 So for secondary structure elements in general, there will be some sort of repeating unit or pattern, right? Let's see if we can understand those patterns. In globular proteins we could argue that this pattern is just, do you prefer to be an alpha helix or not? Or do you prefer to be a beta sheet or not? In membrane proteins the obvious pattern is, are you hydrophobic? That is in the membrane or hydrophilic in the loops. And for fibrous proteins, well we're not really sure what the pattern there is, but the point is that there is some sort of short unit that we're then repeating many times. If we're looking at loops we have kind of the same concepts here. We have some very common patterns that if you're having an alpha helix that then crosses over to another alpha helix. Remember that I said that alpha helixes were usually anti-parallel? Technically, formally you can of course have two parallel helix, but then you need to have this long interconnect that shouldn't be an alpha helix. Again, technically possible, but that's going to be expensive and rare. If you're having two loops next to each other, I'm going to argue that they will always look as they are on the left here. Formally you could have a crossover like on the right, but that defect is going to be very costly. So costly that in practice you won't see it. And that's a bit strange, because if you have a gigantic ribosome or whatever, these defects we're talking about costs in the order of maybe losing one or two hydrogen bonds. That seriously can't matter compared to an entire protein, right? You have thousands or tens of thousands of interactions in a protein. And yet when you just end up doing a small mistake like missing two hydrogen bonds, boom, the protein goes. You will not see that fold spontaneously. That's amazingly interesting, although we don't understand it yet. Some small fraction of the total energy in a protein appears to be stable for its entire stabilization. And that's going to have profound implications for protein folding.