Page 1 of 1

IMC data processing-Normalization & Scaling

Posted: Wed Jun 12, 2019 6:34 pm
by sbooth
Hi there, I am wondering if anyone has a some example code that shows how to 1-99% normalize and 0-1 scale the IMC intensity data - in R hopefully :)

After extracting the cell and image intensity values from cellprofiler using the Measure_Mask_Basic pipeline from the Bodenmiller group I have a HUGE csv of the cell and image level intensities. But my specific questions are:
1. Do all intensities first need to be multiplied by 65535? Corresponding to the unit16 division of 2**16 on cellprofiler output
2. Should the 1-99% normalization & 0-1 scaling be done on all files and channels together, or if each channel and each file should be normalized separately?
3. Should the same process be applied to the cell level expression extracted from Histocat csv files?

Thanks for any assistance!
Steve