AntiOxiPred: Predicting Antioxidant Peptides using Machine Learning Techniques and BLAST

AntiOxiPred predicts antioxidant peptides from primary sequences with peak AUC 0.956 using a CatBoost-plus-BLAST ensemble on an independent testing set.

• Why it matters: Antioxidant peptides are promising for chronic-disease oxidative damage, but discovering them efficiently needs accurate computational prediction methods.
• What they did: The study trained on 1,511 experimentally validated AOPs and 1,511 randomly generated non-AOPs, using one-hot, compositional features, PLM embeddings, and fused features, then combined CatBoost with BLAST.
• The result: On external datasets of 150 peptides (75 AOPs/75 non-AOPs) and 248 peptides (124 AOPs/124 non-AOPs), AntiOxiPred outperformed existing state-of-the-art predictors and is provided as a web server and tools.