New study predicts steroid activities with high accuracy, revolutionizing drug development.
The researchers studied 31 steroids to see how well their structures predict their activities. They found that certain molecular descriptors closely matched the binding affinities of corticosteroids. By using these descriptors, they could predict the activities of 10 other steroids accurately. The study highlights the importance of using representative molecules in training sets and not relying on average activity levels for validation.
