Question

FlowCore/FlowViz: How do I skip plotting of empty populations that would otherwise cause an error?

0

Entering edit mode

shieldsy05 • 0

@shieldsy05-13363

Last seen 7.5 years ago

Hey,

I'm trying to process a whole heap of flow data files (obviously), and I've set up gating strategies. My problem is that due to the nature of my research, in some files, a gated population will be empty, and this causes errors when I try to create a density plot (plots that population for each file until it reaches an empty population, then stops). It's inefficient for me to identify and remove the files, and there are other populations of interest aside from the population that is empty.

How can I remove or skip empty populations, so that I don't have to go back each time and find/delete the FCS files?

Example code:

fn <-list.files(pattern=".fcs")
fs <- read.flowSet(fn,alter.names=TRUE)
qg1 <- quadrantGate(fs,stains=c("CD19","CD3"))
fsqg1 <- split(fs,qg1)
bc <- fsqg1$`CD19+CD3-`
dpl <- densityplot(IndividualID~`FITC.A`,bc)

Thanks!

flowcore flowviz R bioconductor flowstats • 1.9k views

ADD COMMENT • link updated 7.5 years ago by SamGG ▴ 360 • written 7.5 years ago by shieldsy05 • 0

score 0 · Answer 1 · 2017-06-30

0

Entering edit mode

Haiying.Kong ▴ 110

@haiyingkong-9254

Last seen 5.7 years ago

Germany

If an empty .fcs file does not have even a header, and the file size is 0, add to between the line 1 and the line 2:

idx = which( file.info(fn)$size > 0 )

fn = fn[idx]

ADD COMMENT • link 7.5 years ago Haiying.Kong ▴ 110

score 0 · Answer 2 · 2017-07-03

0

Entering edit mode

SamGG ▴ 360

@samgg-6428

Last seen 16 days ago

France/Marseille/Inserm

Hi,

bc is a flowFrame. When it is empty, nrow(bc) should return zero.

ADD COMMENT • link 7.5 years ago SamGG ▴ 360

0

Entering edit mode

It's a start. Is there a way that I can make the densityplot function skip over the empty frames within a subset? bc is the flowset resulting from a gate applied to fs. This is the result I get if one of the frames is blank, see how all the plots are generated up until the problem plot, and the remainder aren't generated?

ADD REPLY • link 7.5 years ago shieldsy05 • 0

0

Entering edit mode

I was wrong, bc is a flowSet. Filtering upon the number of rows (aka events) should do the job.

ff.ok <- fsApply(bc, nrow) > 100  # for a better estimation
table(ff.ok)  # always report when filtering
dpl <- densityplot(IndividualID ~ `FITC.A`, bc[ff.ok])  # should be ok

ADD REPLY • link 7.5 years ago SamGG ▴ 360

0

Entering edit mode

Brilliant! This works really well and reports errors! Thanks. I really appreciate it.

ADD REPLY • link 7.5 years ago shieldsy05 • 0