SenSeqNet: A Deep Learning Framework for Cellular Senescence Detection from Protein Sequences

Cellular senescence, characterized by the irreversible cessation of division in normally proliferating cells due to various stressors, presents a significant challenge in the treatment of age-related diseases. Understanding and accurately detecting cellular senescence is crucial for identifying potential therapeutic targets. However, traditional wet lab assays for detecting cellular senescence are time-consuming and labor-intensive, limiting research and drug development efficiency. There is an urgent need for computational tools allowing swift and accurate detection of cellular senescence from protein sequences. We propose SenSeqNet, a novel deep learning framework for detecting cellular senescence directly from protein sequences. The framework begins with feature extraction using the Evolutionarily Scaled Model (ESM-2), a state-of-the-art protein language model that captures evolutionary information and complex sequence patterns. The extracted embeddings are then passed through a hybrid architecture consisting of long short-term memory (LSTM) networks and convolutional neural networks (CNNs) to further refine and learn from the embedded information. SenSeqNet achieved a final accuracy of 83.55% on independent testing, surpassing various machine learning and deep learning architectures. This performance underscoring the robustness and effectiveness of SenSeqNet for detecting cellular senescence from protein sequences. These results provide a solid foundation for future research on aging and age-related therapeutics.