Abstract:Secondary structure prediction is a key step in understanding protein function and biological properties and is highly important in the fields of new drug development, disease treatment, bioengineering, etc. Accurately predicting the secondary structure of proteins helps to reveal how proteins are folded and how they function in cells. The application of deep learning models in protein structure prediction is particularly important because of their ability to process complex sequence information and extract meaningful patterns and features, thus significantly improving the accuracy and efficiency of prediction. In this study, a combined model integrating an improved temporal convolutional network (TCN), bidirectional long short-term memory (BiLSTM), and a multi-head attention (MHA) mechanism is proposed to enhance the accuracy of protein prediction in both eight-state and three-state structures. One-hot encoding features and word vector representations of physicochemical properties are incorporated. A significant emphasis is placed on knowledge distillation techniques utilizing the ProtT5 pretrained model, leading to performance improvements. The improved TCN, achieved through multiscale fusion and bidirectional operations, allows for better extraction of amino acid sequence features than traditional TCN models. The model demonstrated excellent prediction performance on multiple datasets. For the TS115, CB513 and PDB (2018-2020) datasets, the prediction accuracy of the eight-state structure of the six datasets in this paper reached 88.2%, 84.9%, and 95.3%, respectively, and the prediction accuracy of the three-state structure reached 91.3%, 90.3%, and 96.8%, respectively. This study not only improves the accuracy of protein secondary structure prediction but also provides an important tool for understanding protein structure and function, which is particularly applicable to resource-constrained contexts and provides a valuable tool for understanding protein structure and function.

Protein Secondary Structure Prediction Using Multi-Modal Neural Network

Hybrid Model Of Neural Network And Hidden Markov Model For Protein Secondary Structure Prediction

Improving Protein Secondary Structure Prediction Using a Multi-Modal BP Method

Prediction Protein Secondary Structure by Improved BP Neural Network

Protein Secondary Structure Prediction Based on Multi-scale Convolution and Transformer

Protein secondary structure prediction with a neural network.

Predicting protein secondary structure based on neural network ensemble

Prediction of Protein Supersecondary Structures Based on the Artificial Neural Network Method

Protein Secondary Structure Prediction Using Neural Network and Simulated Annealing Algorithm

Protein Secondary Structure Prediction Based On Bp Neural Network And Quasi-Newton Algorithm

Protein Secondary Structure Prediction Using Deep Multi-scale Convolutional Neural Networks and Next-Step Conditioning

A Dynamic Bayesian Network Approach to Protein Secondary Structure Prediction.

Combining Physico-chemical Properties with PSSM for Protein Secondary Structure Prediction Using BP Neural Network

A Deep Convolutional Neural Network to Improve the Prediction of Protein Secondary Structure

MCP: a Multi-Component learning machine to Predict protein secondary structure

A method based on improved Bayesian inference network model and hidden Markov model for prediction of protein secondary structure

An Algorithm Based On Improved Bayesian Inference Network Model For Prediction Protein Secondary Structure

Prediction of protein secondary structure by the improved TCN-BiLSTM-MHA model with knowledge distillation

Protein Secondary Structure Co-Training Prediction Method

Effective Local and Secondary Protein Structure Prediction by Combining a Neural Network-Based Approach with Extensive Feature Design and Selection without Reliance on Evolutionary Information

Hybrid model of ANN/HMM for protein secondary structure prediction