Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

missing overlaps using eulerr #93

Open
dgcovell opened this issue Mar 7, 2022 · 3 comments
Open

missing overlaps using eulerr #93

dgcovell opened this issue Mar 7, 2022 · 3 comments

Comments

@dgcovell
Copy link

dgcovell commented Mar 7, 2022

Not sure whether I am using this tool properly, but here is what I am getting:

list_euler <- c("agc" = 9, "camk" = 17, "cmgc" = 16, "tk" = 16, "tkl" = 23,
"agc&camk" = 1,
"camk$tk" = 1,
"tk&tkl" = 1,
"camk&cmgc&tkl" = 1,
"camk&tk&tkl" = 2,
"agc&camk&tk&tkl" = 1,
"camk&cmgc&tk&tkl" = 3,
"agc&camk&cmgc&tk&tkl" = 1)
fit <- euler(list_euler, shape="ellipse")

fit (zeros deleted)

agc 9 9 0 0.012
camk 17 17 0 0.023
cmgc 16 16 0 0.021
tk 16 16 0 0.021
tkl 23 23 0 0.030
camk$tk 1 0 1 0.011
agc&camk 1 0 1 0.011
camk&cmgc&tkl 1 0 1 0.011
camk&tk&tkl 2 0 2 0.022
agc&camk&tk&tkl 1 0 1 0.011
camk&cmgc&tk&tkl 3 0 3 0.033
agc&camk&cmgc&tk&tkl 1 0 1 0.011

diagError: 0.033
stress: 0.013

The plot generates 5 non-overlapping ellipses. I see from other comments that small numbers are
excluded. If I have made a mistake in entering the input, please let me know. If there is a way to
display overlaps I would be thankful.
Note that the numbers are taken directly from UpSetR.

Regards,
David Covell, Ph.D.
NIH,NCI

@jolars
Copy link
Owner

jolars commented Mar 7, 2022

There are no guarantees that the resulting fit is perfect. Please see #80, #54, etc.

@jolars jolars closed this as completed Mar 7, 2022
@jolars
Copy link
Owner

jolars commented Mar 7, 2022

If I am to guess, the problem with this example is that the intersections are so small relative to the non-intersecting parts, so it becomes very difficult to find a configuration with these intersections but without the non-existing ones.

@jolars
Copy link
Owner

jolars commented Mar 12, 2022

I've looked into this a bit more and there actually seems to be an inconsistency here, because the fit is often worse using ellipses than it is with circles, which shouldn't happen. I'll reopen this issue and see if I can figure out what's going on.

@jolars jolars reopened this Mar 12, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

2 participants