potential error in flow_cluster_id_score

[Mirin1357]

Hi DrCytometer,

first, thank you so much for trying to make this analysis a bit more user friendly!

I am using two randomly selected ID7000 files gated in FlowJo on LDneg CD45+, exported as FCS. All parameters have been transformed in ID7000 previously as biexp, -300 base width (otherwise, step source( file.path( fcs.src.dir, "flowcytoscript_define_experiment.r") ) is not working, and I was unsure where to specify instrument transformation settings manually in the workflow).

I have encountered an issue when arriving to the source( file.path( fcs.src.dir, "flowcytoscript_clustering.r") ) step. The first part goes through, but after I am asked which clustering parameter to use, independent of the choice, the code breaks with the following error:

I went to the source code and running step by step, scaled.id.score actually produces empty matrix.
scaled.fcs.cluster.median looks like this:

marker.list like this (just a snapshot of the file):

However, once it is run through the scaled.id.score, the matrix is actually empty

I have tried using both original source files and one from the 20260629 folder but with no change.
Please let me know if you would like me to attach any additional files or script.

Thank you in advance for your help!

[DrCytometer]

Hi Mirin,

I haven't been maintaining this because other people have made what I view to be much better tools for the same purpose, namely cyCONDOR and MARMOT. https://github.com/lorenzobonaguro/cyCONDOR https://github.com/peterleary/MARMOT As it may be a bit before I can look into this, you might want to check out those.

[DrCytometer]

So, you're meant to rename your markers into actual marker names rather than relying on the column names generated by FlowJo. This would mean "CD4 : PE-Cy7 - Area" would become "CD4". If you don't do that, the script wouldn't figure out which markers were present, the scoring of marker expression would fail (no markers to score), and it stumbles to a halt. I'm putting up a new version that does a better job of matching the markers for you. Try it and let me know if it solves your problem (ideally, run again without changing the marker names first so we can see if it fixes that problem). You should also now get a "confidence score" for each cluster identification. Keep in mind that the script can really only identify clusters to cell types if you have the relevant markers present, and even then, it's not amazing at it.

[Mirin1357]

Hi DrCytometer,

thank you for looking into this!
I pre-gate on option 1, which is viable cells in immune mouse tissue. I have tried with and without renaming the markers (I chose Ki67, CD8a and CD4), but the script fails in both cases just after I select "1: All" with following error:
Error in FUN(newX[, i], ...) : arguments must have same length

However, as you suggested, there are other R packages and if you are not maintaining this one anymore, I would not want to waste your time for troubleshooting.

Thank you in any case!

[DrCytometer]

Did you try this with the updated script? I just updated it a couple of days ago.

[Mirin1357]

yes, all clean downloaded and ran from the start

[DrCytometer]

great, thanks for checking!