So far as I know, the orientations of the spin axis of galaxies within clusters, and so the inclinations of their flattened disks, is pretty random. In the Hubble Deep Field (above image) which shows hundreds of galaxies, it looks like many of them are aligned in certain favored orientations. But this may only be a statistical fluke since this is just one field, and many of the galaxies do not have well-defined shapes. A careful study using various angular 'bins' would have to be conducted to see if the distribution is truly random or significant given the small number of galaxies used. For example, if you looked at 100 galaxies and binned them in 10 degree increments from 0 to 90 degrees relative to an arbitrary horizontal plane, the statistical uncertainty per bin could be +/- 10 galaxies just from the sampling error alone. It would be hard to argue for the significance of such a result given random sampling errors this large.