Skip to main content


Fig. 6 | BMC Systems Biology

Fig. 6

From: Inference of spatiotemporal effects on cellular state transitions from time-lapse microscopy

Fig. 6

The method’s performance is robust for moderate amount of tracking error. a Statistical power plotted against the amount of tracking error for the density dependent scenario from Fig. 4 a-c (4500 onsets). Solid (dashed) lines correspond to relevant (irrelevant) features in the respective scenario. The correct features are identified reliably (power ≥0.8) up to a tracking error of 3 %. For larger tracking error, there is a high probability to include time (blue curve) into the model even though it is only an indirect influence. Note that tracking error seems to some extent facilitate the detection of ϕ 0,ϕ 1 (see main text for details). b Analogous to a, but for the dataset where the transition rate λ depends on time and local cell density with a Gaussian kernel (4500 onsets)

Back to article page