You are here

Questions on energy oddities

3 posts / 0 new
Last post
Questions on energy oddities
#1

Hi,

I am doing side-chain packing on a fixed backbone over a discrete set of configurations. Given a set of (non-rotamer) side-chain configurations I'm using Rosetta to calculate one-body and two-body energies using the standard scoring function. I observe to things that seem strange in the two-body energies:

1) For a pair of residues sometimes the two-body energies are almost all zero except for a couple configurations. This sort of makes sense if the residues are far enough apart, but does it make sense that they are *exactly* zero?

2) For some configurations the two-body energy depends on only one of the residues and is constant w.r.t. the other. I've observed this in a small protein between Lysine and Tyrosine, but I could probably find other examples. In this case the two-body energy only varies with the Lysine side-chain.

Do these conditions sound normal or should I suspect a bug somewhere in my code?

Jason

Post Situation: 
Wed, 2014-10-15 13:26
pachecoj

To make things such as rotamer packing efficient, most terms of the standard Rosetta energy function have a maximum distance above which they have non-zero energies. Residues pairs which don't have any atoms within that cutoff have zero interaction energy - by definition. (The functional form of the energy terms are adjusted such that the energy and the derivative smoothly transition to zero - there isn't any sharp transition point.) The exact cutoff is term dependent, but normally is sitting somewhere around 6 Ang.

That's probably what you're seeing. The rotamer pair which you're observing is probably just at the edge of that cutoff, so some conformers dip within it and have a small, non-zero energy, whereas the others stay totally outside it, resulting in an exactly zero energy.

For 2), that's the behavior I would expect for interaction between the lysine side chain and the tyrosine backbone. If the tyrosine sidechain is entirely outside the cutoff, changing the atom positions there wouldn't change the energy at all. But changing the lysine sidechain would change how it interacts with the backbone, and you would see an energy change.

Wed, 2014-10-15 13:36
rmoretti

Thanks for the quick response! This all makes perfect sense.

Jason

Wed, 2014-10-15 15:12
pachecoj