Open PaulRegnier opened 7 months ago
@PaulRegnier does changing theme_flowGate work for you? I've tried sourcing your bug fix but still get the same error message.
Hello @Biomiha ,
Yes it works! In addition to what I wrote in my first message regarding the Bioconductor release of flowGate
, I realized that the flowGate
GitHub fork I made also works, so you need to make sure that the right version of the package is installed, at least until the maintainer of flowGate
actually incorporates this fix.
If you don't know what version is used, then I suggest you to manually remove the package flowGate
from your library, then force the installation from the flowGate
fork I made:
library("devtools")
install_github("PaulRegnier/flowGate", force = TRUE)
Then load PICAFlow
and it should work.
That works, thanks!
Hello Andrew,
I am the developer of the
PICAFlow
R package (https://github.com/PaulRegnier/PICAFlow) which basically serves as an all-in-one processing and analysis package for flow/mass cytometry data. Long story short, I discovered yourflowGate
package at the end of the past year and decided to implement your package inPICAFlow
to interactively draw gates on samples.It was working like a charm until I tried again today after a long period without analyzing cytometry data, and realized that your Shiny application was not showing the graph (through your
gs_gate_interactive()
function).The R Shiny application actually opened but a red message showed up which complained that the
legend.text.align
parameter was not defined. This seems to be related toggplot2
. I remembered thatggplot2
got some huge updates in the past weeks/months, so that could be linked. I had to correct a bunch of other functions (unrelated toPICAFlow
) which showed similar errors recently.Nonetheless, I finally managed to fix this strange bug by adding the following code into your
theme_flowGate
variable (right after the line 75 of thepreparePlot.R
file, from the latest Bioconductor version):, legend.text = ggplot2::element_text(hjust = 0.5)
Would you mind to implement this line of code, quickly test it and eventually push it to the Bioconductor repository?
Thank you again for your kind help and again congratulations for
flowGate
, I really like it :)NB: I also have a suggestion for you to implement in
flowGate
, but this is not urgent at all. Actually, it would be very useful if one could directly specify in the arguments of thegs_gate_interactive()
function what manual X/Y limits we want to use. To illustrate, inPICAFlow
, I allow users to draw a global gate using the first sample of the flowset, but afterwards I also allow users to draw specific gates for some samples if they are very different from the template sample. So I iterate over the samples to re-gate, but users must every time reset the manual X/Y limits even if they are the same for the samples to re-gate. What do you think of this?