Generation of Beta Barrel Protein Sequences with an LSTM recurrent neural network.
This script uses Deep Learning Neural Networks (LSTM) with Keras on TensorFlow, to generate new beta barrel protein sequences. This script generates FASTA sequences.
Example of a generated beta barrel FASTA sequence:
GVEFDEETVDGRKVKSIITLDGGALVQVQKWDGKSTTIKRKRDGDKLVVECVMKGVTSTRVYERACPTLGGVGNQTTVDNGPDNSGGGDNVNGVAVGFVVVVPGGGTVGSTVGGGVISGVGGVTVDVTTIRVNIVVGRSVGTVVVDTTTVGTTIDSGDTNTVDGDDGTVTKAGGVRVDVVNFVGVGEGVNVPSLLVDKNVVIVRGTDGVNPGVRSYDG
EGNKLVGKFKRVDNGKELIAVREISGNELIQTYTYEGVEAKRIFKKECDAFVGTWKLVSSENFDDYMKEGDKVGGVVNTDVDVVGDATTDVVGDSGRVGNPRRTVVVGGTGVYAYVGSVVQSDGPVSKTDDVGDGGTVDRTSGVGPVDRGTDTDSGGVNAPPGVLTQKQ
TVDITVTYEDGSTATGQLTVDSPTQFKFDMTASDGTRFTGTVQRQSMSRAAQLLPGTWQVTMTNEDGQTGVGGVTDVAIVVGVRDDIVTGTGVGDRNTTSSGDGLVVDVDGDVFGTDGGVRGVTSVVNVVSKRGDTNDGAVTTVVNGTEDGVDDDDTVVSNVVVRVV
Clone the directory:
git clone https://github.com/JesslynJanssen/BetaBarrelLSTM
Then, install all requirements:
sudo apt update && sudo apt full-upgrade && sudo apt install python3-pip python3-pandas python3-numpy python3-tensorflow tensorboard && pip3 install keras
Finally, train the model or generate new beta barrel protein sequences.
Beta barrel FASTA sequence data from the Protein Data Bank is already provided, but you can train the model with other beta barrel proteins or any other type of protein.
To train the network yourself, run the following command:
python3 train betabarrel
To generate new beta barrel FASTA sequences, run the following command:
python3 generate betabarrel
To determine whether the generated sequences may be plausible, do a BLAST and PSIPRED run to align with database protein sequences.