MRSLpred

MRSLpred predicts multi-label subcellular localization of mRNA sequences to support analysis of mRNA distribution and cellular compartmentalization.

Key Features:

• Multi-Label Prediction Capability: Predicts multiple subcellular localization labels per mRNA molecule, enabling comprehensive characterization of mRNA distribution.
• High Performance and Efficiency: Employs an XGBoost-based classifier with an average area under the receiver operator characteristic (AUROC) of 0.709 for distinguishing subcellular localizations.
• Hybrid Methodology: Integrates alignment-free mRNA sequence composition analysis with alignment-based motif search, with the combined hybrid model achieving an AUROC of 0.742.
• Scalability: Supports genome-scale processing for large transcriptome datasets.

Scientific Applications:

• mRNA Dynamics: Supports studies of mRNA dynamics by predicting localization patterns across cellular compartments.
• Cellular Compartmentalization: Aids mapping of mRNA distributions to subcellular locations to inform compartment-specific functions.
• Gene Expression Regulation: Facilitates investigation of gene expression regulation through localization-informed functional inference.
• Complex Cellular Processes: Enables analysis of processes involving multiple localizations, such as signaling pathways and cellular stress responses.

Methodology:

MRSLpred combines alignment-free analysis of mRNA sequence composition and alignment-based motif search, integrating these features into an XGBoost-based classifier for multi-label subcellular localization prediction.