Microscopy data, analysis code, and segmentation models for Phenotypic drug susceptibility testing for Mycobacterium tuberculosis variant bovis BCG in 12-hour doi-10-17044-scilifelab-26927146-0 https://doi.org/10.17044/SCILIFELAB.26927146 Swedish National Data Service Svensk nationell datatjänst Landing page Microscopy data, analysis code, and segmentation models for Phenotypic drug susceptibility testing for Mycobacterium tuberculosis variant bovis BCG in 12-hour doi-10-17044-scilifelab-26927146-0 https://doi.org/10.17044/SCILIFELAB.26927146 Tran, Buu Minh Karempudi, Praneeth Swedish National Data Service Svensk nationell datatjänst Landing page The study aimed to improve the speed of drug susceptibility testing for tuberculosis (TB), particularly drug-resistant strains. Current phenotypic drug susceptibility testing (pDST) methods take at least two weeks. This study used microfluidic chips, microscopy, and deep neural network algorithms to monitor the growth of Mycobacterium bovis Bacillus Calmette–Guérin (BCG) and Mycobacterium smegmatis in the presence of antibiotics. The findings suggest that pDST for TB could potentially be completed in less than 12 hours for slow-growing mycobacteria, offering a significant improvement in diagnosing drug-resistant TB. The entry contains raw microscopy data, analysis code and output, and code to generate figures. Access to data through an external actor. Åtkomst till data via extern aktör.