EP-pred

Platform based on machine learning likely to have the following characteristics:

• It will extract features from input esterase sequences in FASTA format.

• It will filter those sequences that are too different from the training dataset because of the applicability domain phenomenon, since using a model to predict on proteins that the system has not seen and trained on before will likely produce errors.

• Finally, it will classify the sequences based on its promiscuity, >20 will be considered promiscuous and <20 non promiscuous.