Using mTrackJ to quantify cell motility behavior

  1. Open a stack of images in ImageJ or Fiji. If you are doing this on your own computer, you may not have mTrackJ installed. You can download it from the link:
  1. The system should automatically know the calibration if you pick the correct objective in the presets before you start imaging. If you did not, I think you can change the objective and resave the image and it will correct it. It does not know the time interval. To input this, go to Image/Properties and enter the time interval and check the size calibration.
  2. In the Plug-ins folder, open mTrackJ (the full manual is bookmarked in the menu of Safari).
  3. Click Add Tracks. Pick a cell to follow and click on it. The program puts a dot where you clicked and advances to the next frame. Click again and repeat until you reach the end or decide to stop. When you are done with that cell, double click or hit Escape to end that track. You can then take the movie back to the beginning and start clicking on a new cell.
  4. When done, click on Measure Tracks and check the first two boxes (Display point measurements, Display Track Measurements). You will get two new windows open that show the analysis of the data. The track measurements is for each entire track. The point measurements is the timepoint by timepoint analysis of each track.
  5. This data can be saved and imported into Excel or copies and pasted into Excel.
  6. The columns of data (described in manual) are:

•ID: The ID number of the track to which the point belongs.

•PID: The ID number of the point.

•x [unit]: The calibrated x coordinate of the point. The pixel width and unit of length used here can be set as described above.

•y [unit]: The calibrated y coordinate of the point. The pixel height and unit of length used here can be set as described above.

•z [unit]: The calibrated z coordinate of the point. The voxel depth and unit of length used here can be set as described above. This quantity is not displayed in the case of 2D image sequences.

•t [unit]: The calibrated t coordinate of the point. The frame interval and unit of time used here can be set as described above.

•Len [unit]: The length of the track from the start (first) point of the track to the current point (inclusive). Thus the value of this quantity for the end (last) point of a track equals the total length of the track. The pixel calibration and unit of length used here can be set as described above.

•D2S [unit]: The distance from the start (first) point of the track to the current point. Thus the value of this quantity is always less than (or at most equal to) the value of Len. The pixel calibration and unit of length used here can be set as described above.

•D2P [unit]: The distance from the current to the previous point of the track. The pixel size and unit of length can be set as described above.

•v [unit]: The velocity, at the current point, of the object represented by the track. This quantity is computed as the distance from the previous to the current point, divided by the frame interval. The pixel calibration, unit of length, frame interval, and unit of time can be set as described above.

  1. You can now plot the speed of each cell over time. You have to look at the track ID to find where one cell track ends and the next begins.
  2. One measure of the movement is the distance a cell moves from frame to frame. Another is the total distance moved over the time you tracked the cell (average speed). Note that you are measuring speed, not velocity. Velocity is a vector measurement of speed in a particular direction.

Things to think about as you experiment with how to analyze your data:

  1. Are you going to click on the center of a cell, the front, the back?
  2. Which cells are you going to analyze? Does it make sense to pick cells at random to analyze or will you use some criteria for choosing cells? Be sure to describe how you made the choices in your Discussion.
  3. The interval you choose will affect your data. You do not have to analyze every frame you captured. You might find that not much movement happened from frame to frame, so you would waste alot of time and effort clicking many frames with little movement. If you do this, then the accuracy of your mouse clicks will determine what you measure as speed. We call this “instantaneous velocity” as opposed to “translocational velocity” and it is often used to measure shape change in a cell rather than movement. As the cell changes shape, the mathematical center (centroid) moves around and that is what you are measuring. If you pick time points too far apart, the cell may have changed direction between time points and you may miss some of its movement. You can try analyzing the data with different time intervals and see what happens. mTrackJ is smart enough to know if you skip frames to correct the time accordingly (check to make sure it is correct- do not assume).
  4. The program gives you the net distance a cell moves (D2S: first point to last point) and the total length of the track (Len: the total distance moved along the track). The ratio of these two (D2S/Len) is a measure of the “persistence”. If a cell moves in a straight line, the two numbers will be the same (Persistence=1). If a cell changes direction frequently, then Len will be much greater than D2S. Thus there are two issues: (1) is the cell moving in random directions or all in the same direction and (2), how persistent is the movement in whatever direction it chooses to move?
  5. To answer the question of direction, a nice way to show this is to use a Rose plot. The program gave you the X-Y position of the cell at each time point. So you could plot the positions as a scatter plot to show the track of movement. To do a Rose plot, in Excel you subtract the first X and Y values from the rest of the values for that cell. That means the cell starts at position 0,0. If you do that for each cell and then plot the cells on the same graph, they will all have the same starting point and either all go in the same direction, or go off randomly in different directions from that starting point.
  6. In order to view how the cells moved, you can plot speed vs. time. That tells you whether the cells move in a continuous fashion or they start and stop frequently. Sometimes the data can be pretty noisy. You can smooth the data in several ways.
  7. One way is to average groups of points. If you choose a cell next to the speed column, type = then choose Mean, then select the first 5 cells, Excel will return the average of the first 5 cells. If you now select that cell and drag to the bottom of the column, then choose Edit/Fill/Down, Excel will do the calculation for each point.
  8. A better way to do this is with a weighted average where you count the middle cell the most, and the adjacent cells less in the average. To do this, select the middle of the first 5 cells. In the next cell over set up a calculation where you take the first cell + 2x the next cell + 4x the middle cell + 2x the next cell + 1x the 5th cell and divide the sum by 10. You can then select that cell and the rest of the column and fill down to complete for all time points.
  9. The Configure Tracking and Configure Displaying have some useful options you can play with for analyzing and displaying the data. Make Movie will use those options to create a movie of the tracking you did (see the movie on the top of the main web site page).

The manual explains other options