DBP-CNN: Deep learning-based prediction of DNA-binding proteins by coupling discrete cosine transform with two-dimensional convolutional neural network

作者：

Highlights：

• Developed a novel deep learning based predictor namely DBP-CNN for prediction of DBPs.

• The local decisive information from PSSM was investigated by tetra-slicing.

• DCT compression technique was applied to eliminate noisy data.

• 2D CNN, XGB, ERT, and RF were used as classification algorithms.

• Our novel feature encoder PSSM-TS-DCT with 2D CNN achieved the best performance.

摘要

•Developed a novel deep learning based predictor namely DBP-CNN for prediction of DBPs.•The local decisive information from PSSM was investigated by tetra-slicing.•DCT compression technique was applied to eliminate noisy data.•2D CNN, XGB, ERT, and RF were used as classification algorithms.•Our novel feature encoder PSSM-TS-DCT with 2D CNN achieved the best performance.

论文关键词：DNA-binding proteins,Convolutional neural network,Discrete cosine transform,eXtreme Gradient Boosting,Position-specific scoring matrix

论文评审过程：Received 29 June 2021, Revised 5 February 2022, Accepted 21 February 2022, Available online 1 March 2022, Version of Record 9 March 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.116729