pLMSNOSite

pLMSNOSite predicts protein S-nitrosylation (SNO) sites by integrating supervised embeddings with ProtT5 contextualized protein language model embeddings to identify nitric oxide-mediated post-translational modification sites.

Key Features:

• Model architecture: Intermediate fusion-based stacked generalization ensemble for SNO site prediction.
• Embedding integration: Combines embeddings from supervised embedding layers and the ProtT5 contextualized protein language model.
• Prediction target: Predicts protein S-nitrosylation (SNO) sites, a nitric oxide-mediated post-translational modification.
• Evaluation: Validated on an independent test set of experimentally identified SNO sites with MCC 0.340, sensitivity 0.735, and specificity 0.773.
• Biological scope: Applicable to studying SNO-related signaling and regulation across animals and plants.

Scientific Applications:

• Signaling and regulation studies: Mapping S-nitrosylation sites to investigate nitric oxide-mediated signaling and regulatory effects on protein function.
• Functional analysis of PTMs: Studying SNO involvement in immune response, protein stability, transcription regulation, DNA damage repair, redox regulation, and protection against oxidative stress.
• Mechanistic investigation: Advancing understanding of physiological and pathological mechanisms underlying S-nitrosylation.

Methodology:

Uses an intermediate fusion-based stacked generalization approach that integrates supervised embedding layers with ProtT5 contextualized protein language model embeddings and is evaluated on an independent test set of experimentally identified SNO sites (MCC 0.340, sensitivity 0.735, specificity 0.773).