Skip to main content
Fig. 4 | BMC Bioinformatics

Fig. 4

From: Automatic detection of diffusion modes within biological membranes using back-propagation neural network

Fig. 4

Analysis of the motion modes within synthetic trajectories. a Detection probability using BPNN. 200 trajectories of 400 frames including one directed motion segment with 1.2 μm/s velocity and one confinement segment with 1 μm diameter were analyzed with BPPN (confined in light grey, directed in dark grey; D = 0.25 μm2/s, integration time = 100 ms; each 50 frames segment is always localized at the same position). The percentage of decision based on BPNN corresponds to the number of positive decision for a specific motion mode detected for a given frame over 200 trajectories and normalized to 1 or -1 for confined or directed trajectories, respectively (black, confined trajectories; grey, directed trajectories). The algorithm was also tested with a 30 nm localization noise (dotted lines in the graph). b The upper panel shows a synthetic trajectory of 40 s (400 frames) including a transient confinement (from 10 to 15 s) with a diameter of 1 μm (red trace, zoomed in the red circle) and a transient directed motion (30 to 35 s) with a velocity V = 1.2 μm/s (blue trace). The Brownian part is in black. The diffusion coefficient D is 0.25 μm2/s and the integration time 100 ms. Scale bar, 1 μm. c The Lower panel is the plot of the probability of detection of motion mode (BPNN output values) as a function of the duration time of the trajectory for confined (red) or directed (blue) motion. Confined (bold red trace) and directed (bold blue trace) segments are respectively detected between 10.0 and 14.6 s and between 30.6 and 35.3 s as shown by grey lines. The detection threshold values correspond to a probability of detection with a 95 % confidence (see Fig. 4). The calculated diameter of confinement is 1.06 μm

Back to article page