Protein Corona Formation Prediction on Engineered Nanomaterials

Recent nanotechnology advances have catalyzed sev-eral different types of engineered nanomaterials (ENMs). The nanomaterial classification interprets to identifying any particle that is smaller than hundred nanometer. Protein corona (PC) is an agglomeration of proteins that form on an ENM in organic fluids. Machine learning techniques can be useful to predict the PC formation and interaction within an ENM. In this paper, we develop a random forest model for PC formation prediction on ENMs. Further, we leverage the deep neural network (DNN) technique to accurately and efficiently predict PC formation. We also present an architecture optimization of the trained DNN model to create practically instantaneous inferences. We preform simulation study to show effectiveness of our proposed model. Experiments show that the DNN model can achieve 83.81% accuracy in PC classification on ENMs, while can significantly improve the classification performance.