Deciphering Sub-Cellular Processes in Live Imaging Data Sets via Dynamic Probabilistic Networks

Department of Neurobiology, Yale University School of Medicine, New Haven, CT 06510, USA.
Bioinformatics (Impact Factor: 4.98). 08/2010; 26(16):2029-36. DOI: 10.1093/bioinformatics/btq331
Source: PubMed


MOTIVATION: Designing mathematical tools that can formally describe the dynamics of complex intracellular processes remains a challenge. Live cell imaging reveals changes in the cellular states, but current simple approaches extract only minimal information of a static snapshot. RESULTS: We implemented a novel approach for analyzing organelle behavior in live cell imaging data based on hidden Markov models (HMMs) and showed that it can determine the number and evolution of distinct cellular states involved in a biological process. We analyzed insulin-mediated exocytosis of single Glut4-vesicles, a process critical for blood glucose homeostasis and impaired in type II diabetes, by using total internal reflection fluorescence microscopy (TIRFM). HMM analyses of movie sequences of living cells reveal that insulin controls spatial and temporal dynamics of exocytosis via the exocyst, a putative tethering protein complex. Our studies have validated the proof-of-principle of HMM for cellular imaging and provided direct evidence for the existence of complex spatial-temporal regulation of exocytosis in non-polarized cells. We independently confirmed insulin-dependent spatial regulation by using static spatial statistics methods. Conclusion: We propose that HMM-based approach can be exploited in a wide avenue of cellular processes, especially those where the changes of cellular states in space and time may be highly complex and non-obvious, such as in cell polarization, signaling and developmental processes.

Download full-text


Available from: Rafael Sebastian,
26 Reads
  • [Show abstract] [Hide abstract]
    ABSTRACT: We have applied an artificial neural network (ANN) using the backpropagation learning algorithm to the biomagnetic inverse problem. A forward model was used to calculate the magnetic fields from propagating action potentials (APs) as would be seen in nerve or muscle bundles. This forward model depicted two design schemes of a high-resolution SQUID magnetometer, whose three pickup coils were in two configurations, a stacked array and a planar array with respect to the direction of propagation in the idealized bundle. The ANN was tasked with determining the location of the source in two dimensions, with the third being the direction of propagation which was assumed to be known. The ability of the network to determine the source location was dependent on the arrangement of the magnetometer pickup coils
    Engineering in Medicine and Biology Society, 1995., IEEE 17th Annual Conference; 01/1995
  • Chinese Medical Sciences Journal 03/2011; 26(1):60-3. DOI:10.1016/S1001-9294(11)60021-7
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Accurately detecting subcellular particles in fluorescence microscopy is of primary interest for further quantitative analysis such as counting, tracking, or classification. Our primary goal is to segment vesicles likely to share nearly the same size in fluorescence microscopy images. Our method termed adaptive thresholding of Laplacian of Gaussian (LoG) images with autoselected scale (ATLAS) automatically selects the optimal scale corresponding to the most frequent spot size in the image. Four criteria are proposed and compared to determine the optimal scale in a scale-space framework. Then, the segmentation stage amounts to thresholding the LoG of the intensity image. In contrast to other methods, the threshold is locally adapted given a probability of false alarm (PFA) specified by the user for the whole set of images to be processed. The local threshold is automatically derived from the PFA value and local image statistics estimated in a window whose size is not a critical parameter. We also propose a new data set for benchmarking, consisting of six collections of one hundred images each, which exploits backgrounds extracted from real microscopy images. We have carried out an extensive comparative evaluation on several data sets with ground-truth, which demonstrates that ATLAS outperforms existing methods. ATLAS does not need any fine parameter tuning and requires very low computation time. Convincing results are also reported on real total internal reflection fluorescence microscopy images.
    IEEE Transactions on Image Processing 07/2015; DOI:10.1109/TIP.2015.2450996 · 3.63 Impact Factor