NeuriteTracer plugin

[Thomas Hering]

Dear ImageJ group,

I have a question about the NeuriteTracer plugin, for automatically quantifying neurite outgrowth in neuron cultures. We are having problems obtaining accurate counts of neuronal nuclei. Reference for this plugin is below:

Journal of Neuroscience Methods: Pool M, et al., NeuriteTracer: A novel ImageJ plugin
for automated quantication of neurite outgrowth, J Neurosci Methods (2008), Volume
168, Issue 1, Pages 134-139 (http://dx.doi.org/10.1016/j.jneumeth.2007.08.029).

In the paper describing this plugin, rat cerebellar and hippocampal neurons and chick DRG neurons were analyzed. A sample image set was provided with the plugin that we were able to analyze successfully. We are using a clonal derivative of the PC12 cell line (NS-1), which appear to have larger nuclei and more cytoplasm, in addition to a more complex pattern of outgrowth than the cells mentioned above. We have not been able to get accurate automated counts of the nuclei in these images.

I have resorted to using thresholding nuclei in an image stack, and using "analyze particles" under the "Analyze" menu. If I set the size to "100 to infinity" and the circularity to 0 to 1.0 I can get reasonably accurate counts. There are instances where a clump of several nuclei are counted as one, and I correct for these manually.

I believe that the "Analyze Particles" function is used in the plugin, but the parameters (size and circularity) are fixed, and are not appropriate for PC12 cells. How can I adjust the parameters in the plugin to count nuclei automatically?

Another problem is that the cytoplasm surrounding the nuclei is "skeletonized" along with the neurites to give a circular tracing  that is added to the neurite tracing. Any way to deal with this issue also?

Look forward to suggestions...Thanks in advance!

Tom

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[Citron, Bruce]

Tom-
I just went through a lot of trial and error with NeuriteTracer, which we
last used for a project a couple of years ago.
As a first shot, try using ImageJ 1.44o (64-bit).  For some reason, some
nearby versions were giving us very low nuclei counts, but just switching to
1.44o fixed it and it is downloadable from the ImageJ site.

But that might not be the issue, since you mention that the sample image set
counted OK.  During the past project, we found that if we converted what we
were measuring then (512x512 images) to 1024x1024 with no interpolation or
math, then all nuclei were detected when we had a similar problem.

Finally, Madeline Pool did address some size ranging for the nuclei
evaluation but I don¹t think that it is in the current macros from the site.
Does a window appear that includes a query about minimum and maximum nuclear
area?  The macros dated Nov 17, 2010 (except for
batch_RGBmergetracings.txt=Aug 21, 2010) include this feature and might
solve your problem.  The Macros of Aug 21, 2010 do not.
-Bruce
--
Bruce A. Citron, Ph.D., Bay Pines VA



> On 7/12/12 12:25 , "Thomas Hering" <[hidden email]> wrote:

> Dear ImageJ group,
>
> I have a question about the NeuriteTracer plugin, for automatically
> quantifying neurite outgrowth in neuron cultures. We are having problems
> obtaining accurate counts of neuronal nuclei. Reference for this plugin is
> below:
>
> Journal of Neuroscience Methods: Pool M, et al., NeuriteTracer: A novel ImageJ
> plugin
> for automated quantication of neurite outgrowth, J Neurosci Methods (2008),
> Volume
> 168, Issue 1, Pages 134-139
> (http://dx.doi.org/10.1016/j.jneumeth.2007.08.029).
>
> In the paper describing this plugin, rat cerebellar and hippocampal neurons
> and chick DRG neurons were analyzed. A sample image set was provided with the
> plugin that we were able to analyze successfully. We are using a clonal
> derivative of the PC12 cell line (NS-1), which appear to have larger nuclei
> and more cytoplasm, in addition to a more complex pattern of outgrowth than
> the cells mentioned above. We have not been able to get accurate automated
> counts of the nuclei in these images.
>
> I have resorted to using thresholding nuclei in an image stack, and using
> "analyze particles" under the "Analyze" menu. If I set the size to "100 to
> infinity" and the circularity to 0 to 1.0 I can get reasonably accurate
> counts. There are instances where a clump of several nuclei are counted as
> one, and I correct for these manually.
>
> I believe that the "Analyze Particles" function is used in the plugin, but the
> parameters (size and circularity) are fixed, and are not appropriate for PC12
> cells. How can I adjust the parameters in the plugin to count nuclei
> automatically?
>
> Another problem is that the cytoplasm surrounding the nuclei is "skeletonized"
> along with the neurites to give a circular tracing  that is added to the
> neurite tracing. Any way to deal with this issue also?
>
> Look forward to suggestions...Thanks in advance!
>
> Tom
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html