Biofilms cause complications and high costs in both industry and medicine. Of particular interest are bacterial infections of prosthetic materials, which usually cannot be eliminated due to the high antibiotic resistance known for bacteria forming biofilms. The search for new materials and coatings with lower colonization potential and antibacterial activity is of great importance to reduce biofilm formation. However, there is no standardized procedure to examine the colonization characteristics of bacteria in the biofilm state in situ. Biofilms can be analyzed with automated epifluorescence microscopy systems for the semi-quantitative analysis of three-dimensional (3D) biofilms on various surfaces. Based on our previous publication we used the VideoScan platfrom, which is a fully automatized multispectral imaging plattform that can be used to analyze microbead assays, cell assays or a combination thereof. However, most parts of this platform are driven by a closed source software stack.

To analyze adherent bacteria, three materials (glass, steel and titanium) were incubated with bacteria in a flow chamber system. After fluorescence staining of the bacteria, automated image capturing, quantification of the bacteria, measurement of the colonized area and determination of the 3D biofilm height were carried out by using novel software. Furthermore, the materials were examined for their surface topography using white light scanning interferometry.

One two-dimensional image is analyzed by calculation of the number of bacteria by differences in the gray intensity of pixels of this image. A pixel with higher gray intensity than the background belonged to a bacterium. The higher the gray intensity of one pixel, the more bacteria lay on top of each other. In more detail, to distinguish the bacteria from the background, an intensity threshold (determined experimentally) was set: pixels with a higher fluorescence intensity than the threshold were defined as belonging to a bacterium (a bacterium ~20 pixel). To count bacteria in biofilms, we assume that the bacteria are uniformly stained and brighter areas are caused by the superposition of several bacteria within this area. The bacterium with the lowest fluorescence intensity that is separated from the background is single-layered. Brighter bacteria are multi-layered. This divides the foreground area into brightness classes and represents a linear relation. The first class comprise the areas with intensity values from the threshold up to the doubled threshold. These areas are covered with single-layered biofilms. Intensity values from the double threshold up to the triple threshold are covered with double-layered biofilms. These areas are then divided by the size of a single bacteria and multiplied by the number of superimposed layers. The pixel brightness was used to differentiate between a bacterium and the background, thus determining the colonized area in the grayscale images.

As supporters of reproducible research and open source software (DOI:10.21037/jlpm.2019.04.05), we have reimplemented the algortihm to determine the area, number of bacteria and height of a biofilm by using the presumably most widely used bioimage software ImageJ and its batteries-included distribution FIJI. The approach described above was implemented as an open source plugin, called OBA-Open-Biofilm-Altimeter, for ImageJ/FIJI.

OBA (Open Biofilm Altimeter) is an open source software tool to determine the area, number of bacteria and height of a biofilm with a thickness of up to 20 µm. To determine the thickness of a biofilm, height profiles are created for the gray scale images captured (and could be converted into 3D representations using the ImageJ 1.5 plugin Interactive 3D Surface Plot).

OBA was developed to work with the VideoScan system.

OBO is distributed under the GNU General Public License v3.0.

Core functions of OBA are implemented in the Java library Bacteria_counter.java.

This software was tested under Linux (Ubuntu 19.10) and Windows (8, 10). All systms were 64 bit (32 bit not tested). We assume that it should work under all platfoms that are supported by ImageJ (>= 1.5) /FIJI and Java (>= 1.8.0-12).

1. Create a new folder named in ImageJ/Plugins

2. Open ImageJ or FIJI

3. Click on Plugins/Install

4. Go to the created Folder on choose BacteriaCounter.java. ImageJ will then open a dialog to ask where the plugin should be installed.

5. Go to the created folder and click Yes. ImageJ/FIJI will warn you that Bacteria_Counter.java already exists. Click Yes to overwrite

6. Open an image e.g. one of the examples provided in this repository

7. Then go to Plugins/Name of the install folder/Bacteria counter

Upon start, the plugin will open a dialog for analysis. This dialog contains following parameters:

• Bacteria Height: The height of one bacterium in units
• Pixels per bacterium: The area of one bacterium in pixels
• Stretch Contrast: Stretches the values of the image to a range between 0 and 255. This allows visual control of the detected areas, but might falsify the results. Should not be used if the results should be used further. Nevertheless, useful during preview.
• Preview: Enables the preview mode of ImageJ for parameter tuning

After analysis, following results will be logged by the plugin:

• Approximate Bacteria Count: The approximate number of the bacteria in the biofilm
• Max. Height: The maximal amount of bacteria overlapping each other.
• Theoretical Area: The area in pixels the bacteria would need if layed out in a one dimensional layer