Grenoble-Institute-Neurosciences / AutoNeuriteJ

7 stars 3 forks source link

Potential Issue part III #14

Open sonjadin opened 9 months ago

sonjadin commented 9 months ago

Hello, Thanks a lot for this macro and all the work this must have been required !!!

I am trying to make it work on macroscope images (x80) of a primary culture of DRGs (red staining for Tuj and blue for DAPI). Part I and two worked fine but when I try Part III, it just keeps running for hours and hours (since yesterday evening). On the FIJI bar I can see "Thinning: (and then a number that goes up)".

For info, I have selected as a directory the subfolder created by Part II (the folder inside the resultats_autoneurite).

In any case, thanks a lot in advance !!!

Grenoble-Institute-Neurosciences commented 9 months ago

Hi Sonjadin!

Sorry for the answer waiting. I believe this issue is due to the very highly branched shape of DRGs. I have tested AutoneuriteJ on these neurons and it's a bit tricky.

- How long would you expect Part III to run ? It really depends on the number of neurons analysed and their complexity.

- When I did Part II on 2 different images at once, it created a subfolder for each of these, but only one is not empty... Maybe I should run part II independently for each image ? Part2 is supposed to create a subfolder for each image but if a well remember they need to be in separated folder to begin with so that the second subfolder do not erase the first one.

- Bonus question: do you have any tips for adapting this for the type of images I have ? I would recommend to be careful on the accuracy of the outcomes since DRGs are heavily branched there is strong chance of overlapping neurites. There macro will cut at the weakest staining intensity to try to open the connected neurites and make them seperate but this strategy is not necessarily the best.

So, it is possible if you have a lot of neurons analyse that AutoneuriteJ take ages to quantify all the tiny neurites. I would recommend to try to lower the resolution of your images a much as possible (On imageJ you can go on "Image > Scale..." and try to rescale the image to a lower size. This will save you a lot of Part III running time.

Please let me know if you need further help!

Best, Benoit