biostudies-otherUnknownFernando SolaMICIU/AEI/10.13039/501100011033 and FSE+Spanish Ministry of UniversitiesMICIU/AEI/ 10.13039/501100011033 and ERDF/EUSpanish Ministry of Science and Innovation (Health Institute Carlos III)Homo sapiens (human)https://www.ebi.ac.uk/biostudies/studies/S-BSST2698biostudies-otherPI23/01522FPU23/04004FPU22/00336PID2022-139798OB-I00RYC2023-045030-IMarina R. PulidoInma HernándezRafael AyalaLorena López-CereroDavid RuizDaniel AyalaFernando SolafalseBAKTA genome annotations and AMR gene profiles for ~20,000 NCBI bacterial isolates from NDAROThis dataset contains genome annotations and antimicrobial resistance (AMR) gene features for approximately 20,000 bacterial isolates obtained from the NCBI National Database of Antibiotic-Resistant Organisms (NDARO). Apart from that, derived datasets for AMR prediction are provided. Data contents: - ncbi_isolates_bakta_annotations.7z: BAKTA genome annotations for all isolates, including GFF3, GenBank, and protein sequence files. - ndaro_baseline.csv: AMR gene presence/absence features from NDARO dataset. - UniRef50 and UniRef90 clustered gene annotations datasets: bakta50 and bakta90 (files .npz, _assemblies.pkl and _columns.pkl for each one). - UniRef50 and UniRef90 clustered AMRFinder genes annotations datasets: bakta50_amr and bakta90_amr (files .npz, _assemblies.pkl and _columns.pkl for each one). The annotations were generated to evaluate alternative genomic feature representations for machine learning-based AMR prediction. The dataset enables reproducibility and extension of methods comparing compact AMR-focused features versus genome-wide representations for predicting antibiotic resistance. Note: The organism field was set to "Homo sapiens" due to form limitations, but this dataset contains bacterial isolates (multiple species) from NCBI NDARO, not "Homo sapiens".2026-02-09T00:00:00Z2026-06-19T13:58:24.126Z2026-02-09T11:28:37.346ZS-BSST2698