Predicting subcellular localization of multisite proteins using differently weighted multi-label k-nearest neighbors...

Area of Science:

• Bioinformatics
• Computational Biology
• Proteomics

Background:

• Accurate protein subcellular localization is crucial for protein function.
• Existing bioinformatics methods struggle with large datasets and multisite protein localization.
• Novel feature extraction is needed for complex protein localization prediction.

Purpose of the Study:

• To develop an advanced bioinformatics approach for predicting multiplex protein subcellular localization.
• To achieve high prediction accuracy for proteins with multiple cellular locations.
• To enhance the understanding of protein function through improved localization prediction.

Main Methods:

• Developed a feature fusion strategy using pseudo amino acid composition (PseAAC) and entropy density.
• Employed a multi-label k-nearest neighbors algorithm with weighted attributes for prediction.
• Evaluated the method on a dataset including single-site and multisite protein localizations.

Main Results:

• The proposed method demonstrated a significant improvement in the overall accuracy of multiplex protein subcellular localization prediction.
• Feature fusion effectively addressed challenges posed by multisite localization.
• The approach proved superior to existing single-feature extraction methods.

Conclusions:

• The developed method offers a more comprehensive prediction of protein subcellular localization.
• Improved localization prediction aids in a deeper understanding of protein functions.
• This advancement facilitates better identification and monitoring of potential drug targets.