Late last year, a study revealed that the amount of ordinary matter in ellipitical galaxies has been greatly underestimated. The implication was that the amount of dark matter inferred from lensing effects and galactic rotation curves, adjusted for mass attributable to luminous matter, was greatly overestimated. Rather than being much more common than dark matter, there were about equal amounts of visible and dark matter in the universe, as a result of this data point alone.
But, this isn't the only development casting doubt on the dark matter paradigm. A number of physicists, most prominently F.I. Cooperstock and S. Tieu at the University of Victoria in British Columbia, in a series of papers including this one, but also C. F. Gallo and James Q. Feng (outside academia), and German academic scientists Aleksandar Rakic and Dominik J. Schwarz, have suggested that traditional estimates of the amount of dark matter necessary to produce observed galactic dynamics based on a Newtonian gravitational paradigm are overestimates that ignore significant relativistic effects in rotating galaxies.
Researchers like Tobias Zingg, Andreas Aste, and Dirk Trautmann have criticized the models of Cooperstock and Tieu, as have others (to which they in turn have prepared responses), but acknowledge that the models that have been used to data to estimate dark matter amounts and distributions in galaxies have been inadequate, as a result of their failure to consider relativistic effects, to provide sound answers either. A 2009 paper by H. Balasin and D. Grumiller suggests that properly considering relativistic effects reduces the amount of dark matter necessary to explain galactic rotation curves by 30%.
Meanwhile, the cold dark matter paradigm has proven inadequate to fit large scale galactic structure observations and requires a cuspy halo distribution for dark matter that does not arise naturally with any of the prevailing cold dark matter candidates (for example, WIMPs). Direct WIMP (weakly interacting massive particle) searches like Xenon 100 are also increasingly developing the experimental power to rule out the existence of WIMPS with a wide variety of proposed properties.
This doesn't mean that the dark matter paradigm is dead. Observations like the bullet cluster collision have ruled out simple versions of modified gravity theories in which gravity tracks luminous matter, but have also put limits on the cross-section of interaction for dark matter that rule out a typical cold dark matter hypothesis. Theorists seem to be migrating to a "warm dark matter" hypothesis, in which dark matter is both weakly interacting and moving at marginally relativistic speeds, or mixed dark matter models, but still have no good candidate particles that have been shown to really exist.
But, I have yet to see any papers that really integrate these new developments into a comprehensive whole. If the amount of ordinary matter in the universe was previously greatly underestimated, and the amount of dark matter necessary to produce observed galactic rotation effects in galaxies has been underestimated by some amount due to a failure to consider significant relativistic effects in galactic dynamics, then ordinary matter must actually account for most of the matter in the galaxy rather than a mere minority portion of it, as is commonly asserted, even before anyone embarks on new physics. Moreover, the residual distribution of dark matter given these findings should be quite different than in "old school" cold dark matter models that are starting to lose credibility.
More refined estimates of how much dark matter is out there, how it is distributed, and what its cross-section of interaction must be from multiple sources, in turn, will narrow the parameter space for potential dark matter candidates (which are also, one by one, being ruled out by LHC results). And, shrinking estimates of the percentage of matter that is dark, particularly as they fall below 50%, are increasingly hard to fit to theories that naturally include stable, very weakly interacting particles that aren't too massive individually, which are dark matter candidates.
Yet another part of this stew is a conclusion based on gamma ray astronomy observations that predictions that there might be quantum effects at the Planck scale, reflecting quanta of length, have not appeared, suggesting that many string theory based and loop quantum gravity based theories of quantum gravity may be flawed. Polarization effects expected if there were quanta of gravity in very high energy gamma ray bursts were not observed (citing P. Laurent, D. Götz, P. Binétruy, S. Covino, A. Fernandez-Soto. Constraints on Lorentz Invariance Violation using integral/IBIS observations of GRB041219A. Physical Review D, 2011; 83 (12) DOI: 10.1103/PhysRevD.83.121301). This impacts dark matter theories because gravity modifications relative to general relativity in the weak field, if they exist, would likely be a result of quantum gravity effects associated with quanta of length in the fabric of space-time. Continuity at a level below the Planck scale cases doubt on the entire concept of discrete units of space-time because there is no other natural scale at which such discontinuities should exist.