This is a repost from June 13, 2006, that brings back some experimental issues I was having then…
The journal for the International Society for Analytical Cytology has an interesting methods paper out titled Functional proteomics for cell migration. It’s opening premise is:
Determining the activity-function relationships of signaling proteins to cellular phenotypic changes in living cells over hours to days can be daunting and is often inaccurate. Conventional cell biology relies primarily on manual imaging of one or a few cells, or on whole-population molecular methods. We report progress on a set of high-throughput microscopy (HTM) computer vision tools that rapidly acquire, process, and analyze biosensor ([1-3]) and HTM () images from many cells over multiple fields-of-view for tracking cell migration and division at high-resolution and quantifying the spatiotemporal activation of signaling proteins in living cells.
Unenthusiastic about the prospect of interactively acquiring and subjectively interpreting many time-lapse images to correlate Rho GTPase [or any signaling molecule’s] activity with cell migration, we explored techniques for automating scanning, processing, measurement, and analyses. To correlate cell behavior to protein activation, we developed software to automatically trace cell tracks and identify mitoses for measuring cell velocities and interdivision times. Automated measurements of spatiotemporal protein activity patterns from many cells in a population may accelerate causative pathway discoveries and improve the confidence and breadth of our understanding of the underlying mechanisms.
Indeed, this is addresses a daunting problem (and flaw) in my own research, as I’m currently using conventional cell biology to image and manually track whole-populations of cells – and it IS a daunting, slow and imprecise methodology that I’m currently using. Check below the fold for info and comments on the methods discussed in Shen et al.
Specifically, the paper’s goals were to explore programming methods of resolving and aligning different flourescent channels for quantifying and correlating protein activity to cellular functions; and to refine motion estimation of optical flow for tracking a dense velocity field and feature-based tracking of a relatively small number of features or objects in long time-sequences. Included in the study were the use of a microstepper motorized stage-equipped microscope, strobe- and shuttered-illumination for bright-field and flourescence imaging (respectively), a CCD camera with autofocus capabilities, and a stage-encompassing cell culture incubator. (side note: for myself, I really really need a microstepper motorized stage)
Overall, Shen et al. nicely demonstrate a packet of software tools for data extraction and mining in an automated method that can aptly be described as high-throughput microscopy (HTM). These techniques – including both HTM component utilization and software implementation aspects – are vital for further work in cell migration dynamics, and the Cell Migration Consortium would do well to encourage further workshops into the use of these methods.
“The time-lapse HTM and data processing tools will result in rapid screening and scoring of thousands of cells per experiment with statistical significance not possible from more laborious interactive observations.”
- Shen F, Hodgson L, Rabinovich A, Pertz O, Hahn K, Price JH. Functional proteometrics for cell migration. Cytometry A. 2006 Jul;69(7):563-72. DOI: 10.1002/cyto.a.20283
- Software tracks proteins inside living cells – New Scientist Tech.