Extended Data Fig. 3: Estimation of voxel-based PSF model of a lattice light-sheet system from bead data.
The data were collected by imaging beads in agarose gel at sample stage positions from −50 µm to 50 µm, with a step size of 50 nm. Translation of the sample stage will translate the beads both in x and z dimensions in the coordinate of the detection objective. For this data, translating the sample stage by 50 nm, will translate the beads by 40.8 nm in x and 28.9 nm in z relative to the detection objective. (a) Localization in x of the deskewed data used for inverse modelling. (b) The raw data stack of each bead was deskewed using the above translation relationship between x and z dimensions. Only shifts of integer pixels were applied to maintain the photon statistics of the raw data. The deskewed bead stacks were used for inverse modelling. The sub-pixel shifts were incorporated in the forward model so that the estimated PSF model has no skew as in (d). (c) localization of the data used for inverse modelling, sub-pixel shifts were applied to the localization result in (a) to remove the skew effect. The z/x/y bias vs. z are defined as in SI Note 2 Eqs. 2.37 and 2.38. Black lines and shaded areas in localization plots indicate mean and standard deviation of the x, y and z biases over >20 beads. (d) Estimated PSF model. (e) Comparison of an example bead stack and its corresponding forward model. Scale bars, 1 µm (b,d,e).
