Machine Learning Model for Identifying Antioxidant Proteins Using Features Calculated from Primary Sequences.

Area of Science:

• Biochemistry and Molecular Biology
• Computational Biology
• Bioinformatics

Background:

• Antioxidant proteins are vital for cellular protection against reactive oxygen species (ROS).
• An imbalance between ROS and antioxidant defenses contributes to aging and disease.
• Accurate identification of antioxidant proteins is crucial for research and therapeutic development.

Purpose of the Study:

• To develop and validate a machine learning-based computational model for predicting antioxidant proteins.
• To evaluate the performance of various machine learning and deep learning algorithms using sequence features.
• To establish a robust model for identifying potential antioxidant candidates.

Main Methods:

• A machine learning approach was employed using a benchmark set of protein sequencing data.
• The Random Forest algorithm was selected and optimized based on sequence features.
• Model performance was assessed using 10-fold cross-validation and validated on three independent datasets.

Main Results:

• The Random Forest model achieved a high accuracy of 84.6% for antioxidant protein identification.
• The model demonstrated balanced sensitivity (81.5%) and specificity (85.1%) on the training dataset.
• Validation on independent datasets confirmed the model's superior performance compared to existing methods.

Conclusions:

• The proposed machine learning model, particularly Random Forest, is effective for identifying antioxidant proteins.
• This computational approach significantly aids in the rapid detection of antioxidant candidates.
• The model holds promise for advancing research in oxidative stress and related health conditions.