SyMBac training data, Omnipose checkpoints, and morphology tables for Escherichia coli dormancy-exit imaging

Computational resources accompanying the BioImage Archive study S-BIAD3676, a mother-machine timelapse dataset of Escherichia coli exiting stationary phase. This record contains: symbac_data.zip: two SyMBac synthetic phase-contrast training datasets, comprising 5,000 image-mask pairs representing exponentially growing cells and 5,000 image-mask pairs representing stationary-phase cells. These data were used to train the released Omnipose checkpoints. omnipose.zip: 17 Omnipose checkpoints used to generate the multi-checkpoint segmentation hypotheses deposited with the imaging data. The final checkpoint is stored as epoch 3999 and corresponds to hypothesis index 16, the primary-analysis hypothesis. hypothesis_16_trace_measurements_long.parquet: long-form trace-level morphology measurements for hypothesis 16 (86,653,385 rows). hypothesis_16_cell_cycle_measurements_long.parquet: long-form cell-cycle morphology measurements for hypothesis 16 (4,794,120 rows). analysis_code.zip: all analysis and figure generation code used for this study. The raw Nikon ND2 acquisition, extracted trench-image Zarr, multi-checkpoint segmentation-mask Zarr, and curated lineage tracks are deposited in the BioImage Archive under accession S-BIAD3676 and DOI 10.6019/S-BIAD3676.