I'm very very curious why you would decide to limit your explore options Brow42? And also define your 'discard' options?

....and actually how you have decided how to define that 'margin of error'?

I'm just curious :)

CFC

Mainly, it started out with frustration at the mutate sidechains/mutate combo/mutate all turning everything into Pro or Gly if I moved the ligand too close, and I'd have to undo or mutate by hand to get polar sidechains back, plus I'd lost 20 minutes to brute force mutate just to get 2 gly and a pro.

Lately, I've been thinking about a mutate function that was configurable...pass in any combination of: all hydrophobe (except pro, gly), all hydrophilic, MALEK, , pro, and gly, or a custom list.

As for the TNT puzzles...before my time.

Obviously, this can all be done in pure Lua, but I suppose the C implementation might do something smarter, like pick-random-until-improvement or something, and be significantly faster at that.

I just wanted to perma-cross-link this feedback here: http://fold.it/portal/node/992394 as another example. I think the docking puzzles have stimulated a lot of new ideas. That feedback notes a preference for aromatic sidechains at many interfaces.