Jeffrey Segall's Lab

Eliminating x-y drift in time lapse movies

It can be useful to mark sites of interest in time lapse movies; for example we will mark where tumor or cells are moving as part of quantitating tumor cell motility.  We have written a simple series of macros/plugins that make use of another valuable plugin written by Philippe Thévenaz, called Pointpicker (http://bigwww.epfl.ch/thevenaz/pointpicker/).  Pointpicker keeps a record of the x-y coordinates of the mouse when clicked and allows you to save or copy them.  We typically are just counting instances of tumor cell movement and thus wish to mark their locations and kept track of the total number.  Usually this is most accurately done viewing just the tumor cell channel in black and white, however, pointpicker will either work on color or black and white movies. 

Since we have GFP labeled cells, we have a simple plugin: Greentopointpicker, that selects the green channel and begins the Pointpicker plugin. 

Items to put in Plugins folder:

PointPicker (from Thevenau web site): folder with PointPicker plugin and files in it

Procedure:

  1. Start ImageJ.
  2. Open time lapse movie of slice to analyze (preferably already registered to remove drift, jitter, etc). 
  3. Option: Select the color you would like the crosses to be by going to Edit-options-colors and change the selection color to the color you like – i.e., red
  4. Option: Run the plugin Greentopointpicker: this extracts the green image (typically the cell image that we want to track) and then starts Pointpicker. 
  5. Option: convert to crosses to just 1 color (double click on the pen+ (leftmost) button and choose monochrome). 
  6. Click on the Apply to all slices button (6th from the left, has a sliced cube image on it) so that the crosses will be on all slices.
  7. Option: It can be useful to magnify the image 1x (click on magnifying glass, click on image, then click on pen+ (leftmost) button to go back to marking mode).
  8. Option: It can be helpful to play the movie  at high speeds and looping back and forth continuously to see motions (Stacks-animation options).
  9. Option: Expand the image to be tracked to the full screen (I like this for seeing small movements)
  10. Start the movie running in a continuous loop.
  11. Mark all the sites of interest (i.e., tumor cell motility) by clicking on the pen + (leftmost) button and clicking with the mouse of the specific sites.  A cross should appear at each marked site. 
  12. Option: Errors can be removed by clicking on the pen – button (third from left) and clicking the mouse anywhere on the screen.  However, this simply removes the marks in reverse order to how they were made – specific marks cannot be removed out of this sequence.
  13. Stop the movie, shrink the image back down to view the ImageJ tool bar.
  14. Reveal the points list by clicking on  the document button (8th from left) and clicking on the show button from the drop down menu.
  15. The results box will appear with point, x, y, slice, and color columns.   The information in each slice should be identical.  Highlight the rows with the appropriate slice numer corresponding to the slice currently being analyzed and paste them into an excel file.  For different analyses, the file can have different worksheets (ie., tumor cell motility, host cell motility, movement on matrix, orientation towards blood vessel, etc).  This stores then the number and coordinates of the locations of cell movement together with the slice number.  I then close the results box.  If you will be using jsDrawDot1 to mark the sites on a z stack, copy and save the results with the number in the slice column being the slice number+1 (this provides the correct registration of the dots with the z series using jsdrawDot2).
  16. It is usually good to save the excel file at this point.  Using a file name indicating the date of the experiment and other distinguishing information is helpful for keeping the data organized.
  17. Option: To further analyze the same movie for a new category, remove the current markings by selecting the pen- button (3rd from the left) and click on the image until all the crosses are gone (and the pen+ button will automatically depress).
  18. Repeat steps 10 – 16 for as many categories as desired.
  19. Close the movie image. 
  20. Click the return button to return to the standard imageJ menu, open the next file to be analyzed, and repeat the process.
  21. Option:  save the excel data file with the file name of the experiment in the same directory as the registered files. 

Marking specific locations with red circles in a z-stack using the data acquired from Pointpicker

Plugins used (put both just in the Plugins folder of ImageJ):

  • JsdrawDot1 : from this web site
  • IntLoader2 : from this web site

To superimpose the dots on a z series of the data, copy the columns (not the header column) and paste as unformatted text into a word file, and save the word file as plain text.  This makes the data readable for IntLoader2 (called by jsdrawDot2).

  1. Open the zseries upon which you wish to superimpose the sites
  2. Run jsdrawDot2.  It will remind you what is required and after you click ‘OK’ it will ask for the location of the plain text file that has the data acquired from PointPicker.
  3. It will note ‘All done’ when finished.

FYI: The Manual Tracking plugin by Fabrice Cordelieres can be useful for following individual cells in 2D and 3D.  Link from ImageJ site -  http://rsb.info.nih.gov/ij/plugins/track/track.html