jacobpwagner
commented
5 years ago Hi @JimboMahoney , I'll address the issues in order:
1) Yeah, I understand it's a little frustrating. The choice to use the recursive=TRUE option in list.files within diva_to_gatingset is intentional as it makes it easier for users to have the xml in a top directory and have the FCS files in a nested data directory, without having to explicitly point it there with the path arg. If that's a real problem, we could maybe allow the user to pass in a value to recursive that would be honored by list.files so recursive=FALSE would only look in the top level. Either way, it's dangerous to silently guess how users would want the issue of duplicate filenames resolved, so it's safer to just error and prompt them resolve it.
2) If you could attach the Diva experiments and FCS files or a location where I could access each, that would help. Until I have a chance to look at it, a few quick things to try/examine:
First, are the gate definitions for the new gates getting read in correctly? Is this successful (and new_gate looks as expected?:
new_gate <- gs_pop_get_gate(gs, "<new node name>")` Second, what is the result after recompute(gs)?
JimboMahoney
I have some ideas about what is causing this, but it would help me if you could paste in the output of 
Hi there!
Firstly, MANY thanks for this code! I've used it successfully to compare manual gates with automated clustering.
However, I've found a couple of issues, one of which I can work around and one I cannot.
Here is my code:
ws <- open_diva_xml(file)diva_get_sample_groups(ws)gs <- diva_to_gatingset(ws, name = 1, subset = 1, worksheet="global")There is also this, but I'm not entirely sure it's necessary:
gh <- gs[[1]]Issue 1:
The third line "breaks" if the directory containing the XML / FCS files contains subdirectories with files of the same name. For some reason, it also looks in the subdirectories, finds files with the same name and then reports that it cannot open the specified FCS file.
At first I was stumped, but after removing the subdirectories, it opens the file OK.
Issue 2:
I'm not sure if this is an issue with Diva (I'm using 8.0.2) or with CytoML. I'll try and explain.
I have an experiment with the original FCS files captured in Diva. This opens fine, as well as all gates and cell populations.
However, I get very strange results if I do the following:
1) Export a subset of that experiment (e.g. after doing basic gating, such as lymphocytes and singlets) to an FCS file (so that I have only the cells I'm interested in). 2) Open that FCS file as a new experiment in Diva. 3) Set my gates. Looks good in Diva, with all stats being as I expect. 4) Export the experiment. 5) Import using CytoML.
The gates look good using this code:
plot(gs)and also this:
gs_get_pop_paths(gh,path="auto")However, the stats are only correct for the first two gates:
gh_pop_compare_stats(gh)Any further gates are unpopulated, except for one that only contains 4 events (it should contain 598).
In addition, the plotting of the graphs doesn't work:
plotGate(gh)It gives the following error:
"Error in valid.viewport(x, y, width, height, just, gp, clip, xscale, yscale, : invalid 'yscale' in viewport"
And a bunch of the following warnings:
"3: In min(x) : no non-missing arguments to min; returning Inf 4: In max(x) : no non-missing arguments to max; returning -Inf"
Any ideas on what's going on would be appreciated! Let me know if you want / need the Diva experiments (they're only small).
Thanks!