Gating on a single unstained control

[hrj21]

Hi team, thank you again for the fantastic package.

We encountered a situation today that we could not progress past. I'll describe it and perhaps you can tell me if we are not following the workflow correctly.

We have a reference group made up of PBMCs, including an unstained PBMC sample used as their universal.negative. The experimental samples are of another tissue however, so there is a single unstained sample to capture their autofluorescence. We have this unstained sample in the control file with its fluorophore column set to "AF", and it is the only sample in the table with its particular gate.name. When we run tune.gate(), we get the following message and error:

Unstained samples were included in the files to be used for gating. These have been removed: NA Error: No samples left after excluding unstained samples. Please check control file and try again.

It seems this function drops unstained samples and as the only sample associated with this gate is unstained, there is nothing left for gating. So my questions are:

• Why does tune.gate drop unstained samples in the first place?
• How should we be gating a singular unstained sample for autofluorescence?

I can share an image of an example control file that results in this, if my description isn't clear.

Best wishes
Hefin

[nlaniewski]

Hi Hefin,

Sorry that you have encountered some issues!

To reiterate:

1. Reference group of single-color controls, all stained using PBMCs (including an unstained PBMC)
2. Experimental samples of another tissue type (including an unstained tissue-specific control)

For generating the spectra, you would use only the reference group (stained single-color PBMCs and the unstained PBMC); during the unmixing process is when you would specify the tissue-specific unstained for per-cell AF extraction unmix.fcs().

Are both unstained types (PBMC and tissue-specific) included in the control file? It should look something like this example with a single row defining the unstained PBMC (exclude the tissue-specific control for now); then try the automated workflow using get.spectra.automated().

To better answer your first question: as you suggested -- if it's possible to paste images in these comments -- then sharing an image of the control file would be helpful, thanks.

For your second question, the 'gating' for experimental/tissue-specific autofluorescence is done using self-organizing maps (via FlowSOM); this article includes the workflow for generating your tissue-specific AF profiles. However, you need the spectra first, so let's try to get that solved!

Best,
-Nate

[hrj21]

Hi Nate,

Thanks for such a quick response. Aha I see. So all events are used to build the SOM from the tissue-matched unstained sample. I can see how this and the cell-specific AF extraction make pre-gating redundant.

Yes, we had thought to put both the PBMC and tissue-specific unstained files in the control file as below (rows highlighted for convenience), but I understand now this isn't the expected workflow.

I'm awaiting feedback from my colleague who's experiment this is to see how she gets on after doing as you say, but I thought I would reply in the meantime.

Best wishes
Hefin

[nlaniewski]

Hi Hefin,

Thanks for attaching an image of the control file -- this really helps.

1. As you highlighted, you would remove the affected row containing F1 Well_031_20260803_TMNC 2D_TCE_d3_D1 unstained.fcs
2. You would then set H9 Negative_cells (Cells)_Single Stains.fcs to AF under the fluorophore column
3. Set H11 Negative_beads (Beads)_Single Stains.fcs to Negative under the fluorophore column

There is expected naming convention (unless Oliver has re-worked the code), so unstained beads are set to just Negative and unstained cells to AF -- everything else in your control file looks good.

With the edited control file, you should be able to derive spectra using the legacy or automated workflow -- please let us know if that's not the case.

Best,
-Nate