Has anyone had an issue with flipping of eigenvectors (e.g. before the A compartments were in one direction, and now they are flipped with the B compartments)?? and if so -- could anyone help trouble shoot this with me? I have tried a few things but nothing is working and i am not sure why it is happening. I am using previously published data.

That's expected. There is no order in the eigenvectors from the PCA. Basically, you have to assign A and B, using additional features like GC content, genes, etc.