GraphDTA for predicting drug–target binding affinity

Identifier:  TFM:2103

Handle: https://hdl.handle.net/20.500.11797/TFM2103

Authors:  Gálvez Rísquez, Natzaret

Abstract:
Drug development is costly and time-consuming, but drug repurposing offers a quicker alternative by finding new uses for approved drugs. Knowing drug-protein interactions is crucial for this process. GraphDTA, a model that predicts drug-target affinity using graph neural networks, outperforms established models like DeepDTA and WideDTA in accuracy and adaptability. While GraphDTA relies on SMILES representations, incorporating 3D structural data could retain vital molecular information, enhancing predictions. This highlights GraphDTA's potential to improve with further development, making it a robust and versatile tool for drug-target affinity predictions