The PathoFact2_datasets.tar.gz archive contains datasets used for training, validation, and benchmarking of PathoFact 2.0. The dataset is organised into the following folders: 1. BenchMarking/ Contains test datasets used to benchmark the performance of PathoFact 2.0 against other prediction tools. Toxin_module/ 1000_TEST_non-toxin_ToxinPred2BenchMarking.faa Non-toxin test dataset used when benchmarking PathoFact 2.0 against ToxinPred2. 1000_TEST_toxin_ToxinPred2BenchMarking.faa Toxin test dataset used when benchmarking PathoFact 2.0 against ToxinPred2. VF_module/ VF_Test_dataset_VirulentHunterBenchmarking.faa Virulence factor test dataset used when benchmarking PathoFact 2.0 against VirulentHunter. non-VF_Test_dataset_VirulentHunterBenchmarking.faa Non-virulence-factor test dataset used for benchmarking PathoFact 2.0 against VirulentHunter. 2. Toxin_module/ Contains datasets used for training and validating the PathoFact 2.0 Toxin prediction module. Toxin-related.faa non-toxin.faa splits/ — datasets divided into 80% training and 20% test sets: TRAIN_Positive_TOX.faa TRAIN_Negative_TOX.faa Test_Positive_TOX.faa Test_Negative_TOX.faa 3. VF_module/ Contains datasets used for training and validating the PathoFact 2.0 Virulence Factor prediction module. VF_dataset.faa non-VF.faa splits/ — datasets divided into 80% training and 20% test sets: TRAIN_Positive_VF.faa TRAIN_Negative_VF.faa Test_Positive_VF.faa Test_Negative_VF.faa The DOME-ML_PathoFact2.json is the DOME, the community standard for transparent machine learning json file created for PathoFact 2.0.