https://www.youtube.com/watch?v=fvzdehnJA9k
I'm trying to stay out or at least limited here, but here are some general and specific comments. Not trying to give any answers, just information for those interested.
There is a ton of bad info on the internet. It's a challenge to sort through some of it.
Stress-strain of steel (and other similar materials) is a very important and pretty well studied and understood science. Important for it to be well understood for building bridges, pressure vessels, pipelines, etc.
That science can be extended and applied to help understanding of the specifics of threaded fasteners - nuts vs. bolts, yield, proof load, torsional vs. axial stress, etc. That's where things can get a little complicated and it may seem to be purely empirical and more black magic than it really is. Thinking back to the basic materials science of stress-strain helps bridge the gap.
For example, "yield" means some of the steel plastically deforms, which is a permanent change in shape, and strengthens the material. From that point on, it can still function elastically, including returning to the previously reached deformation point repeatedly in an elastic state, with the same elastic modulus but a now higher limit before yield will occur. But pushing beyond that point has its limits, and will eventually lead to fracture.
Those were the general comments. Now for some specific stuff:
Be careful if torquing with a non-flanged nut and bolt. The BMW ones are both flanged. Due to the increased area and torquing radius vs. non-flanged fasteners, and the fact that about 50% (depends, of course on surface prep, smoothness, etc.) of tightening torque is used to overcome that surface friction, using non-flanged fasteners will provide more stretch (and I use that term without distinguishing between plastic and elastic deformation) than will flanged ones.
Similar to how using really well lubed surfaces and threads might result in excessive tightening.
So that could mean if you're using non-flanged fasteners, even on the initial installation you are over torquing the 56 Nm stage, possibly causing plastic deformation where it would otherwise not occur.
Different thread pitch
If you try using SAE fasteners or a fine thread, the bolt stretch you get from applying a 90* angle will be different. The torque will be different too, but that is less clear.
Variable nut-thread loading
That is a basic fact of threaded fasteners. It happens regardless of any plastic deformation.
Fasteners are generally designed to one one hand have the nut be more compliant than the bolt to allow for better load distribution, trying to reduce this effect.
And on the other hand, the nut is always designed to be stronger than the bolt. The reason for that is that a nut would fail by having the threads progressively strip out, vs. the bolt would just fracture. You need to worry about the first nut thread stripping out, because if it starts to strip it will pass that load down the line and all the threads will strip out. Vs. the bolt fracture would most likely happen during assembly which is not such a big problem, vs. having a nut progressively strip out would happen literally "down the road" when it would be more of a problem.
For similar reasons, you should never use a weaker nut than bolt if there is any concern for nut failure. So using a 10.9 nut on a 12.9 bolt is not something to do casually. Doing that, you will be ensuring that the nut will fail before the bolt, which is the opposite of what is safer. Here's a pretty good reference for that and other things:
Frequently Asked Questions on Bolting Matters
On the other hand, my M54 engine uses Class 12.9 head bolts, which do yield along their whole 110 mm length, threaded into an aluminum engine block ... or during a typical head gasket repair threaded into a ~15-thread long steel insert, which is probably not Class 12.9. So if it's designed carefully, it can be made to work.
When terms like plastic deformation, yield, proof, ultimate, etc. are used, they sometimes refer to total system failure (like "ultimate tensile strength" tells you when the bolt will snap [system failure] when increasing load is applied).
And sometimes they refer to localized material effects - where for example, the first thread in a nut may plastically deform very slightly - and it's not a problem at all, but it did deform. BTW, by the time the nut threads noticeably plastically deform, the bolt will have deformed a lot more.
True "stretch bolts" or TTY bolts - by which I mean bolts that are made specially for a TTY application, rather than regular bolts that are used in a TTA/TTY(?) application (as we have here) - are typically designed with relatively oversized heads and threaded sections, but with a long, thinner, uniform portion that is designed to yield, and to have all the yield there, and none anywhere else. Those are easier to analyze.
On figuring out whether your previously used nuts and bolts have deformed enough to significantly affect reinstallation by requiring more torque to overcome the bad fit caused the deformation ...
One pretty simple test is to see if you can finger-thread the nut onto the bolt. If you can do that, your torque wrench will probably not be fighting too hard to overcome that, and the measured tightening torque will have the same clamping effect as when they were initially used.