AMP Autoencoders

Generative models for antimicrobial peptide design: auto-encoders and beyond This repository contains the datasets and generative models for our paper. The code for training and peptide generation can be found on GitHubOur Study can be found here: Springer Nature Link (BioData Mining) Our study presents a comparative analysis of contemporary deep learning models’ generative performance for generating novel antimicrobial peptides. The models examined include Variational Auto-Encoders, a Wasserstein Auto-Encoder, a Recurrent Neural Network and a Language Model. The primary focus of this study is the systematic comparison and evaluation of various methods and sampling options to identify the most suitable model and sampling strategy combination for different use cases. The datasets used for training and comparison are located in the datasets.tar.gz archive.There are three peptide datasets, with precomputed peptide descriptors.- amp_descriptors.csv: The training dataset with precomputed descriptors- uniprot_descriptors.csv: The comparison dataset with UniProt sequences and precomputed descriptors- eskape_pathogens_train_dataset.csv: The training dataset with sequences only The models are located in separate tar.gz archives. For all VAE models the second part of the name denotes the type ofannealing schedule used. The encoder and decoder of the auto-encoder models are saved as separate models.- vae_normal.tar.gz: VAE without annealing- vae_logistic.tar.gz: VAE with logistic annealing- vae_linear.tar.gz: VAE with linear annealing- vae_cyclic.tar.gz: VAE with cyclic annealing- rnn.tar.gz: The RNN model- language_model.tar.gz: Language model and its tokenizer- wae.tar.gz: The Wasserstein auto-encoder Important: All auto-encoders and the RNN are only compatible with Keras below version 3.0! The language model can be used or trained using more up-to-date versions of the frameworks. Problems may occur due to the different Keras and TensorFlow versions (Keras 2.15.0 and TensorFlow 2.15.0 were used for the auto-encoders).