labeling in dittoDimPlot overrides factor order during faceting about dittoseq HOT 5 CLOSED

I just tried it out and it works for me - thanks for the fix!

from dittoseq.

Ah, this may be a bug in the dataframe created to hold locations of the labels after splitting -- that's manual unfortunately, not through ggplot. I'll look into this. Thanks for pointing it out!

from dittoseq.

Indeed. I did not keep the split.by-groupings columns, in the label-locations data, as factors at all. So then ggplot defaults back to alphabetical/numerical ordering because it notices that there is disagreement between the per-cell split.by data versus the per-group split.by data.

In the split-label branch, I've added code to retain factor levels which should fix the issue; it looks right for me with the example you give above. If you install from that branch with
BiocManager::install("dtm2451/dittoSeq@split-label"), then restart R, and then try with your data, could you confirm that the issue is fixed for you?

from dittoseq.

Thanks, Dan for your prompt response! I am teaching a workshop tomorrow on 10X and I don't want to change anything just yet, but I will try in the next day or two.

from dittoseq.

No worries, but whenever you get a non-inconvenient chance, I'd appreciate it!

I've checked a bunch myself and am pretty sure it's good. I just feel like there's no going back once something is pushed to BioC, even though that's not at all actually true, so I like to be double/triple/qraudruple -sure lol.

from dittoseq.