Standardized Framework and Quantitative Benchmarking: Nuclear Segmentation platforms for Multiplexed Immunofluorescence data in Translational Studies

Paired nuclear ground truth annotated binary masks and segmentation binary masks predicted from different segmentation platforms can be found in external.zip and inhouse.zip. internal.zip was used for development of the nuclear segmentation evaluation pipeline while external.zip hosts data that was/can be used for benchmarking of nuclear segmentation performance for multiplexed immunofluorescence workflows in translational studies. external.zip contains ground-truth annotated binary masks from another dataset in synapse (https://doi.org/10.7303/SYN27624812). README_DAPI.md contains information about where to locate the raw DAPI nuclear intensities used by the different segmentation platforms to predict nuclei. The naming convention of the ROIs used is as follows: "Sample ID_ULHC" where ULHC refers to the coordinates of the upper-left hand corner pixel of the ROI within the whole sample field, delimited by an underscore ("_"). E.g. For ROI "Bph2_45174_14597_160_246": "Bph2_45174_14597" is the sample ID and (246,160) is the ULHC pixel. Citing Please cite these sources if you use this data Sankaranarayanan, A., Khachaturov, G., Smythe, K.S. & Mittal, S. Quantitative benchmarking of nuclear segmentation algorithms in multiplexed immunofluorescence imaging for translational studies. Commun Biol 8, 836 (2025). https://doi.org/10.1038/s42003-025-08184-8 Sankaranarayanan, A., Khachaturov, G., Smythe, K. S. & Mittal, S. Standardized framework and quantitative benchmarking: nuclear segmentation platforms for multiplexed immunofluorescence data in translational studies. Zenodo (2024). https://doi.org/10.5281/zenodo.12775616