A Stochastic Multiscale Model that Explains the Segregation of Axonal Microtubules and Neurofilaments in Neurological Diseases Supporting Information

To calculate the RDF, for each time frame we used the centers of the neurofilaments in the domain [−0.2μm, 0.2μm] × [−0.2μm, 0.2μm] as reference points. We chose to use reference points in a smaller domain in order to avoid boundary effects. For each reference point we binned the center-to-center distances (r) between the reference neurofilament and all other neurofilaments with a bin size ∆r = 1nm, and normalized by the factor 2πr∆rρN where ρN is the density of neurofilaments in the whole domain. We then took the average over all reference points for each time frame and then over all time frames to obtain g(r). We note that g(r) is noisy due to under sampling and the small bin size (1 nm). Averaging over more time frames or using a larger bin size gave smoother g(r). To compute the OPD pn, we used circular windows with fixed radii 60 nm as in [1]. For each time frame, we sampled centers of 10nN circles according to the uniform distribution in the domain [−0.3μm, 0.3μm] × [−0.3μm, 0.3μm] and calculated the particle occupancy number for each circle. We then produced a histogram of occupancy numbers obtained for all circles and normalized it by the total number of circles to obtain pn.