Biovolume interpretation

I calculated global biofilm properties for my time series and multi position data.
I was particularly interested in biofilm_volume parameter because it translates to overall physical volume the biofilm is covering. However, I am having a hard time in reconciling the output data with the actual images. For example, for 2 image stacks I got biovolumes as 297 and 5000 respectively. But visually, I am not sure if the difference is 20 fold.
I used most of the default settings till global properties calculation. Please let me know if something is wrong with my analysis or interpretation.
attaching the relevant imagespos8_0uM_10hrs.tif (3.7 MB) pos15_500uM_10hrs.tif (3.6 MB)

PS: reassuringly, a few stacks were out of focus for which there is very low biovolume as expected.

Hi Omkar,

yes, you are right: The images do not indicate a 20x biovolume change.

Have you checked the foreground/ background segmentation via the button: “Overlay”?
image

I suspect that the (default) 2 class Otsu threshold does not work in your case and that a large number of voxels (3D pixel) are misclassified. If this is the case there are three possible options to proceed:

  1. Try to use the 3 class Otsu threshold.
  2. Experiment how the other (provided) automatic threshold approaches.
  3. Since you only have two images: Use the manual thresholding (Not recommended since it can introduce biases into the analysis)

I hope that this brings the analysis back to a more reasonable output.

Best wishes,

Eric

1 Like

Thanks, Eric!
I checked ortho view after segmentation. While the default sensitivity was quite low for my purpose, it shouldn’t have biased the samples in different directions (2 visually similar samples but very different biovolumes). I will be repeating the analysis with appropriate sensitivity and method.
But I have 2 followed questions

  1. Merging: I only have one fluorescent channel and the other one is transmitted light channel so I don’t wish to merge these. But I saw that default option is to merge channel 2 with channel 1. If I don’t click on “merge” tab and directly click on “segment cells”, am I actually merging anything? Is there an option of not merging the channels.
    image
  2. Overlay: I am a bit confused about what does the overlay show exactly? In some slices red colour only outlines the cells but in some slices it fills the cell completely. Does this indicate good segmentation?

Hi Omkar,

maybe I can quickly answer your two questions:

  1. No merging will be performed unless you press the “Merge” button. If you do the normal segmentation, you don´t need to worry about merging. Similarly, no transfer will be done unless you hit “Transfer”.

  2. In the overlay, the outline of cubes are shown red. Since cubing is also performed in the z-direction, all objects appear red when you scroll to a layer just at the top of the cubes. This is perfectly normal and does not have any implication on your segmentation quality.

Hope this helps,

Hannah