Data and analysis code for "Fused deep learning enables 6D single-molecule localization in polarization-resolved microscopy"

This dataset contains the processed data used to produce the figures, together with the analysis and plotting code, for a study of six-dimensional (6D) single-molecule localization microscopy, in which the three-dimensional position and three-dimensional orientation of individual dipole emitters are measured at the same time. The imaging was performed on a polarization-resolved fluorescence microscope using a double-helix point spread function (DHPSF) phase mask, and the two polarization channels were analyzed with a fused deep-learning pipeline that combines an orientation-estimation network (mDeep-SMOLM) with a 3D-localization network (mDeepSTORM3D). The data are organized by figure and provided in non-proprietary formats (.csv, .txt, .tif). The 6D localization tables give the lateral position in nanometers, the axial position in nanometers, the polar and azimuthal orientation angles in degrees, and the rotational wobble solid angle in steradians, each accompanied by a column dictionary. Two experimental systems are represented, Nile red on 2-micrometer DPPC-coated silica beads and single Rhodamine B molecules in poly(methyl methacrylate) films, along with simulation studies used to characterize network accuracy and DHPSF orientation degeneracies. A snapshot of the analysis and plotting code is included so that the analyses and figures produced with it can be regenerated. The dataset may be useful for reproducing figures from the associated publication, for benchmarking 6D localization and orientation-estimation methods, for training or evaluating single-molecule orientation networks, and for reusing the 6D localization tables in downstream analyses of molecular position and orientation.