pCysMod

pCysMod predicts cysteine post-translational modification sites in proteins, computationally identifying S-nitrosylation, S-palmitoylation, S-sulfenylation, S-sulfhydration, and S-sulfinylation to support studies of autophagy, redox homeostasis, and cell signaling.

Key Features:

• Deep Learning Framework: Employs a deep learning model that integrates several protein sequence features and is optimized using particle swarm optimization algorithms.
• Comprehensive Prediction Capabilities: Predicts multiple cysteine modifications including S-nitrosylation, S-palmitoylation, S-sulfenylation, S-sulfhydration, and S-sulfinylation, which are relevant to autophagy, redox homeostasis, and cell signaling.
• Benchmark Dataset: Trained on a benchmark dataset comprising experimentally verified cysteine sites curated from literature and other databases.
• Performance Metrics: Evaluated by cross-validation with reported average AUCs of 0.793 for S-nitrosylation, 0.807 for S-palmitoylation, 0.796 for S-sulfenylation, 0.793 for S-sulfhydration, and 0.876 for S-sulfinylation.

Scientific Applications:

• Regulatory mechanism analysis: Facilitates identification of cysteine modification sites to study regulatory mechanisms of cysteine-mediated processes.
• Disease-related studies: Aids investigation of roles of cysteine modifications in diseases, including cancers and diabetes.
• Protein biochemistry and molecular biology research: Supports studies on protein function and signaling by providing multi-type cysteine modification predictions.

Methodology:

Computational methods include a deep learning model integrating protein sequence features, optimization with particle swarm optimization, training on an experimentally verified benchmark dataset of cysteine sites, and performance evaluation via cross-validation reporting AUCs for each modification type.