Define Rosetta matrix for compensation on MIBIBlue slides

[alex-l-kong]

Relevant background

More lab members have started running cohorts on MIBIBlue slides. While this will remove a lot of work needed for compensation (especially on Au), it does introduce other organic contamination that needs to be addressed.

The main culprit is the Barium channel, which bleeds into a lot of essential antibodies. Jon Pollack encountered this on his cohort, Patricia + Mamon are exhibit B, and Marc + Abdallah are most certainly going to be exhibit C. Barium (mass 138) is not currently defined in the Rosetta matrix and needs to be added.

Other masses may also need to be adjusted depending on how the Rosetta images turn out. For example, on Jon's cohort:

1. CD45 introduced more contamination on CD20
2. CD11c was unsalvageable
3. dsDNA introduced more contamination on FOXP3

Good news is the existing Noodle coefficients seem unaffected by the MIBIBlue slide.

Design overview

Mamon is doing some work specifically on the Barium channel to see if there's a correlation coefficient that can help define the specific multipliers needed. I helped him extract Barium as an intensity image instead of a pulse image to see if that will also help.

More likely than not, we'll need insight from Erin and Noah to see how they empirically defined the subtraction coefficients for each source-output channel pairs. The process they used isn't detailed in any paper released by the lab.

With the Noodle range extraction correction, we'll build off of the Rosetta v1 matrix. The fact that Au will be much dimmer is further justification for this.

Code mockup

Most of the code will be kept the same as the only addition is a new file. The only real change is we'll need to explicitly add "Barium" to the default list of intensities pending Mamon's research.

Additionally, there should be an option in the Rosetta notebook to define whether a MIBIBlue slide was used or not:

# change to True if a MIBIBlue slide was used
mibiblue = False

On a side note, people call the masses extracted as intensities by default all sorts of things (especially 197: I've seen "au", "Gold", etc. instead of the default "Au"), so a separate PR will need to make it more flexible for users without having to explicitly set the intensities parameter in the extraction notebook.

Required inputs

The final Rosetta file will be in commercial_rosetta_matrix_mibiblue_v1.csv. We should also change the old Gold-based input files to commercial_rosetta_matrix_goldslide_..., something along these lines.

Output files

Outputs will be the same.

Timeline
Give a rough estimate for how long you think the project will take. In general, it's better to be too conservative rather than too optimistic.

• A couple days
• A week
• Multiple weeks. For large projects, make sure to agree on a plan that isn't just a single monster PR at the end.

Estimated date when a fully implemented version will be ready for review:

Optimistically before winter closure. This will depend a lot on further testing, however.

Estimated date when the finalized project will be merged in:

Optimistically before winter closure. This will depend a lot on further testing, however.