Unknown
RBX1_design_40
Sequence (98 AA)
MEELKKKLEELKQEAEQLEEEAARAKAEAEALRARAAELAARLAALPESLSEERAALAAEIAALKAEAEQKEKEAERLRELKKQKEEEAEKLEKKLTG
Structure prediction unavailable
It might take a few hours to appear.
No experimental data
This protein hasn't been validated in the lab yet.
This protein was designed using TProtein language model guided de novo binder generation + novelty filtering
We developed a de novo computational pipeline to generate compact protein binders targeting RBX1 using a sequence-first strategy guided by protein language model representations and downstream structural plausibility filtering. The workflow was designed to prioritize novelty, foldability, and compatibility with the structured C-terminal RING domain of RBX1.
Candidate binders were initially generated as α-helical repeat–like scaffolds enriched for amphipathic packing patterns commonly observed in small interaction modules. Sequence space exploration was guided using protein language model embeddings to retain fold-consistent residue environments while encouraging diversity across candidates.
Generated sequences were filtered using physicochemical constraints including charge distribution, hydrophobic balance, and secondary-structure compatibility to favor soluble, monomeric binders. Redundant or low-complexity sequences were removed prior to downstream evaluation.
To enforce the competition novelty requirement, sequence similarity screening was performed against reference protein databases using profile-based homology search. Designs showing detectable similarity to known proteins below the required threshold were retained, ensuring compliance with the ≥25% edit-distance constraint relative to UniRef50-like sequence space.
Additional filtering emphasized structural compactness and interface-compatible residue organization to favor interaction with the RBX1 RING domain surface rather than nonspecific aggregation-prone contacts. Final candidates were ranked based on sequence diversity, physicochemical stability indicators, and absence of database-detectable homologs.
The resulting binder set represents a collection of short, de novo helical proteins optimized for novelty, solubility, and structural plausibility as RBX1 interaction partners.