Robustness of ExTrack to various sources of bias. Heatmaps of relative errors on d₁, k_u, and k_b in case of two-state parameter fits to two-state simulations with one immobile state and one diffusive state, for different sources of bias, as a function of the source of bias (y axis) and transition rates k. (a and b) We simulated track-to-track variations of localization error σ (a) or diffusion coefficient D₁ (b). Varied parameters followed chi-square distributions (white graphs in b) re-scaled so the mean localization error equals 0.02 µm (a) or the mean diffusion coefficient equals 0.25 µm².s⁻¹, which corresponds to a diffusion length of $d1*$ = 5 (b). (c) Membrane proteins diffuse on a cylindrical surface and leave the field of view on the sides (see cartoon). We varied the width w of the field of view as indicated in the cartoon, while maintaining d₁ and σ fixed. (d) Cytoplasmic proteins can leave the focal plane anywhere (see cartoon). We varied the focal depth while maintaining d₁ and σ fixed. (e) Particles confined in a symmetric cube. We varied the box size while maintaining d and σ fixed. 10 replicates per condition. If not stated otherwise, d₀ = 0 µm, d₁ = 0.1 µm, and σ = 0.02 µm. ExTrack settings: window length = 7, no sub-steps. CV, coefficient of variation.