Help interpreting CellBender output and choice of parameters

[AClab-sgarcia]

Hi CellBender team,

Thank you for providing such a powerful tool!
I’ve recently run CellBender on some of my 10x single-cell RNA-seq samples, and I have a couple of questions about the interpretation of the results and the choice of parameters.

In particular:

1. Are the parameters I’ve used within the typical or recommended range?
2. For samples L04 and L05, I ran CellBender using different values for --total-droplets-included. I'm unsure whether my initial runs were sufficient, or if the ones where I increased that parameter (and which now produce “cleaner” or more expected plots, highlighted in blue in the attached PDF) are more appropriate.

Additionally, I’ve noticed that the “Determination of which barcodes contain cells” plots don’t always show the expected behavior, and in general, I’m not very confident in the results I'm seeing — they appear noisier or less defined than I anticipated. For sample L07, in particular, the plot shows an unusual white gap that makes me question whether the output can be trusted, or if it indicates a problem with the run.

To clarify all of this, I’ve attached a short PDF summarizing the parameters used and the output plots for each sample. I would really appreciate it if someone could have a quick look and help me determine whether the results look reasonable and which runs should be prioritized for downstream analysis.

Thanks again for your time and support!

CellBender outputs summary.pptx