For several decades, the striking contradiction between the Huang diffuse scattering experiments, resistivity recovery data, and predictions derived from density functional theory (DFT) remained one of the mysteries of defect physics in molybdenum. Since the nineteen seventies, observations of Huang X-ray diffuse scattering appeared to indicate that a self-interstitial atom (SIA) defect in Mo adopts a $langle 110rangle$ dumbbell configuration. However, the low temperature defect diffusion data supported the DFT prediction of a different, highly mobile $langle 111rangle$ SIA defect structure in the same metal. Using symmetry-unconstrained DFT simulations, we show that an SIA adopts a symmetry-broken configuration in all the group 6 metals: chromium, molybdenum and tungsten. The symmetry-broken defect structure, a $langle 11xirangle$ dumbbell, where $xi$ is an irrational number, agrees with nudged elastic band analyses of $langle 110rangle$  to $langle 111rangle$ transformations. Direct simulations of Huang diffuse scattering by symmetry-broken defect configurations predicted by DFT explain why no zero intensity lines were observed in experiment and resolve the long outstanding question about the structure of defects in Mo and similar metals. A $langle 11xirangle$ defect migrates {it on average} one-dimensionally through a sequence of three-dimensional non-planar [11$xi$] to [$xi$11] or [1$xi$1] transitions. Barriers for defect migration in non-magnetic Cr, anti-ferromagnetic Cr, Mo and W derived from DFT calculations, 0.052, 0.075, 0.064 and 0.040 eV are well correlated with the onset of defect migration temperatures observed experimentally.