The Milky Way's Formation History

The observed differences between the actual makeup of the Milky Way and what is predicted by simulations is more or less exactly what MOND would predict. 

The early phases of galaxy formation are in the Newtonian regime so they behave as a Newtonian model predicts that they would. 

But once the galaxy reaches a point where some parts of it are in the MOND regime, the pull of gravity binding satellites to it gets stronger than predicted by a Newtonian regime, so dwarf galaxies at the galactic fringe look like recently acquired outlier galaxies in a Newtonian model.

Galactic halos are known to grow hierarchically, inside out. This implies a correlation between the infall lookback time of satellites and their binding energy. In fact, cosmological simulations predict a linear relation between infall lookback time and log binding energy, with a small scatter. 

Gaia measurements of the bulk proper motions of globular clusters and dwarf satellites of the Milky Way are sufficiently accurate to establish the kinetic energies of these systems. Assuming the gravitational potential of the Milky Way, we can deduce the binding energies of the dwarf satellites, as well as of the galaxies previously accreted by the Milky Way, which can, for the first time, be compared to cosmological simulations. 

We find that the infall lookback time vs. binding energy relation found in a cosmological simulation matches that for the early accretion events, once the simulated MW total mass within 21 kpc is rescaled to 2*10^11 solar masses, in good agreement with previous estimates from globular cluster kinematics and from the rotation curve. However, the vast majority of the dwarf galaxies are clear outliers to this re-scaled relation, unless they are very recent infallers. In other words, the very low binding energies of most dwarf galaxies compared to Sgr and previous accreted galaxies suggests that most of them have been accreted much later than 8 or even 5 Gyr ago. We also find that some cosmological simulations show too dynamically hot sub-halo systems when compared to identified MW substructures, leading to overestimate the impact of satellites on the Galaxy rotation curve.

F. Hammer, et al., "The Milky Way accretion history compared to cosmological simulations -- from bulge to dwarf galaxy infall" arXiv:2411.07281 (November 11, 2024).

Another new paper likewise finds a strong correlation between galaxy mass as estimated by lensing and galaxy shape. This is something that Deur's work predictions, but which is not found in standard Newtonian gravity with dark matter type simulations.