Week 5 HW: Protein Design Part II
Part A: SOD1 Binder Peptide Design
| Peptide | Perplexity | ipTM score | N terminus | B-barrel | Dimer interface |
|---|---|---|---|---|---|
| WRYPAAAAALKX | 4.30808 | 0.3 | Close | No | Surface bound |
| WRYGATVAAHKX | 5.811953 | 0.48 | Far | No | Partially buried |
| WLSGAAALALKX | 5.716131 | 0.45 | Close | No | Surface bound |
| WLYPAAALALKX | 8.30171 | 0.36 | Far | No | Partially buried |
| FLYRWLPSRRGG | 0.38 | Far | No | Surface bound |
The predicted protein–peptide complexes produced relatively low ipTM scores overall, indicating weak confidence in the modeled interactions. The PepMLM-generated peptides showed ipTM values ranging from 0.30 to 0.48. The highest score was observed for the peptide WRYGATVAAHKX (ipTM = 0.48), followed by WLSGAAALALKX (ipTM = 0.45), both of which exceeded the ipTM score of the known SOD1-binding peptide FLYRWLPSRRGG (ipTM = 0.38). Despite these slightly higher scores, none of the predicted peptides appeared to strongly interact with the β-barrel region of SOD1, and most were either surface-bound or only partially buried on the protein surface. Overall, while some PepMLM-generated peptides showed marginally higher ipTM scores than the known binder, the predicted interactions remain weak and uncertain.
| Peptide | Predicted binding affinity | Solubility | Hemolysis probability | Net charge | Molecular weight (Da) |
|---|---|---|---|---|---|
| WRYPAAAAALKX | 5.437 | Soluble | Non - hemolitic | 1.76 | 1199.6 |
| WRYGATVAAHKX | 5.440 | Soluble | Non - hemolitic | 1.85 | 1241.6 |
| WLSGAAALALKX | 6.550 | Soluble | Non - hemolitic | 0.76 | 1082.6 |
| WLYPAAALALKX | 6.693 | Soluble | Non - hemolitic | 0.76 | 1198.7 |
| FLYRWLPSRRGG | 5.96 | Soluble | Non - hemolitic | 2.76 | 1507.7 |
The peptide property predictions were broadly favorable, since all candidates were predicted to be soluble and non-hemolytic. However, the AlphaFold3 results showed only modest ipTM values, suggesting weak to moderate confidence in the predicted protein-peptide interactions. The peptide with the highest ipTM score was WRYGATVAAHKX (0.48), while the best predicted binding affinity value was observed for WRYPAAAAALKX (5.437), indicating that higher ipTM did not perfectly correlate with stronger predicted affinity. Overall, WRYGATVAAHKX appears to offer the best balance between structural binding potential and therapeutic properties, so it would be the strongest candidate to advance.
I would choose WRYGATVAAHKX, because:
- it has the highest ipTM
- it is soluble
- it is non-hemolytic
- its charge is moderate
- it outperformed the known binder in ipTM