Extended Data Fig. 4: Toy model illustration for inner workings of DDC.
Toy model illustration for inner workings of DDC (See text within SI for in depth description): a, Simple toy model with 4 true localizations and 2 repeats (color coded), with the number showing the frame of each localization (can also be used to identify each localization for this example). b, The true pairwise distance distribution (PT (Δr)) and the distribution of distances between loci given that at least one is a repeat (PR1(Δr∣Δn = 1)) for the localizations within (a) The number (and probability) for ‘small’ distances and ‘large’ distances for each distribution is above each bar, with an assigned variable (a, b, c, d) used in the calculation of the Likelihood (Lik). We also show the specific pairs of loci under the bars to illustrate how assigning a particular loci to a certain set influences the likelihood calculation. Note: for this specific example blinks only appear with Δn = 1, and hence we ignore the distributions with Δn > 1 (See text). c, Simplified illustration of how Alg. 1 and Alg. 2 work together and assign localizations as a true localization or repeat localization. Multiple steps of the MCMC are shown with different rows (1 to 3) (See Text). Alg. 1 essentially calculates the probability that a localization is a repeat (green bars), if this value is above .5 it is assigned to that set. Alg. 2 varies this calculation by a small amount each step, generating new sets d, The sets assigned in (c) lead to different likelihoods (due to the particular distribution the distance between each pair is assigned (changing (a,b,c,d), note how the specific distances between each pair change with each assigned set), when the distributions of the assigned sets match the correct distributions (those in (B)) Lik is maximized. (See text for further details).
