Gabe Rocklin, formerly at Baker lab (birthplace of Foldit), has a new preprint out about what factors contribute to stability in mini protein designs. Foldit designs are typically a little bigger than these mini proteins, but the pointers might still be helpful for Foldit players.
Here are some highlights from Dr. Rocklin's twitter thread about the new work:
"One buried unsatisfied polar atom is almost as damaging as one hydrophobic residue is favorable."
"Designs worked better with common loop types, longer helices, and shorter strands."
"So, how do you design αββα miniproteins? Make sure they are hydrophobic enough, use good loop types and long helices, and make sure the helices have nonpolar residues near the termini."