Prediction of S-sulfenylation sites using mRMR feature selection and fuzzy support vector machine algorithm

Area of Science:

• Biochemistry
• Bioinformatics
• Computational Biology

Background:

• Cysteine S-sulfenylation is a vital protein post-translational modification.
• It influences critical cellular processes like transcriptional regulation and cell signaling.
• Accurate identification of S-sulfenylation sites is essential for understanding its molecular mechanisms.

Purpose of the Study:

• To develop a novel bioinformatics tool for predicting protein S-sulfenylation sites.
• To improve the accuracy of S-sulfenylation site identification.

Main Methods:

• Developed Sulf_FSVM, a tool utilizing multiple feature extraction techniques.
• Incorporated amino acid factors, binary encoding, and k-spaced amino acid pair composition.
• Employed the maximum relevance minimum redundancy method for feature selection.
• Utilized a fuzzy support vector machine algorithm to address data imbalance and noise.

Main Results:

• Sulf_FSVM demonstrated a Sensitivity of 73.26%, Specificity of 70.78%, and Accuracy of 71.07% via 10-fold cross-validation.
• Achieved a Matthew's correlation coefficient of 0.2971.
• Independent tests confirmed Sulf_FSVM's superior performance compared to existing predictors.

Conclusions:

• Sulf_FSVM is a valuable tool for the accurate prediction of protein S-sulfenylation sites.
• This tool can advance the study of protein S-sulfenylation and its biological roles.