Fig. 4: Experimental determination of the optimal parameters for the Anderson–Darling null hypothesis test.
From: Utilizing the predictable binding kinetics of DNA-PAINT to denoise super-resolution images
a Montage image of the GATTA-PAINT nanorulers exhibiting the characteristic linear arrangement of DNA-binding sites in 3 marks, with a mark-to-mark spacing of 80 nm. b A histogram showing the distribution of the number of binding events per individual cluster of the GATTA-PAINT nanorulers. c The fraction of non-rejected clusters as a function of the significance level for the Anderson–Darling test. d A super-resolved image of an E-cadherin negative control sample that has not been exposed to docking strands. Image is representative of 7 independent experiments. e The percentage of clusters with a given number of binding events (orange line) and the percentage of these clusters that are rejected by the Anderson–Darling test (blue line). f Total percentage of non-specific clusters rejected as a function of the minimum number of binding events required per cluster. Clusters with fewer binding events are directly rejected, while the Anderson–Darling test is applied to clusters with more binding events. Error bars indicate mean ± standard deviation (E-cadherin negative control samples n = 7).
