One needs three things to successfully use predictions of off-target activities in the drug-discovery process:
1. Consistent assays of the relevant activities, preferably with high throughput. 2. A software infrastructure to generate, validate, and publish QSAR models. 3. Interfaces by which chemists and biologists can generate predictions.
Once it is possible to make predictions given only a chemical structure, there are many possible applications of prediction to the drug discovery process.
Predictions do not necessarily need to be very accurate to be useful as long as the goal is to prioritize many compounds rather than make a decision about a single compound.
In this talk I will discuss a specific set of nine off-target activities, “the global counterscreens”, how QSAR models for these are built, and how the models are applied.
Robert Sheridan, Cheminformatics Department
Merck Research Laboratories
One needs three things to successfully use predictions of off-target activities in the drug-discovery process:
1. Consistent assays of the relevant activities, preferably with high throughput.
2. A software infrastructure to generate, validate, and publish QSAR models.
3. Interfaces by which chemists and biologists can generate predictions.
Once it is possible to make predictions given only a chemical structure, there are many possible applications of prediction to the drug discovery process.
Predictions do not necessarily need to be very accurate to be useful as long as the goal is to prioritize many compounds rather than make a decision about a single compound.
In this talk I will discuss a specific set of nine off-target activities, “the global counterscreens”, how QSAR models for these are built, and how the models are applied.