Quantitative Image Analysis of Cell Behavior and Molecular Dynamics During Tissue Morphogenesis
The cell behaviors that drive tissue morphogenesis, such as division, migration, or death, are regulated at the molecular scale. Understanding how molecular events determine cell behavior requires simultaneous tracking and measurement of molecular and cellular dynamics. To this end, we have developed SIESTA, an integrated tool for Scientific ImagE SegmenTation and Analysis that enables quantification of cell behavior and molecular events from image data. Here we use SIESTA to show how to automatically delineate cells in images (segmentation) using the watershed algorithm, a region-growing method for boundary detection. For images in which automated segmentation is not possible due to low or inappropriate contrast, we use a minimal path search algorithm to semiautomatically delineate the cells. We use the segmentation results to quantify cellular morphology and molecular dynamics in different subcellular compartments, and demonstrate the whole process by analyzing cell behavior and the dynamics of the motor protein non-muscle myosin II during axis elongation in a Drosophila embryo. Finally, we show how image analysis can be used to quantify molecular asymmetries that orient cell behavior, and demonstrate this point by measuring planar cell polarity in Drosophila embryos. We describe all methods in detail to allow their implementation and application using other software packages. The use of (semi) automated quantitative imaging enables the analysis of a large number of samples, thus providing the statistical power necessary to detect subtle molecular differences that may result in differences in cell behavior.
Key wordsImage analysis Segmentation Watershed Minimal path search Cell morphology Molecular dynamics Planar cell polarity Cytoskeleton
We are especially grateful to Jennifer Zallen for her constant support and advice. SIESTA was initially developed by R.E.G. in the Zallen lab. Our work is supported by a Connaught Fund New Investigator Award to R.E.G., and grants from the University of Toronto Faculty of Medicine Dean’s New Staff Fund, the Canada Foundation for Innovation [#30279], the Ontario Research Fund and the Natural Sciences and Engineering Research Council of Canada Discovery Grant program [#418438-13 to R.E.G.].
Running and using SIESTA: This tutorial demonstrates, step by step using SIESTA, each of the methods discussed in Subheading 3 (MOV 166779 kb).
Dijkstra’s algorithm for minimal path search: To find the brightest path between two pixels in an image (indicated by green crosses), the image is inverted (such that bright pixels have low pixel values, and vice versa). A directed, weighted graph is built. In this graph nodes represent pixels, edges connect adjacent pixels, and each edge is weighted by the gray value of the destination pixel in the inverted image. Dijkstra’s algorithm for minimal path search (Box 1) is used to find the optimal path between the nodes that represent the initial and final pixels. The final path is transferred to the original image (MOV 1687 kb).
- 7.Rasband WS (1997) ImageJ. U.S. National Institutes of Health, Bethesda, MDGoogle Scholar
- 11.Mortensen E, Morse B, Barrett W et al. (1992) Adaptive boundary detection using ‘Live-Wire’ two-dimensional dynamic programming. Paper presented at the Computers in Cardiology, Durham, NC, October 11–14, 1992.Google Scholar