dSCOPE

dSCOPE predicts sequence segments critical for liquid-liquid phase separation (LLPS) in proteins to identify Sequence Critical fOr Phase sEparation (SCOPEs) that contribute to formation of membraneless organelles such as P-bodies, nucleoli, and stress granules.

Key Features:

• Curated experimental SCOPE dataset: Curates experimentally validated sequence segments driving LLPS from the literature to form the training dataset.
• Sliding-window feature extraction: Uses a sliding window approach to extract physiological, biochemical, structural, and coding features from protein sequences.
• Random forest integration: Integrates extracted features using a random forest algorithm to predict SCOPEs.
• Predictive performance: Demonstrated satisfactory performance in predicting SCOPEs with precision.
• Proteome-scale analysis: Applied to large-scale analyses of the human proteome, revealing enrichment of protein post-translational modifications and cancer-associated mutations within predicted SCOPEs.
• Pathway association: Identifies that proteins containing predicted SCOPEs are involved in cellular signaling pathways.

Scientific Applications:

• LLPS region identification: Identification of sequence segments critical for LLPS to study formation of membraneless organelles.
• Proteome-wide discovery: Large-scale screening of the human proteome to locate candidate SCOPEs.
• PTM and mutation enrichment analysis: Assessment of enrichment of protein post-translational modifications and cancer-associated mutations within predicted SCOPEs.
• Signaling and disease research: Linking predicted SCOPE-containing proteins to cellular signaling pathways and investigating their roles in cellular organization and disease.

Methodology:

Curates experimentally validated LLPS-driving sequence segments from literature; applies a sliding window to extract physiological, biochemical, structural, and coding features from protein sequences; and trains/applies a random forest algorithm to integrate features and predict SCOPEs.