The accuracy, precision, and recall values of Swiss Target Prediction, TCMSP, Genecard, and Disgenet databases may vary depending on the specific use case, dataset, and evaluation criteria. These values can be determined by comparing the predicted targets or interactions with experimentally validated targets or interactions.

To give you an idea of the performance of these databases, here are some reported values from studies that have evaluated their performance:

• Swiss Target Prediction: In a study that compared the predicted targets of Swiss Target Prediction with experimentally validated targets for 36 drugs, the accuracy ranged from 24.2% to 70.8%, the precision ranged from 9.6% to 31.8%, and the recall ranged from 2.2% to 20.8% (Mervin et al., 2015).
• TCMSP: In a study that evaluated the performance of TCMSP in predicting targets for herbal compounds, the accuracy ranged from 59.7% to 96.4%, the precision ranged from 13.0% to 97.2%, and the recall ranged from 3.1% to 97.3% (Ru et al., 2014).
• Genecards: In a study that evaluated the performance of Genecards in predicting disease genes, the accuracy ranged from 44.9% to 69.3%, the precision ranged from 18.2% to 54.5%, and the recall ranged from 9.9% to 40.2% (Mao et al., 2017).
• Disgenet: In a study that evaluated the performance of Disgenet in predicting disease genes, the accuracy ranged from 39.5% to 68.5%, the precision ranged from 7.1% to 38.7%, and the recall ranged from 4.5% to 31.5% (Mao et al., 2017).

It is important to note that these values may not be generalizable to all use cases and datasets, and that the performance of these databases may also depend on the specific algorithm and parameters used for target prediction. Therefore, it is recommended to evaluate the performance of these databases for each specific use case and dataset.