marker expression maximums change across batch

[No2Ross]

Hi,

I am analysing Aurora data and have put the compensated unmixed data (from three batches) through cytoNorm to correct for the batch effects.

The batch corrected expression values for the markers are mostly normal looking, but for one marker, there is a massive increase in expression for some of the cells which seems to be batch dependent. The following plot shows the maximum expression (arcsinh transform of the batch corrected values, cofactor = 6000) of the markers across the A, B and C batch. The marker of interest is highlighted in a green box.

When you look at the expression of this marker across the cells of the reference datasets, they seem to be comparable:

When you just take the arcsinh (cofactor = 6000) expression values (i.e. not batch corrected), the maximum expression across the batches seems equal:

I'm thus wondering what is causing this markers expression to jump so high/how I can figure out why this is happening. Or perhaps this is expected behaviour, if so what does the interpretation of the batch effect mean?

I should also note, that the majority of the cells have comparable expression of this marker across the batches. For batch C, there is only 39 cells (out of 3,593,692 total cells) which have really high expression of this marker.

Best wishes,
Ross

I'm running R version 4.4.1 and CytoNorm version 2.0.2