From my understanding, the CNN model trained on HER2 receptor binders and non binders is able to predict the binding of computational library of approximately 1 × 10^8 trastuzumab variants.
Would this model be able to generalize and predict the binding on other antigen peptides or must the model be retrained with new binding data for classifying binders and non binders of a new antigen of interest?
From my understanding, the CNN model trained on HER2 receptor binders and non binders is able to predict the binding of computational library of approximately 1 × 10^8 trastuzumab variants.
Would this model be able to generalize and predict the binding on other antigen peptides or must the model be retrained with new binding data for classifying binders and non binders of a new antigen of interest?