Skip to main content

Table 3 Results of prediction on non-redundant data sets obtained in standard cross validation procedure. Results for both for models using gene expression only (columns 2, 3 and 6, 7) and models built on integrated data sets (columns 4, 5 and 8, 9) are shown. Results for model built on molecular descriptors shown in the last row

From: Integration of human cell lines gene expression and chemical properties of drugs for Drug Induced Liver Injury prediction

   AUC    MCC
Cell line    GE all GE base    GE all GE base
  GE all GE base + chem + chem GE all GE base + chem + chem
A375 0.48 0.48 0.59 0.62 -0.01 0.00 0.05 0.12
A549 0.57 0.47 0.66 0.64 0.08 -0.07 0.13 0.09
ASC 0.47 0.50 0.59 0.63 -0.01 0.02 0.05 0.13
HA1E 0.58 0.59 0.65 0.66 0.10 0.12 0.17 0.19
HCC515 0.56 0.45 0.66 0.63 0.09 -0.03 0.16 0.07
HPEG2 0.51 0.53 0.62 0.63 0.02 0.05 0.09 0.09
HT29 0.48 0.49 0.59 0.62 -0.01 0.02 0.07 0.10
MCF7 0.62 0.62 0.70 0.70 0.23 0.18 0.29 0.23
NPC 0.42 0.43 0.57 0.64 -0.07 -0.07 0.02 0.09
PC3 0.43 0.44 0.59 0.60 -0.06 -0.03 0.06 0.07
PHH 0.42 0.44 0.56 0.62 -0.10 -0.08 0.03 0.07
SKB 0.50 0.51 0.61 0.67 0.03 0.14 0.08 0.17
VCAP 0.58 0.51 0.66 0.63 0.06 0.03 0.12 0.11
molecular descriptors 0.66 0.15