ImageJ & Mapping Oyster Shells

Co-authored by Madeline Eppley and Lisa Gouralnik (Summer 2024 Intern)

This tutorial provides basic guidance for taking high-quality images, identifying polydora worm blisters, and using ImageJ to calculate total area of blister infection.

Equipment

  • DSLR camera or digital camera. An phone camera may work if it focuses on very close objects.

  • Camera stand or tripod that allows the camera lens to point directly down

  • Level table in a room with consistent, bright, indoor lighting. No natural light is ideal.

  • 3-4 black beanbags, ideally with sand or small particles

  • Color scale swatch card with grey scale

  • An object of known invariable length (we use the length of a square in our color scale swatch card)

  • A computer to run the freely-available ImageJ software

  • A spreadsheet or app to collect data in

Helpful but not required materials

  • A tablet or computer with touch screen and a stylus

  • Computer mouse (instead of laptop trackpad)

  • Large external hard drive to store image backups

Taking high-quality images

The first step in the protocol is to take high-quality images of each shell. We photograph both valves together in one image. We also include a label with the ID number of the individual in the image for easy reference and a color scale card to ensure the color balance is consistent across images.

Scale + Level

It’s crucial that all images of your study organism are taken at the same scale and are level. Any differences in the height of the object to the camera will translate into measurement differences in ImageJ. For this reason, use a level surface and do not vary the height of the tripod and camera. Keep all objects in the frame consistent in their distance to the camera lens across all images.

Tips: If possible, do not move the tripod setup until all imaging is complete. If needed, use a prop (e.g. beanbags) to support each object on a level plane. Do not change the height of the supportive materials - for example, do not vary the number of beanbags stacked under each object.

Photograph Protocol

  1. Attach camera to tripod setup. Focus the camera and take a trial image to test lighting.

  2. Remove the oyster shells and ID tag from the bag. Ensure the ID tag matches the label on the bag. Clear any dust or particles off of the inner shells.

  3. Stabilize the shells on top of black beanbags to ensure they sit parallel to the table and about ~5-6 inches below the camera lens.

  4. Focus the camera and take a few images with both shells, ID tag, and the color swatch in the frame. Make small adjustments in between images if necessary to ensure there is no glare on the shells.

Example image showing both valves, ID tag, and color scale card
Data Collection and ImageJ Protocol
  1. Upload the images from the camera into a folder on the computer that will use the ImageJ software. Label the folder based on the ID_Site_Date population info. Name each of the images a unique name. We use SG_XXXX_PD to stand for seascape genomics, individual ID number, and Polydora.
Collecting Data

We collected our data using an app designed with AppSheet and linked to two Google Sheets - one for whole oyster shell measurements, and one for individual blister measurements. From this point on, if you are not using an app, record your data in whichever spreadsheet(s) you are collecting measurements in, we will refer to “app” as the data collection location.

  1. Upload the full color image of each individual into the app under the Image_valves_full_color field.

  2. Input TRUE or FALSE for Infection_status based on the individual. If there is at least one blister, the infection status is TRUE.

  3. Upload the full color image to ImageJ twice.

  4. Convert one of the images to 8-bit grey scale by selecting Image->Type->8-bit in ImageJ for color comparison.

  5. Set the measurement scale in ImageJ by using the linear selection tool to draw a diagonal line across a square on the color swatch. This is our scale object of known invariable length. Select Analyze -> Set scale and add the length of the diagonal line in the Known Distance text box and the unit in the Unit of Length text box.

The linear selection tool is used to calculate our known distance
  1. Measure the area of each shell. Use the free draw selection tool to select the entire area of each value with the stylus, then press “m” on your keyboard to get the area. Enter the data from the pop up measurement window into the app.

The free draw selection tool is used to map the entire area of each valve

  1. Leave the Blister_count, Total_area_infected, and Total_area_infected_current_blisters_only fields blank (for now), save the data, and open the blister data field in the app.

  2. Fill in the polydora_blistersLabel field with SG_XXXX_PD_n, with n being the count of the blister that you are processing.

  3. Measure the area of the blister. Use the free draw selection tool to draw around the perimeter of the blister then press “m” on your keyboard to get the area. Enter the area into the app.

Overlapping Blisters

Abiotic or biotic environmental stress on oysters may reduce the ability of the oyster to secrete new layers of shell to recover from infestation. It has been hypothesized in the literature that as oysters secrete new layers of shell over blisters, the color of the blister becomes lighter Dorgan et al 2021. Due to this hypothesized dynamic, we have a specific protocol for collecting data from overlapping blisters.

If a white blister and a yellow blister are overlapping, record the area of the entire yellow blister, and crop the area of the white blister to record only the part that does not have overlapping blisters. We collect data this way in order to (1) avoid collecting the same blistered area twice, and (2) accurately estimate the extent of recent/current infections. Assuming that opaque white blisters indicate shell repair over fully healed blisters, yellowed and dark spots are current blisters that are not yet covered with shell repair.

An oyster shell shows overlapping healed white and current infection yellow blisters This image shows the correct way to collect area data from overlapping blisters. The yellow blister area is calculated in its entirety, the white area is cropped, and we avoid collecting the area of overlapping blisters twice.

Multicolored Blisters

In the case of a large, complex blister that has multiple colors, it would take a lot of time to collect data on each small section of it. Instead, view the image in 8-bit grey scale. If they are the same hex code designation in the 8-bit photo, you can count it as one large blister.

An oyster shell shows a multicolored blister This image shows how a complex large blister with multiple colors (yellow, black, brown) is the same shade in grey scale. We count this blister as one occurrence and collect area and hex code data accordingly.

  1. Using the 8-bit photo, record which hex code and number (1-6) is closest to the blister color in Blister_hex_code app field. We determine best hex code match by eye, with the grey scale hex color palette open on the computer screen next to the 8-bit image.

  2. Record Blister_location based on the position in the shell (hinge, center, edge). Record the Blister_valve which specifies if the blister was on the left or right valve in the image.

  3. After completing steps 13-16 for all of the blisters on both shells, go back into the main spreadsheet (related polydoras) to update the Blister_count with the count of total count of blisters processed across both valves.

Data Analysis Protocol

  1. To calculate the Total_area_infected, reference the polydora_blisters tab on the SeascapeSamples spreadsheet. Copy all of the data from the individual just processed and paste into a running excel sheet. Reference the polydora tab on the SeascapeSamples spreadsheet and copy the individuals data into the excel sheet. Highlight all of the blister area values for the individual and record the sum value. Divide the sum value by the total area of the individual (both valves added together). Add this value to the app under Total_area_infected.

  2. To calculate the Total_area_infected_current_blisters_only, highlight all of the blisters except the 1-FFFFFF colored blisters and record the sum value. Divide the sum value by the total area of the individual (both valves added together). Add this value to the app under Total_area_infected_current_blisters_only.

Spreadsheet Format for Data Analysis

In this running excel sheet setup, the NEXT line indicates the start point for data from a new individual. The data from each blister is listed in rows. Below the data, WSUM equals the sum without white blisters (removing any 1-FFFFFF hex code areas). To the right of the data, P1/P2 stands for percentage 1, the percent area of all blisters on the shell, and percentage 2, the percent area of only current blisters on the shell.

Example excel spreadsheet format for calculating total area infected
Editing in ImageJ for Publication

Optionally, you can use features in ImageJ to prepare publication-quality images. We used the freehand selection tool to select each oyster valve, and then removed everything in the background with Edit -> Clear Outside.

Then, we transferred images to powerpoint and used freehand markup tools to create publication-quality composite images like this one!

Composite image and hex code scale