Hepatitis Associated with Chlormezanone

Ohsawa, Takaaki; Konishi, Kyoji

doi:10.1177/106002808602000619

Cited by 3 publications

(1 citation statement)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The entire results and evidences are listed in Supporting Information Table S8. Except three drugs—mibefradil, phenolphthalein, and sulfametopyrazine—having ADRs in the categories we did not predict, 25 out of 27 withdrawn drugs (92.6%) having severe ADRs are successfully predicted by our approach, and also have been found by previous reports . The overall recall of the prediction is 78.8% (calculated by using the results in Supporting Information Table S8).…”

Section: Resultssupporting

confidence: 65%

Predicting adverse drug reaction profiles by integrating protein interaction networks with drug structures

Huang

Chen

2013

Proteomics

View full text Add to dashboard Cite

The prediction of adverse drug reactions (ADRs) has become increasingly important, due to the rising concern on serious ADRs that can cause drugs to fail to reach or stay in the market. We proposed a framework for predicting ADR profiles by integrating protein-protein interaction (PPI) networks with drug structures. We compared ADR prediction performances over 18 ADR categories through four feature groups-only drug targets, drug targets with PPI networks, drug structures, and drug targets with PPI networks plus drug structures. The results showed that the integration of PPI networks and drug structures can significantly improve the ADR prediction performance. The median AUC values for the four groups were 0.59, 0.61, 0.65, and 0.70. We used the protein features in the best two models, "Cardiac disorders" (median-AUC: 0.82) and "Psychiatric disorders" (median-AUC: 0.76), to build ADR-specific PPI networks with literature supports. For validation, we examined 30 drugs withdrawn from the U.S. market to see if our approach can predict their ADR profiles and explain why they were withdrawn. Except for three drugs having ADRs in the categories we did not predict, 25 out of 27 withdrawn drugs (92.6%) having severe ADRs were successfully predicted by our approach.

show abstract

Section: Resultssupporting

confidence: 65%