A multi-label predictor for identifying the subcellular locations of singleplex and multiplex eukaryotic proteins

Area of Science:

• Bioinformatics
• Computational Biology
• Proteomics

Background:

• Subcellular protein localization is crucial for understanding protein function.
• Existing predictors often fail to account for proteins with multiple or dynamic locations.
• Accurate prediction of subcellular localization is essential for high-throughput proteomic annotation.

Purpose of the Study:

• To develop a novel predictor, Euk-ECC-mPLoc, capable of handling both single and multiple subcellular locations in eukaryotic proteins.
• To improve the accuracy and efficiency of predicting eukaryotic protein subcellular localization.
• To provide a user-friendly web server for accessing the predictor.

Main Methods:

• Developed Euk-ECC-mPLoc using a multi-label learning approach.
• Integrated gene ontology and dipeptide composition information to exploit correlations between subcellular locations.
• Validated the predictor on a stringent benchmark dataset of eukaryotic proteins using jackknife cross-validation.

Main Results:

• Euk-ECC-mPLoc achieved an average success rate of 69.70% and an overall success rate of 81.54%.
• The predictor demonstrated significantly improved success rates for small subsets of proteins.
• The developed predictor can identify eukaryotic proteins across 22 distinct subcellular locations.

Conclusions:

• Euk-ECC-mPLoc is a promising tool for predicting subcellular locations of both single and multiplex eukaryotic proteins.
• The predictor offers a significant improvement over existing methods, particularly for complex cases.
• Euk-ECC-mPLoc is available as a free, user-friendly web server, facilitating high-throughput biological research.