Sequence (91 AA)
GLCNAWATLGKELKKKGYQELAAILNWFALGNRYFPLLQEKLRERLTELGLPAEAAALVAAMEACDVAAAGAALAAIAARLGLAEAAALAA
No experimental data
This protein hasn't been validated in the lab yet.
This protein was designed using PRBoltzGen → PyRosetta → GlycoShape → BioEmu-MD
Generative design: BoltzGen Use BoltzGen to generate candidate miniprotein sequences + 3D atomic models. We specify design constraints appropriate for our target: binding-site residues. We run the “protein-anything”, requesting a large initial set of designs (e.g. tens of thousands) to ensure broad sampling.
Physical scoring & packing filter: PyRosetta (or Rosetta-based scoring) For each generated candidate, compute predicted folding free energy (ΔΔG), analyze residue–residue contacts, core packing quality, and solvent-accessible surface exposure. Discard designs exhibiting unfavorable energetics, poor packing, exposed hydrophobic patches, or other red flags that may cause misfolding or aggregation. This filter ensures that only biophysically plausible designs pass to the next stage.
Glycan-aware binding/context check: GlycoShape For targets that are glycoproteins (or predicted to be glycosylated), restore plausible glycan chains at known or predicted glycosylation sites using GlycoShape’s glycan-structure database (500+ unique glycans derived from MD sampling) and its Re-Glyco algorithm to graft them onto the target structure. Re-evaluate each miniprotein-target binding pose in the presence of glycans: discard any designs that would clash sterically, have binding interfaces blocked, or require unrealistic conformational rearrangements to accommodate glycan shielding. This step ensures compatibility with the physiological, glycosylated form of the target, a frequent source of failure in binder design.
Dynamic stability validation: BioEmu For top-ranked designs from stage 3, check the stability assessment of the miniprotein (alone). Generate equilibrium ensembles and perform trajectory analyses: Backbone RMSD over time, to assess whether the protein remains folded and close to the design. Per-residue RMSF, to identify flexible or unstable regions that may indicate partial unfolding or poor packing. Radius of gyration (Rg), to monitor overall compactness and fold integrity. Principal Component Analysis (PCA) of coordinates, to capture dominant collective motions and detect potential large-scale drift or unfolding events. Only designs that remain structurally stable (low RMSD drift, reasonable RMSF, stable Rg, no major PCA-detected unfolding) over the simulated timescale are selected as final candidates.
Final selection & export for experimental testing: Custom ranking including ipSAE score From the stable designs, compute ipSAE on predicted complex models to assess binding-interface quality, along with classical and dynamic stability metrics. Use ipSAE to rank binding-propensity, favoring designs with high ipSAE (e.g. ipSAE > ~0.6, or among top-scoring subset), combined with good BioEmu-stability and packing. Select a final, diverse set of miniproteins balancing binding confidence, stability, and structural diversity for synthesis and experimental validation.