ANOX

ANOX predicts antioxidant proteins using machine learning to identify proteins involved in oxidative stress and related conditions such as anti-aging, altitude sickness, coronary heart disease, and cancer.

Key Features:

• Feature Extraction: Uses Frequency Matrix Features (FRE), Amino Acid and Dipeptide Composition (AADP), Evolutionary Difference Formula Features (EEDP), k-Separated Bigrams (KSB), and PSI-PRED secondary structure (PRED).
• Feature Optimization: Ranks and selects features with the Max-Relevance-Max-Distance (MRMD) algorithm and attains optimal performance using the top 1170 ranked features.
• Performance Evaluation: Validated by 5-fold cross-validation and jackknife testing, reporting AUCs of 0.930 and 0.935 and outperforming the AOPs-SVM benchmark (AUCs 0.869 and 0.885).

Scientific Applications:

• Antioxidant protein prediction: Predicts antioxidant proteins to facilitate studies of oxidative stress and their roles in anti-aging, altitude sickness, coronary heart disease, and cancer.
• Structural and compositional insight: Provides information on structural and compositional characteristics that define antioxidant proteins.
• Functional exploration: Enables investigation of protein functions related to oxidative stress and its implications in various diseases.

Methodology:

Extracted sequence- and evolution-derived features, ranked and selected features using MRMD, applied machine learning for prediction, and validated performance with 5-fold cross-validation and jackknife testing.