Hi Kristopher,

This cluster analysis is somewhat similar to stuff that happens in our lab, and we have a Post-Doc (Thibault Lagache) working specifically on calculating the clustering nature of X (vesicles, cells, you name it) using radial distributions (Ripley's K function actually). He wrote a couple of plug-ins for the open-source Icy imaging software, might worth a try:

http://icy.bioimageanalysis.org/plugin/Compaction_profiler

http://icy.bioimageanalysis.org/plugin/Spatial_Analysis

(has relevant pointers to associated publications)

It's Java-based but has a script editor where you can do some further hacking in Javascript, Jython or plain Python (R under development).

Thanks for the info, will check it out.

I have a similar approach to this plugin but I use the spatstat package in R and the statistical test written by Marcon, Traissac and Lang (https://www.researchgate.net/publication/236331582). This will get me to the result of whether my cells are distributed in a random, clustered, or dispersed form.

What I was hoping to achieve was the classification of cells based upon their HIK or K function, such that members of D or T clusters could be split into separate groups and ultimately reveal the patterning of 2,3 or more distinct cell classes with overlying distributions. Sort of a kmeans approach using the K function rather than the intensity.

Article A Statistical Test for Ripley's K Function Rejection of Pois...