At minimum, to the extent that mutations in DNA affect gene expression variability, this absolutely accounted for in population genetics. The fact that some mutations cause their effect by changing expression is already part of population genetics. One important fact is that gene regulatory sequences can function like “dimmer switches” that do not solidly turn on or off a gene when they are mutated. That makes it safer to build up variation this way.
Here, when I say “mutations”, I’m not meaning merely point mutations, but a whole host of different mutations (e.g. transposon jumping, copy number variation, structural variation), all of which can dim in or dim out expression of genes at different point of development. This is all included in the theory, but not always modeled in practice. Part of the reason why is that we do not have enough knowledge of some of the particulars yet (e.g. de novo mutation rate). Give it time though. It is already part of the theory. Soon enough there will be software out there to model it with increasing detail.