For this protein-protein interaction prediction challenge, we present three target proteins: KRAS, RAF (the KRAS binding domain), and Mdm2. For each target protein, we are providing you a list of 8 potential proteins that may or may not bind the target, all built from the same protein scaffold. We are giving you the information that exactly 3 of the binders for each target do not have detectable binding to the target (they are red herrings), and 5 of the potential binders bind the target, with varying affinities from pM to nM to uM (affinities provided, measured by SPR
in vitro).
So the challenge is two-fold:
#1 Identify which are the non-binders (three for each target)
#2 Rank order the real binders by affinity
That's it! Submit your answers below and we will contact you if you get them correct! Email dickinson@uchicago.edu with any questions. Otherwise, we will post an answer key sometime later on. Or, if no one is able to get it, maybe we will just leave this challenge out there like the chalkboard in Good Will Hunting, waiting for the savior of protein interaction prediction to emerge.
