The Ring Box Protein 1 is a part of the Cullin-RING E3 Ubiquitin ligase (CRL) complex. The complex controls degradation of numerous cellular proteins. To design binders specifically against RBX1 we follow four strategies: Free design on only the RBX1 protein not in complex from the PDB 2LGV. Epitope specific design on only the RBX1 protein not in complex from the PDB 2LGV. The epitope will be chosen such that it is exposed in the primary CRL complex. Free design on the RBX1-Cullin 1 complex from the PDB 7Z8R. The epitope will be chosen such that it is exposed in the primary CRL complex. Epitope specific design on the RBX-1 Cullin 1 complex from the PDB 7Z8R.

Initial runs that used the full structure from 2LGV, failed on RMSD-based metrics — especially the target-aligned RMSD. We realized that the rather long disordered region (at the N-terminus) could be tripping up the structure prediction model and precluding the prediction of structures in the context of this region. We hypothesized that truncating the structure to remove the disordered region will lead to better results.

For design, we use the mosaic (escalante-bio) framework adorned with some more losses to hallucinate minibinders. These minibinders were then subject to structure prediction via the Protenix model. We evaluate the quality of the designed binder based on a number of metrics such as the iptm, min-ipae, mean-ipae, binder-rmsd and target-aligned-binder-rmsd. Structures predicted by Protenix that pass thresholds set by us for these metrics are submitted to the competition.