Skip to main content
. 2012 Nov 15;2012:bas042. doi: 10.1093/database/bas042

Table 1.

Dataset

Dataset Chemical names Positives Negatives Total
Training 2-Acetylaminofluorene 81 97 178
Amsacrine 37 32 69
Aniline 100 126 226
Aspartame 46 110 156
Doxorubicin 138 61 199
Indomethacin 76 9 85
Quercetin 392 150 542
Raloxifene 161 109 270
Test Cyclophosphamide 107 47 154
Phenacetin 65 21 86
Urethane 106 98 204

The training and test sets include eight and three target chemicals, respectively. Because the ratio of positive and negative examples varies with target chemicals, our system is tuned to achieve high MAP score on the training chemicals.