Biofilm_surface area-Roughness

Hi Hannah

Attached are my results from biofilm outer surface to volume. The biofilms on HA discs represent biofilms on 2D substrates, whereas biofilms on MEW 100 and MEW 250 represent biofilms on 3D substrates.

Also with regards to biofilm roughness some areas show “Nan”, not sure what it represents.

I do not see any major difference in the biofilms formed on 2D versus the 3D substrates.

Is it a correct interpretation of the results??

How to avoid this “Nan” in the biofilm roughness column??


Hi Srinivas,

am I assuming correctly that the three columns in each condition correspond to three different replicates? Without knowledge of your experiments and the detailed steps of analysis I cannot really judge if your interpretation is correct. But from a purely statistical point of view, plotting the mean and standard deviation for each condition should help you determine whether or not a significant difference exists.
Regarding the image analysis, if the segmentation looks good and the cube size is reasonable, the quantification should also be fine.

To be honest I am also a bit puzzled by the NaN values in the biofilm roughness. These values occur when a calculation does not return a reasonable result (e.g. dividing zero by zero). To get to the bottom of this, could you please do the following and tell me the result:

  • Check the segmentation: Pick an image where the NaN value occurs and double-check its segmentation by selecting the image, clicking on “Load segmentation results”, selecting some meaningful property (e.g. Surface_LocalRoughness) from the table that opens (might take a few seconds to open) and click on “Ortho view” (see screenshot below). The view that will open shows a representation of your biofilm in which you should be able to also see the cubes. Alternatively, you can use the “Overlay” option in the image preview area to view the segmentation. Now, please confirm that this segmentation looks fine, and that there are cubes visible of a reasonable size (see for example bottom screenshot).

  • Use the visualization described above (load segmentation results, then select property and click on ortho view) to check if the properties “Surface_LocalRoughness_rangeX”, “Surface_LocalThickness” and “Surface_PerSubstrateArea” look reasonable (you should again be able to see the outline of your biofilm and cubes, if this is not the case there might be something wrong). By clicking on the “Data Cursor” at the top of the window and then clicking on any of the cubes of your biofilm, you will be able to directly see the value of the property for that cube (see screenshot below, the value is given by “Index”).

  • Finally, can you check the property “Biofilm_MeanThickness” and let me know if this one is also NaN?

I am looking forward to your reply - thank you for making us aware of this potential bug.