Lab 1, Part B: Analysis of Cell Motion Using ImageJ.


This is the second week of a two-week lab studying cell motion. Last week we analyzed the 1-D motion of an amoeba. This week we will be using ImageJ to analyze videos of cell motion. Although you are using canned data, this is a trial run for a whole video analysis of a biological phenomenon.

The Scenario: A patient has a wound, in the process of healing, that is infected with bacteria. By looking at the relative speeds of (1) wound healing, (2) neutrophil motion, and (3) bacteria motion, you will be able to give one answer to the question “Why is an immune response necessary, but not always sufficient, to prevent bacterial infection of a wound?”  This can be more succinctly phrased as “Why are we really, really glad that we have antibiotics?”

Materials and Process

Your lab group has been provided with three video files, of wound healing, neutrophil motion, and bacteria motion. Each video is a sequence of images called frames. Taken together, each video is an image sequence or (in ImageJ terminology) a stack. The wound healing videos shows breast tissue cell sheet migration. The Neutrophils video shows white blood cells responding to a high concentration of fMLP — the chemical indicator of bacteria. The bacteria video shows E. coli motion.

  • Wound healing: Wound healing
    • Scale: 0.65 μm/pixel
    • Recording rate: 6.0 min/frame
  • White blood cells: Neutrophils
    • Scale: 1.326 μm/pixel
    • Recording rate: 7.2 sec/frame
  • Bacteria: E. coli
    • Scale: 31.5 pixels/μm
    • Recording rate: 0.050 sec/frame

Your task is to perform a quantitative analysis, with ImageJ and Excel, of the rates of motion of these cells. This quantitative analysis should help you problem-solve within this scenario. Today you will practice and master the skills necessary to analyze motion using Fiji / ImageJ. After today, you will ALL be expected to be experts at these skills, so take turns and help each other learn. Take notes for the future if you are worried that you will forget.

  1. Start with the neutrophil file. Use the “Multi-Point” tool to track a single cell for at least 30 frames. “Measure” your clicks, export to Excel (using cut/paste), convert to real units (um and s) and calculate the “instantaneous” velocity and speed of the cell
  2. Later on we will ask you to compare the average speed of the different types of cells. Tracking cells frame by frame is probably overkill if you are just going to end up averaging all the data. How many frames can you reasonably skip between tracking clicks? How did you decide this? You will need to make a similar judgment when analyzing the other videos, so make sure your answer can be applied to any type of video.
  3. Track 10 neutrophils. and calculate the average speed of each cell.
  4. Grab the E. coli file and track 10 cells. Calculate the average speed of each cell.
  5. Grab the wound healing file and track 10 cells. Calculate the average speed of each cell.
Lab writeup

Use this blank Word document as a template for your writeup. It’s really just a glorified cover page with some hints about what you might include.

By the next lab, your group will submit one collective lab report. This will be graded by the TA according to our lab rubric. Good attention to detail now will save you time later! Remember, your TA is here to help you with equipment and ImageJ, but the physics is up to you and your group!

Questions to be answered
  1. What criterion did you use to decide how many frames you could skip while tracking the cells? Explain why this is justified.
  2. Based on the average (or typical) speeds you measured, why is an immune response necessary, but not always sufficient, to prevent bacterial infection of a wound?
  3. The average speed isn’t always the most relevant physical quantity. How does your answer to question (1) change when you consider the fastest (or slowest) speeds you measured? To see why it’s the slowest / fastest speed that matters, think of the joke about the two hikers trying to outrun the bear.
  4. Think back on last week’s question: are all these types of cells equally “fast” or “slow” according to your criterion from last week?
  1. If all goes well you can just “File > Open” the video files in ImageJ. If not (and in the future) you may need to pull out the big guns to get video into ImageJ: use “File > Import > Movie (FFMPEG)”. If your ImageJ doesn’t have the FFMPEG importer you will need to add it; see the instructions here.
  2. Don’t trust the frame rate information that may (or may not) be embedded in the file: it is easily corrupted when changing video formats. Use the numbers at the top of this page.
  3. Choose cells that are easily tracked – that don’t run into other similar-looking cells, get obscured by dirt or obstacles, etc.
  4. Summarize your cell speed data in a table in the writeup. We don’t want to see all the cells’ speeds, but you could (for each type of cell) present the average, maximum and maybe the minimum and/or standard deviation. This isn’t a big improvement in data reduction since you only had 10 data points to begin with (so winnowing down to 3 or 4 doesn’t help enormously) but this approach scales well with increasing data size: if we had time you’d track a couple of hundred cells to get better quality data and 200 cells condensed to 3 or 4 statistical measures is much more useful.
  5. Once you figure out what procedure you want to use to track cells, you may want to farm out the tracking of the three types of cells to three members of your lab group. Use parallel processing.

Originally developed by: K. Moore, J. Giannini, B. Geller & W. Losert (Univ. of Maryland, College Park)