Detecting Succinylation sites from protein sequences using ensemble support vector machine

Area of Science:

• Biochemistry
• Proteomics
• Computational Biology

Background:

• Lysine succinylation is a crucial post-translational modification regulating protein conformation and cellular functions.
• Accurate identification of succinylation sites is essential for understanding its mechanisms.
• Experimental methods for site identification are time-consuming and labor-intensive.

Purpose of the Study:

• To develop a rapid and accurate computational method for predicting protein succinylation sites.
• To improve the understanding of the succinylation mechanism through computational analysis.

Main Methods:

• Integration of multiple features: amino acid composition, binary encoding, physicochemical properties, and grey pseudo amino acid composition.
• Application of feature selection using information gain.
• Training and validation using Support Vector Machine (SVM) and ensemble learning algorithms.

Main Results:

• Achieved 89.14% accuracy and 0.79 MCC on the training dataset via 10-fold cross-validation.
• Demonstrated 84.5% accuracy and 0.2 MCC on an independent dataset.
• The developed method shows high promise for succinylation site prediction.

Conclusions:

• The developed computational method offers a promising and efficient approach for predicting succinylation sites.
• Findings contribute to a deeper understanding of the succinylation mechanism.
• Source code and data are publicly available for further research.