Flow.control error

[kmunozforti]

I keep getting this error

flow.control <- define.flow.control(control.dir, control.file, asp)
Checking control file for errors
Reading control information
Determining the gates that will be needed
Matching and replicating negatives per gate
Determining channels to be used
Defining gates
Error in dimnames(x) <- dn :
length of 'dimnames' [1] not equal to array extent

[DrCytometer]

This is the sort of very unhelpful error message that R produces constantly. Something has gone wrong somewhere, resulting in a mismatch in what exists and what is expected, but we don't know what caused it or when it happened. This, unfortunately, means I can't troubleshoot from this error message.

Did you get plots produced of the gates in the folder "figure_gate"? If you did, did you get a diagnostic gating failure plot?

It looks like you've run directly to define.flow.control without first pre-defining the gates. Since the gate definition process can be a bit slow and may not produce the optimal gate for every set of data on the first go, the recommended procedure in workflow example is to first define the gates and pass them to define.flow.control.

[DrCytometer]

I believe I have tracked down the cause of this. If you have multiple samples using the same gate, but one or more of those is wildly different, all the cells in the differing sample(s) will fall outside the gate, returning no events, and triggering this failure. This is the reason you should create gates first. You, as the user, know your controls best. If you are using two different types of beads and just put "beads" in the control file, define.flow.control thinks they're all the same and tries to figure out a single gate, which will match the more abundant particle type. See this gating article for guidance.

I am pushing a patch in v1.5.7 that will issue a warning if this situation occurs. This will not fix the problem. It will only tell you there is a problem.

FYI, this whole process of gating and setting up the workflow should become much easier in the upcoming release of AutoSpectral in Honeychrome.

[DrCytometer]

With the newly released Version 1.6.0, this issue can be avoided entirely by running the new automated spectral extraction pipeline to identify fluorophore profiles. Try get.spectra.automated() and see the new workflow instructions.