Subcellular localization prediction of apoptosis proteins based on evolutionary information and support vector...

Area of Science:

• Bioinformatics
• Computational Biology
• Molecular Biology

Background:

• Accurate prediction of subcellular localization is crucial for understanding protein function and cellular processes.
• Apoptosis proteins play key roles in programmed cell death, making their localization prediction important.

Purpose of the Study:

• To develop a high-quality sequence encoding scheme for predicting the subcellular location of apoptosis proteins.
• To introduce novel evolutionary-conservative information and a golden section-based segmentation method for protein sequence representation.

Main Methods:

• Utilized evolutionary-conservative information to represent protein sequences.
• Applied a golden section mathematical proportion to segment Position-Specific Scoring Matrices (PSSM).
• Employed Support Vector Machine (SVM) for prediction and validated using a jackknife test.

Main Results:

• The golden section segmentation method outperformed approaches without segmentation.
• Achieved high overall accuracy rates of 98.98% for ZD98 and 91.11% for CL317.
• Demonstrated the proposed method's effectiveness and potential complementary role to existing techniques.

Conclusions:

• The proposed feature representation is powerful, leading to significant improvements in prediction accuracy.
• The method achieves state-of-the-art performance in predicting the subcellular location of apoptosis proteins.
• This approach offers a valuable advancement for bioinformatics and related fields.