GraPES

GraPES predicts proteins that localize to cellular condensates, particularly stress granules, to support analysis of phase separation-driven membraneless organelle composition.

Key Features:

• MaGS predictor: Uses experimentally-derived protein features to assess the likelihood of a protein localizing to condensates.
• MaGSeq predictor: Applies a purely sequence-based approach to predict a protein's propensity for inclusion in condensates.
• Species-specific models: Includes models trained on stress granule proteins from human (Homo sapiens) and yeast (Saccharomyces cerevisiae).
• Condensate focus: Targets prediction of proteins involved in phase separation and stress granule formation within membraneless organelles.

Scientific Applications:

• Phase separation research: Supports investigation of molecular determinants and proteins involved in phase separation-driven condensate formation.
• Stress granule composition: Enables prediction and analysis of candidate protein constituents of stress granules.
• Comparative studies: Facilitates studies of stress granule proteins across Homo sapiens and Saccharomyces cerevisiae.

Methodology:

GraPES provides two predictive models—MaGS, based on experimentally-derived protein features, and MaGSeq, using sequence-only features—with models trained on human (Homo sapiens) and yeast (Saccharomyces cerevisiae) stress granule proteins.