Recently, a group from the Baker lab published a paper in Science describing a high-throughput way to test lots of protein designs at once. That paper is now available, but only though a specific link on the Baker lab website. To view the paper, head to the Baker lab site and click on second item listed (Global analysis of protein folding using massively parallel design, synthesis, and testing).

Traditionally, if you wanted to measure a protein’s structural stability, you'd first have to purify the protein (which is expensive and time-intensive) and then run some kind of protein unfolding experiment with your purified protein sample.

The authors of this paper developed a protease assay that can approximate the stability of thousands of proteins in parallel, without having to purify and test each individual protein.

The idea is that unfolded or unstable proteins will be readily chewed up by proteases; but stable, well-folded proteins will resist protease degradation and remain intact. We can then sort out which proteins survived the proteases.

From this large dataset of protein stability, the authors were also able to make some interesting conclusions about protein design. For example, they found that ASP and GLU are especially stable in the first turn of a helix; ARG and LYS are preferred in the last turn.

This experiment was made possible by cheap, high-throughput DNA synthesis. Currently, this type of DNA synthesis is limited to short stretches of DNA that can only encode proteins of <60 residues (still too small for most Foldit-designed proteins). However, this technology is getting better all the time, so we may soon be able to apply this method to test thousands of Foldit-designed proteins!