search

farrellja / URD

URD - Reconstruction of Branching Developmental Trajectories
GNU General Public License v3.0
118 stars 41 forks source link

Weird behavior from buildTree #52

Closed ZhaoxiangSimonCai closed 4 years ago

ZhaoxiangSimonCai commented 4 years ago

Hi @farrellja , I have been using URD for a while but recently I'm having trouble with the buildTree function. It seems no matter if I take a previous processed axial.rds object, or start from the beginning, the tree layout looks like just a mess. More specifically, even if I take a previous axial.rds after processRandomWalksFromTips, then call buildTree exactly the same as before (I kept all my code), I still get a messed tree.

I noticed that now I get some warnings from ggplot such as: Removed 760 rows containing missing values (geom_point). Removed 8 rows containing missing values (geom_segment).

If I directly load a previous axial.tree.rds, then plotTree is still able to get the expected tree layout. I have also tried to uninstall and reinstall the package, but the problem still exists. I suspect it could be due to some dependant that got updated but completely have no better idea.

It would be great if you can shed some light on this! I have also attached the two layouts that are supposed to be the same. Thanks!

image image

farrellja commented 4 years ago

Hi @Olimon660 ,

I think this is related to the ggraph package -- there seems to be a major change in behavior of one of the dendrogram functions I used from it between versions 1 and 2. See the discussion in Issue #45 and let me know if that helps you. If not, we can revisit :)

ZhaoxiangSimonCai commented 4 years ago

Hi @Olimon660 ,

I think this is related to the ggraph package -- there seems to be a major change in behavior of one of the dendrogram functions I used from it between versions 1 and 2. See the discussion in Issue #45 and let me know if that helps you. If not, we can revisit :)

solved it like magic... massive thanks!!