Antioxidative Peptide Predictor

Antioxidative peptide predictor (AnOxPP) is an efficient tool to predict the antioxidative activity of peptides by using the BiLSTM neural network and the optimized amino acid descriptors SDPZ27. The non-redundant code SDPZ27 shows efficient conversion of sequence features, and interprets the importance of four decisive features of AnOxPs: Steric properties > Hydrophobic properties > Electronic properties > Hydrogen bond contributions. By learning the key sequence/structure features of AnOxPs converted by SDPZ27, AnOxPP exhibits more accurate prediction than the existing model. AnOxPP will contribute to understanding the structure−activity relationship of AnOxPs and provide a methodological reference for the application of deep learning to study bioactive peptides.

You can get the following purposes through AnOxPP: (i) Predict unknown peptides in a single or batch manner; (ii) View the sequence characteristics of AnOxPs including length, mass, terminals and amino acid composition; (iii) Screen the predicted AnOxPs (pre-AnOxPs) from the 'Pre-Libraries', which is constructed by using different enzymes to simulate the hydrolysis of 21249 proteins in 60 species; (iv) Download datasets of AnOxPs, non-AnOxPs and pre-AnOxPs.