"High overlap" despite mutually exclusive conditions?

[Eric]

I have defined a few cell types like this:

>CD4+ Effector T

expressed: CD4, CD3E, CD3D, CD3G, B3GAT1 

not expressed: NCAM1

>CD4+ Effector T1

expressed: IL2RA 

not expressed: FOXP3, NCAM1

subtype of: CD4+ Effector T

>CD4+ Effector T2

expressed: FOXP3

not expressed: IL2RA, NCAM1 

subtype of: CD4+ Effector T

When I run check_markers, I see this result:

I am confused that there is "High overlap" for FOXP3, given that this gene is specified as "not expressed" in the T1 cells, but "expressed" in the T2 cells. So naively I would expect no overlap to be possible. Can you help me understand this situation properly?

Thank you!

Eric

[Hannah Pliner]

Hi Eric,

Sorry for the delayed reply. I'm thinking you may have found a bug but I'm having a hard time reproducing it. Can you try running the following code, where "~/Desktop/test.txt" is your exact marker file from above and let me know what you get?

library(garnett)
library(org.Hs.eg.db)
 data(test_cds)

marker_check <- check_markers(test_cds,

                              "~/Desktop/test.txt",

                              db=org.Hs.eg.db,

                              cds_gene_id_type = "SYMBOL",

                              marker_file_gene_id_type = "SYMBOL")
plot_markers(marker_check)

Best,

Hannah

[Robert Alpin]

I've encountered a similar issue to Eric's, even after making sure my script lines up with Hannah's.

[Hannah Pliner]

Hi Robert, 

Can you try running this code: 

library(garnett)
library(org.Hs.eg.db)
 data(test_cds)

marker_check <- check_markers(test_cds,
                              "~/Desktop/test.txt",
                              db=org.Hs.eg.db,
                              cds_gene_id_type = "SYMBOL",
                              marker_file_gene_id_type = "SYMBOL")
plot_markers(marker_check)

But with this test marker file? 

>CD4+ Effector T

expressed: CD4, CD3E, CD3D, CD3G, B3GAT1

not expressed: NCAM1

>CD4+ Effector T1

expressed: IL2RA

not expressed: FOXP3, NCAM1

subtype of: CD4+ Effector T

>CD4+ Effector T2

expressed: FOXP3

not expressed: IL2RA, NCAM1

subtype of: CD4+ Effector T

And post the marker plot output?

[Robert Alpin]

Running it with the pre-downloaded data set and those terms behaves as expected:

I'm not sure why I get this issue with my own data set, however.