Chromatin Project, short-term goals

Issues with current data analysis

In trying to get of order 100 loci per chromatin domain I’ve take a bunch of partially imaged edge of focus domains or other less well defined images (e.g. near edge of field of view). These un-necessarily blur out the size distributions of regions. A more stringent filter will be better.
I think this is maybe even more important for statistics like moment of inertia and total localizations
I haven’t used the z-scan information for number of dots per cell. This is pretty important. We should make this a field of the export data so that each spot has

Let’s get Z information and other quality metrics into the chromatin cropper.
Let’s plot the spots statistics on a scatter plot so we can see where each of them go

the two to three step annotation process is unnecessarily slow. This should all be done in the Chromatin Cropper
Each individual dot should save
- its raw data
  - vlist
  - imaxes
  - impars
  - imdata (Iconv, Istorm, Itime, Ihist XY YZ XZ, Icell, Iarea). For just that dot
- its ID information
  - region ID (e.g. F03). This can be used to retrieve locus size, locus coordinates, ChIP seq details etc.
  - daxfolder
  - binname
- its dot quantification statistics
  - Area
  - Area Map
  - moment of inertia
  - something like distance of local maxima from edge. Distance of internal local minima from edge. Define maxima/minima based on quantiles. Say local max are all the spots that hit 90% or above. Local min are all the local min that remain if we saturate 10% max and above.
  - Coefficient of variation for localizations per pixel
  - Water-filled areas
- png images
  - use export_fig.m function with set(fighandle,'PaperPosition',[0,0,width,height]). This will capture the set(gcf,'color','k') and other properties, and preserve the spacing.
  - conventional and STORM dot pair. Just re-apply an auto-contrast for the conventional image and then convert to the correct colormap. STORM image may require manual contrast adjustment
  - Area