA dataset

From: Exploring the role of stromal osmoregulation in cancer and disease using executable modelling

Data subset Accuracy Accuracy (%) False positives False negatives MCC score
All samples 2290 of 2315 98.92 2 of 2159 23 of 156 0.91
EC origin cells 558 of 564 98.94 0 of 535 6 of 29 0.89
EN origin cells 849 of 857 99.07 1 of 790 7 of 67 0.93
M origin cells 1015 of 1030 98.54 3 of 944 12 of 86 0.9
Random non-channel subset 2221 of 2315 95.94 13 of 2006 81 of 309 0.81
  1. Data are classified using only expression data of ion channels, and shows a high accuracy when compared to a set of random non-channel genes from the same dataset (Matthews Correlation Coefficient of 0.91 vs. 0.81). Included are the false positive numbers and false negative numbers