search

angelolab / toffy

Scripts for interacting with and generating data from the commercial MIBIScope
MIT License
10 stars 3 forks source link

Rosetta coefficient tracking #55

Open ngreenwald opened 2 years ago

ngreenwald commented 2 years ago

Please respond to this thread with the changes you made to the rosetta matrix.

  1. Attach the modified the .csv file of your matrix
  2. Summarize what major changes needed to be made (i.e., a certain channel required significant alteration to produce usable data).
  3. Summarize what minor changes needed to be made (i.e., a certain channel required minor modifications to improve the data).
mkammersgaard commented 2 years ago

  1. commercial_rosetta_matrix_mult_0.5_exceptPD1.csv

  2. PD1 needed to be at multiplier 2 for Noodle to remove bright edges

  3. Aimed for one common multiplier for Noodle across all channels. I explored multipliers between 0.25 and 2, and 0.5 was the overall best fit across all channels for preserved signal with sufficient background removal

TerGla commented 11 months ago

Would it be possible to add another quality control check after the rosetta compensation, e.g. scatter plot of the pixels for all given marker combinations to see if there are potentially some spill overs?

ngreenwald commented 11 months ago

Hi @TerGla, we don't currently run those types of checks, but if you think it would be useful you're welcome to put something together.

We usually rely more on manual inspection to see if it's working correctly.