Hide specific dataset(s) from embedded graphs about conos HOT 4 CLOSED

from conos.

Thanks! I was able to get it to do (mostly) what I wanted! Note: show.na was incorrect, it is `plot.na. However, I can only seem to get it to plot colored by sample, not with clustering or with gene expression.

To dump the full cell->sample relationship table I used:
write.table(con$getDatasetPerCell(), file = "FILENAME", sep="\t", quote= FALSE, col.names = FALSE)

Then manually replaced the sample IDs I wanted to set to NA.
Then I added the plot.na=FALSE parameter to the appropriate conos $plotGraph functions.

Okay, so I think what I'd need is to dump the tables for cell->cluster and cell->gene relationships as well, modify and read back in those tables as groups annotation with plot.na=FALSE, but can't seem to find a function to dump that information.

from conos.

Maybe the answer is a bit late, but if anyone else will be looking for a similar method, I find it the easiest to save the conos plot to a ggplot object and work on that. So in your case you would do

samples_to_plot <- c() # Vector with sample names you want to keep on the plot
p <- con$plotGraph()
p$data$Sample <- con$getDatasetPerCell()
p$data <- subset(p$data, p$data$Sample %in% samples_to_plot)
p

With this approach you can also make separate plots for each of the samples (or just some of them). Simply add:

p + facet_wrap(~Sample)
p

from conos.

Okay, so I think what I'd need is to dump the tables for cell->cluster and cell->gene relationships as well, modify and read back in those tables as groups annotation with plot.na=FALSE, but can't seem to find a function to dump that information.

Did it in #39. Now, if groups are provided, they used to filter which cells are plotted even if you color by gene.

from conos.