Im currently constructing trajectories based on DE genes between a basal state cell and the subsequent phenotype that emerges after injury. I know this change to be true in vivo and in vitro, so i know that my trajectories should be running basal -> injured. But they are consistently backwards, with the injured cells on the left trajectory graphs at components -4, -2 , 0 on the x axis.
I have tried the use the
object <- orderCells(object,reverse=T) argument, but this seems to have worked for others only for simple trajectories with no branches, and has no effect on mine.
Similarly, assigning a specific states as a root state argument in
orderCells does nothing either. I have reordered the comparison in the DE test ( so basal Vs injured, rather than injured Vs basal ) in case it was some sort of non commutative situation, but that has had no effect either.
A line of code suggested on the monocle forum has no effect when i run and then replot the trajectory
pData(object)$Pseudotime <- max(pData(object)$Pseudotime) - pData(object)$Pseudotime # to reverse
Has anybody encountered and solved this issue?
Given the main issue is the directionality of images and DE gene lists, would it be improper to just flip the trajectory in image editing software ( and disclose such obviously) and just account for this change when interpreting gene lists?