Abstract
The transport of antigenic peptides from cytoplasm to the endoplasmic reticulum (ER) via transporter associated with antigen processing (TAP) is a critical step during the presentation of tumor neoantigens. The application of computational approaches significantly speed up the analysis of this biological process. Here, we present a tool named DeepTAP for TAP-binding peptide prediction, which employs a sequence-based multilayered recurrent neural network (RNN). Compared with traditional machine learning and other available prediction tools, DeepTAP achieves state-of-the-art performance on the benchmark datasets. The source code and dataset of DeepTAP are available freely via GitHub at https://github.com/zjupgx/DeepTAP.
Competing Interest Statement
The authors have declared no competing interest.