iProStruct2D: Identifying protein structural classes by deep learning via 2D representations

作者：

Highlights：

• Protein classification from multi-view 2D representation of proteins using Jmol.

• CNNs train 13 2D visualizations emphasizing specific properties of protein structure.

• New protein data augmentation and CNN fusion exploits diversity of 2D representations.

摘要

•Protein classification from multi-view 2D representation of proteins using Jmol.•CNNs train 13 2D visualizations emphasizing specific properties of protein structure.•New protein data augmentation and CNN fusion exploits diversity of 2D representations.

论文关键词：Protein classification,Protein visualization,Deep learning,Convolutional neural networks

论文评审过程：Received 30 June 2019, Revised 12 September 2019, Accepted 10 October 2019, Available online 11 October 2019, Version of Record 17 October 2019.

论文官网地址：https://doi.org/10.1016/j.eswa.2019.113019