AFP-Predictor

The prediction framework utilizes ESM2 embeddings coupled with multilayer perceptron (MLP) models. The prediction only demands FASTA-formatted protein sequences containing the 20 canonical amino acids as input. Typically, each prediction task takes several tens of seconds to finish and return results. A query sequence is classified as an antifreeze protein when its predicted score is no less than 0.7796.

Please input a protein sequence (one sequence each time), with amino acid number ranging from 10 to 2000.