ADNPPro: Enhanced Promoter Prediction in Nannochloropsis Using Attention-Based Densenet

Nannochloropsis is a significant genus of marine microalgae, renowned for its strain diversity and high productivity. However, it faces challenges in achieving optimal yields using wild-type strains and traditional cultivation methods. Promoters play a crucial role in regulating gene expression, but current algorithms for recognizing Nannochloropsis promoters have limitations in accuracy and interpretability. To address this problem, we introduce ADNPPro (Attention-based DenseNet Predict Promoter), merging an attention mechanism with a densely connected convolutional neural network. By interpreting the deep learning model, we process an integrated analysis of sequence importance scores, enabling a deeper understanding of key promoter features across various Nannochloropsis strains. Experimental results demonstrate the robust generalization capability of ADNPPro across six Nannochloropsis strains. Motif analysis provides insights into their diverse performance capabilities and enhances the understanding of Nannochloropsis promoter characteristics. The introduction of this framework is set to advance synthetic biology, offering insights into the biological significance of sequence features.