Problem with declumping object

I d like to know how to set the cube side length during objects declumping. Even though i could use imagej to quantify the length,but i am not quite sure how to determine the voxel of the small clusters further during post-processing.

Hi Yuanhao,

the choice of cube side length strongly depends on your data as well as the type of question you want to explore with BiofilmQ. It determines the spatial resolution of any parameter calculated, so in order to figure out which cube size you need, you may want to think about the following questions:

What is the length scale of the smallest structure of interest (e.g. a single cell or cell cluster) in your data? Half of this value roughly sets a lower threshold for the cube side length you should choose - it simply does not make sense to set the resolution smaller than your smallest structure of interest (I´m proposing half here, because the cube positions won´t perfectly match your objects).

Which property do you want to quantify with BiofilmQ and over what length scale does this property show variation? Half of this value roughly sets the upper limit for your cube side length, because in order to resolve the variation of this property, the cubes will need to be smaller than this length scale.

Since a smaller cube side length results in more cubes, which in turn leads to longer computation times, you want the side length to be as small as necessary, but as large as possible. To get a feeling for how different choices correspond with the structures in your image, I recommend to pick a representative image, perform the segmentation and cubing, and then check either via the overlay function (magenta in the image below) or the segmentation preview (cyan in the image below) - both are also described the general overview tutorial ( - what the result looks like.

I hope this description was helpful to you and please write again if you have any more questions.



1 Like

Thanks for reply.There is another question confused me for a long time, can BiofilmQ quantify live/dead ratio of biofilm in stark? And how?

Hi yuanhao,

you can anyalyze live/dead ratios in BiofilmQ if you have both live and dead cells visible in separate channels (for example a constitutive reporter and a dead stain). One way to go about this is to first perform segmentation on each of those channels, then merge them, and visualize the properties Cube_RelativeAbundance_chX and Cube_Overlap3D_chX_chY (where X and Y are the channel numbers).

The merging can be performed in the post-processing tab of the segmentation - in the screenshot below I marked the area where you can set the channels to merge red and the button you´ll have to click blue. Note, that the result of the merging will be saved in the channel indicated in the second text field and that the segmentation results for the individual channels are no longer going to be available (except as a back-up in case you want to undo the merge). If you wish to calculate parameters both on the merged channel and the individual channels, the easiest way to go is to simply copy your directory and work separately in the original and the copied folder.

Cube_RelativeAbundance_chX describes for each cube the percentage of the segmented biomass of channel X in relation to the added biomass of channels X and Y. This means, that Cube_RelativeAbundance_chX and Cube_RelativeAbundance_chY always add up to 100%, even if they overlap. I believe, this might be the ratio you are looking for? If you want the ratio over the complete biofilm, you can have a look at the statististical properties (mean, median, percentiles,…) that are calculated for the whole biofilm and each property in each time point and available as global parameters. They can be explored in the visualization tab - check out our tutorials or documentation on this, if you haven´t done so already.

Let me know if there is anything else I can help you with or if you´d like a more detailed explanation on one of the above descriptions. I am always happy to help :smiley:



PS: If you have a question regarding a new topic, feel free to start a new forum post about it. This will make it easier for other user to find relevant posts for them, since most of the time people just skim through the titles and don´t take the time to read each thread. Thank you :smiley: