Summarized by Masters of Longevity from Nature - Main.
A report on a proteomics-trained AI ‘virtual cell’ that predicts drug responses in triple-negative breast cancer, highlighting its clinical-application focus.
Key Takeaways
- The model was trained on over 38 million protein measurements from 18 breast-cancer cell lines including 16 TNBC lines.
- Training data included levels of 5,585 protein groups measured before and at 6, 24, and 48 hours after drug treatment.
- The model achieved 88% accuracy predicting responses to 81 drugs withheld from the training set and validated on 501 patient biopsies.
