Does the dark matter halo rotate with the galaxy? If the dark matter halo is stationary related to the arms of the galaxy then tidal effects should slow the galaxy rotation.
If it rotates with the normal matter in the galaxy then shouldn't it flatten out into a disk?
 A: It would probably be more proper to say that the galaxy rotates with the dark matter halo, since the mass of the halo is greater than the baryonic mass of the galaxy that we observe. Generally, dark matter halos have triaxial shapes, and shortest axis of the visible part of the galaxy will also be shortest axis of the dark matter halo. 
EDIT: I should have been rather more careful. My answer is generally correct, I think, but there may be some more variation than I thought. A paper looking at spin of dark matter halos in the presence of baryons in the Millennium simulation sees a median misalignment between halo and baryonic (visible) galaxy of 30 degrees, and quite a total distribution. They further note that this misalignment seen in simulations will complicate mapping of dark matter halos with lensing measurements. 
A: The deduction of Dark Matter is caused by the wrong idea that Kepler's laws would be valid in disc galaxies. In reality however, Kepler is only valid for one large central mass and a tiny orbiting one.
As LeVerrier has proven for our 8 planets, an additional planet will alter that orbit by a precession due to their gravity. For each added planet, all the orbits of the other planets will be altered! This results in Mercury's Newtonian part of its perihelion precession.
However, in disc galaxies, there are many millions of stars that alter each-other's orbits many million times, resulting in orbit speeds that became totally different from Kepler's. 
The only way to do it right is to integrate a plausible mass distribution according to Newton's rules of integration, and to calculate the gravity upon a given orbit. 
The consequence of that calculus is that the velocity curves of disc galaxies become indeed very flat outside the bulge, as observed. 
This is also the reason why no need for Dark Matter was found in the DF2 galaxy cluster. In the case of DF2, the small number of globular clusters in the galaxy are more comparable to our planetary system, which indeed doesn't need any presence of Dark Matter, because the mutual influence of the few cluster's orbits is very limited. 
However, the mainstream theory advances that disc galaxies are not stable with stars alone, and that the need of Dark Matter is in any case legitimate, even if the velocity curves are solved. They conclude this because computer simulations result to that.
In order to solve this issue, we need to come to the insight that moving masses are inducing the surroundings by a second gravitational field, just like moving electrons cause induced magnetic fields and Lorentz forces to other moving charges.
This gravitational effect, caused by the velocity of masses, occurs in the bulge of disc galaxies by many fast spinning stars and (pseudo-) black holes, which generate a global gravitational angular momentum in the bulge, transmitted by gravity to the surroundings.
This is not a fancy theory or hypothesis. It has been proven by the Gravity Probe B experiment, which found that the orbital motion and the spin of the Earth cause gyroscopes' axes in satellites to be altered.
The proof of this behavior can be seen in bar galaxies: the bar is in fact caused by a re-oriented gravitational angular momentum of the bulge, after important collapses of stars, which have dramatically increased the spin rates. Hence, the bar is in fact an inclined part of the disc's central part in formation. The gravitational angular momentum will further be transmitted to the rest of the disc, and the whole disc will eventually be re-oriented.
With galaxy clusters, a similar problem arized with the wrong use of the virial theorem. This theorem is applicable if the side with the virial can be set to zero. In the present case, it would be so if either the system is fully cyclic or if the average velocity can be found if the system can be extended to an infinite time.
None of the cases are valid for galaxy clusters, but it is assumed so by the mainstream, and this gives indeed a wrong use of the virial theorem.
Since the virial theorem is used to determine the mass of the system, this estimation will be totally wrong.
With best regards,
Thierry De Mees
http://gsjournal.net/books/De-Mees-Gravitomagnetism-and-Coriolis-Gravity-2011-A4.pdf
http://gsjournal.net/Science-Journals-Papers/Author/58/Thierry,%20De%20Mees
