BINMAN
bridging ligands & neosubstrates Selection state Nothing pinned yet. GLUE E3 TGT

Degron Scan

The β-hairpin-with-exposed-glycine geometry that underlies CRBN neosubstrate recognition, found by geometry rather than by sequence motif. This is a geometric filter and its score is a rank, not a calibrated probability.

21,717
candidates
8,983
proteins
21,717
high confidence
tip pLDDT ≥ 70
0
known neosubstrates
Two scores, two questions. Both are shown; neither replaces the other.
degron_geometry_score hairpin geometry. Measured against the Sievers screen at AUC 0.44, below chance: a C2H2 zinc finger is a hairpin with an exposed glycine, so this largely detects the fold. A hypothesis generator, not a classifier.
imid_degradation_score predicted probability of degradation by pomalidomide. Nested held-out AUC 0.830, sensitivity 0.867 at specificity 0.613, which clears the spec 9.2 floors. Empty for 17,067 of 21,717 rows by design: those carry no zinc finger.
Degradability is compound-specific, which is a finding rather than a limitation: the same features predict lenalidomide well and the promiscuous analogs barely at all. Contingency tables and the threshold sweep are on the About tab.
nothing pinned
Pin a degron candidate to load its AlphaFold model. The cartoon is coloured by pLDDT on the standard four-band scale, the candidate hairpin is highlighted and the tip glycine renders as spheres.