DNA promoter recognition study based on fishing-optimized TCN-Attention

DNA promoters are key regions in the regulation of gene expression, determining the location and frequency of transcription initiation. Accurate identification of promoters is essential for understanding gene regulatory mechanisms, disease diagnosis and treatment. However, existing methods still face challenges in dealing with complex sequence information and spatio-temporal dependencies, especially when facing different types of promoters, the generalization ability of the models needs to be improved. In this paper, a two-branch input model based on Fishing-Optimized Temporal Convolutional Network with Attention mechanism (Fishing-Optimized TCN-Attention) is proposed for efficient identification of DNA promoters. The model adopts a combination of image input and feature sequence input, extracts Gram's Corner field image features by TCN and temporal sequence features by BiGRU, respectively, and introduces a fishing algorithm to optimize the relevant hyperparameters. The experimental results show that the model exhibits high accuracy and robustness on the public dataset, which provides a valuable reference for subsequent research.