Fig. 2
From: Ultrafast data mining of molecular assemblies in multiplexed high-density super-resolution images
Validation of the dTC algorithm on simulated SMLM data. a Simulated three-color single-molecule localization image in which two different molecular patterns were mixed rand randomly orientated. ~ 1000 such patterns in total were seeded onto 10240 × 10240 nm2 canvas, and each pattern has ~10 localizations in each channel. b The triple-correlation profile in which two local maxima were found representing the two input geometric configurations. c Comparison between the ftTC and dTC algorithm in computing the triple-correlation. Three-color images of the same canvas size (10,240 × 10,240 nm2) but different molecular densities were submitted for dTC computation. These coordinates were rendered onto 2048 × 2048 pixelized image (pixel size = 5 × 5 nm2) for ftTC computation. Error Bars in dTC computation stands for the SD of five computation replicates while ftTC computation was only performed once due to its computational costs. Source data for Fig. 2c are provided as a Source Data file
