Lot's of people have spent lots of decades looking for nonrandom variation without finding it. I've seen FReepers argue that most mutations are harmful, but then the chances against a germ cell achieving conception are a hundred million to one. There's an enormous selection factor prior to birth, much greater than after birth.
I just wanted to mention one thing concerning your remark about the inability to find evidence of nonrandom variation. You might find the interviews and discoveries by Gehring concerning master control genes - in his case the evolution of eyeness across phyla - to be illuminating. Ditto for Peter Weiss' commentary on "how the eye got its brain". More later...