benchmark_data: Data used to train and evaluate the classification models. giant_virus_data: Sequences and metadata of giant viruses identified in public metagenomes. neural_network_training: Code used to train geNomad's neural network-based classification model. provirus_data: Data used to train and evaluate the conditional random field model employed by geNomad to identify provirus regions. reference_sequences: Sequences of chromosomes, plasmids, and viruses that were used to build geNomad's marker dataset and to generate the training data for the classification models.