If you haven't already, please read through our contributing guidelines before opening your PR
What is the purpose of this PR?
After testing the mixing score on 100 images and across various populations, we determined that total cell count was more useful to restrict rather than counts for the individual populations. As a result, we need to be more restrictive with the ratio of cells that we allow. The defaults that seem to work best are ratio_threshold=5 and cell_count_thresh=200.
For plotting the ratios in the mixing notebook, we should drop any bigger than 15. Large ratio values just mess with the x axis and makes the plots hard to interpret.
How did you implement your changes
Change the threshold default values in both the notebook and the code. For the output file, only return one cell_ratio value per image, instead of both the pop1/pop2 ratio and pop2/pop1 ratio.
Adjust the notebook to
Check for and automatically generate the specified neighbor matrix, rather than rely on the user to run the correct cell.
Tweak the ratio plots to have x-limit 15 and the mixing histogram plot to always start at 0.
If you haven't already, please read through our contributing guidelines before opening your PR
What is the purpose of this PR?
After testing the mixing score on 100 images and across various populations, we determined that total cell count was more useful to restrict rather than counts for the individual populations. As a result, we need to be more restrictive with the ratio of cells that we allow. The defaults that seem to work best are
ratio_threshold=5
andcell_count_thresh=200
.For plotting the ratios in the mixing notebook, we should drop any bigger than 15. Large ratio values just mess with the x axis and makes the plots hard to interpret.
How did you implement your changes
Change the threshold default values in both the notebook and the code. For the output file, only return one
cell_ratio
value per image, instead of both the pop1/pop2 ratio and pop2/pop1 ratio.Adjust the notebook to
Remaining issues
N/A