Caribou pipeline for the alignment-free bacterial identification and classification in metagenomics sequencing data using machine learning

This dataset contains sequencing data used to train the models of the Caribou pipeline. We developed this pipeline for alignment-free bacterial identification and classification in metagenomics sequencing data using machine learning. The datasets were derived from the GTDB v.202 database (https://data.gtdb.ecogenomic.org/releases/release202/202.0/) and include training steps using the species representatives, as the benchmark datasets used non-representative whole genomes. We also simulated sequencing reads to evaluate and compare performance on whole genomes and sequencing reads. We provide models and encoding files of CNN-trained models; datasets used for training, validation and testing of models, randomly sampled from representative genomes; and datasets used for benchmarking the method against state-of-the-art methods, randomly sampled from non-representative whole genomes and simulated reads.