That is a very interesting question, and one I'd bet is on the minds of many, particularly the off modal folks. I wish there was more study on strs, their mutations, and further utility development. I kind of think that because of the issue of random mutation and signatures being rather indigenous to individual family groups that perhaps it has been an area that it is considered too much of a tiger to take on. That's just a guess. I think there are some things that could reasonably be determined given the right data. It would seem that to first determine which alleles mutate up and which mutate down would be a good start. I would think that a base would need be set, and the dna evidence extracted from archeological human artifacts to use as a base would be preferable. There are some out there; however, I could see a conflict there if it were considered sacrilegious to dig up the dead unless it were strictly an archeological find. Secondly, and unfortunately all too often the artifacts found are degraded past the point of good dna extraction -- maybe and depending. I think this field is yet so young relatively that there could be many aspects under consideration but just haven't developed yet. There seems to have been some rather substantive work done on mutation rates among various analytical computer models, but mutation direction up or down yet seems fuzzy and uncertain; and perhaps it is that way because the alleles are unpredictable of whether any given one mutates up, down or both. I think that a contradictory premise, but from a novice observer there are many things accepted which seem contradictory to me.