内蒙古民族大学学报（自然科学版）2025，Vol.40Issue(3)：62-70,9.DOI:10.14045/j.cnki.15-1220.2025.03.009
基于SE-Connection Pyramid Network网络的蛋白质-DNA结合位点预测
Prediction of Protein-DNA Binding Sites Based on SE-Connection Pyramid Network
摘要
Abstract
The binding process of protein and DNA is very important for gene expression and regulation in bio-logical systems.The accurate protein-DNA binding sites on protein is of great significance for understanding life ac-tivities.In this paper,a protein-DNA binding site prediction model based on the SE-Connection Pyramid Network(SECP-Net)network is proposed.The model combines the pyramid structure and the Squeeze-and-Excitation mod-ule,which can effectively extract multi-scale features and dynamically adjust the weight of feature channels.Experi-mental verification on the PDNA-62 and PDNA-224 datasets shows that the SECP model outperforms traditional models in multiple evaluation indicators,demonstrating its good performance in protein-DNA binding site prediction.The experimental results show that the method can accurately predict protein-DNA binding sites.关键词
卷积神经网络/特征提取/蛋白质序列Key words
convolutional neural network/feature extraction/protein sequence分类
计算机与自动化引用本文复制引用
张辰瑞,姜静清,赵芳,宋佳智..基于SE-Connection Pyramid Network网络的蛋白质-DNA结合位点预测[J].内蒙古民族大学学报（自然科学版）,2025,40(3):62-70,9.基金项目
国家自然科学基金项目(62162050) （62162050）
内蒙古民族大学博士科研启动基金项目(KYQD23006) （KYQD23006）
