Research Article

Joint Decision-Making Model Based on Consensus Modeling Technology for the Prediction of Drug-Induced Liver Injury

Table 5

Top 10 descriptors with a significant difference between DILI positives and DILI negatives.

No.NameDescriptionVPMVNMDI

1MATS3cMoran autocorrelation—lag 3/weighted by charges−3.72e-36.88e-352.32
2ALogPGhose–Crippen LogKow3.74e-3−0.4352.77
3ATSC1pCentered Broto–Moreau autocorrelation—lag 1/weighted by polarizabilities−0.0688.86e-31.94
4MATS2cMoran autocorrelation—lag 2/weighted by charges0.0216.62e-30.92
5MATS2iMoran autocorrelation—lag 2/weighted by first ionization potential0.0210.0370.60
6MATS7sMoran autocorrelation—lag 7/weighted by I-state−0.035−0.060.56
7AATSC6vAverage centered Broto–Moreau autocorrelation—lag 6/weighted by van der Waals volumes−0.954−1.4740.45
8AATSC2vAverage centered Broto–Moreau autocorrelation—lag 2/weighted by van der Waals volumes1.2621.9000.43
9ATSC4iAverage centered Broto–Moreau autocorrelation—lag 4/weighted by first ionization potential−9.774−13.8310.36
10AATSC1pAverage centered Broto–Moreau autocorrelation—lag 1/weighted by polarizabilities−2.00e-3−1.51e-30.32