A Smartphone-Based Imaging Method for C. elegans Lawn Avoidance Assay

This article describes a simple, low-cost method of recording the lawn avoidance behavior of Caenorhabditis elegans, using readily available items such as a smartphone and a light emitting diode (LED) light box. We also provide a Python script to process the video file into a format more amenable for counting.

Among the many behaviors displayed by C.elegans, pathogenic lawn avoidance is the worm's defense strategy to minimize its exposure to harmful pathogens. In our lab, we use this behavior as a tool to identify genes involved in the proper functioning of the neurons that mediate this behavior. The lawn avoidance behavior is a simple assay.

However, it develops over many hours. Hence, it is convenient if the assay can be recorded to observe behavior over long time periods. Also, if multiple assays can be recorded at the same time, that's even better.

Although there are many imaging solutions described in the literature, even a few modestly-priced cameras may be out of reach for labs with limited resources. Also, some may want to try the assay, but do not plan on doing it long term. For such cases, short-term solutions, such as this protocol, might help.

Our imaging protocol offers an easy, low-cost way to record worm behavior. We use smartphones to record the assay, then process the video through a Python script that makes it suitable for counting. To set up the imaging apparatus, use a smartphone with a minimum specification of 12-megapixel camera, 1080-pixel resolution, five-gigabyte storage space, and a time-lapse video app from the application store.

Place the LED light box on the bottom rack of an incubator set at 25 degrees Celsius where the assay will take place. To hide the dotted pattern on the LED light surface, spread two sheets of tissues to cover the entire surface of the LED box. Pick an elevated stage for the specimen comprising of a clear plastic sheet supported by hollow rectangular tunnels.

Ensure that the tunnels'walls are somewhat dark to minimize light scatter. Place another rack above the stage to place the phones for recording. Each phone will record three to six plates, so adjust the rack height accordingly.

Put a power strip inside the incubator to plug in the phones during overnight recording. To prepare for the recording, plug in the smartphone to the power strip connected to a power outlet. Ensure to disable the autolock setting to prevent the phone from returning to the lock screen while recording.

Open the time-lapse camera app and set the video quality to 1080 pixels at 30 frames per second. Set the time-lapse interval to two seconds. Place the smartphone with the screen facing up to record with the back camera, and ensure that the paper box tunnels fit within the field of view.

Using a platinum wire pick, transfer 30 synchronized L4-stage worms to the PA14 plate. Place the worms in the middle of the PA14 bacteria lawn. In this study, for each condition, two plates were tested.

Place the two plates on the elevated stage of the recording apparatus with the lid facing down such that the agar side is facing up toward the camera. On the smartphone screen, tap where the plate is so the camera can focus on the assay plates. Start the recording, then add more plates to the stage.

Record for 20 hours from the last set of plates placed on the stage. In the final time-lapse video, 20 hours of recording will result in a 20-minute-long video. Wild-type N2 worms were compared against npr-1 mutants.

The wild-type worms progressively left the bacterial lawn and stayed outside. Worms outside the thick bacterial lawn were clearly seen in the video, but worms inside the bacterial lawn were harder to distinguish. Identify the dimensions of a PA14 plate containing the L4-stage worms by running the Python script extract_frame.py.

Then, open the JPEG file in ImageJ. From the menu, choose Analyze, then Set Measurements. Ensure that the Display label box is checked and close the window.

Using the straight line tool, measure the diameter of a plate by drawing a line across it. Then choose Analyze, then Measure from the menu. If the video is in 1080 pixels, each plate will be about 480-pixels wide.

Enter this information in the Notepad to be used later and close the Results window. Using the multi-point tool, mark points on the upper-left side of each plate, which will become the upper-left corner of the cropped videos. Mark it in order of when the plates were started.

After creating a point for all the plates, choose Analyze, then Measure from the menu. Measurements, including the X and Y coordinates of the points, will appear in the Results window. To process multiple videos, repeat the process in ImageJ with other JPEG files.

All X and Y coordinates will be listed in the same Results window. Save the Results window into a CSV file. The file should be saved to the same directory as the movie files.

To find the starting time for each plate, play the movie either on the computer or phone and take note of the starting times of each set of plates placed under the camera. Open the Results. csv file with the coordinates and add a start column.

For each row corresponding to individual plates, enter appropriate start time in seconds under the start column. Save the file. To crop and trim the videos, run the crop_n_trim.

py script. Choose the Results. csv file.

Enter the plate dimensions and the pixel value noted earlier. The script will now read each row of the Results. csv file to find the correct movie file after the script finishes running.

A folder will appear with the same name as the movie followed by the start time in which videos corresponding to the assay will be saved. For manual counting, open each AVI file in ImageJ. Count the worms that are visible outside the lawn and then calculate the occupancy rate for each time point.

The wild-type N2 worms progressively left the bacterial lawn and stayed outside as noted in the occupancy rate over time. The counts made directly from the plates were compared against counts from imaged worms. The counts made from imaged worms turned out to be highly accurate.

When three trials for each strain were averaged together, the N2 and npr-1 strains yielded 99.5%and 96.2%accuracy, respectively.