Week 5: Protein Design Pt 2
SOD1 binder peptide design
Generate binders with PepMLM
The original SOD1 sequence[1] is as follows:
The A4V mutation changes the alanine to valine at codon 4, which results in
| Binder | Pseudo Perplexity |
|---|---|
| WRYPVAGLAHWE | 15.095015 |
| WRYPVAAVAHKE | 11.070548 |
| WRYPAVALRHKK | 16.016262 |
| WRYGPAALAWGE | 12.227762 |
| FLYRWLPSRRGG | - |
Evaluate binders with AlphaFold3
All of the ipTM values were higher than the known binder except for the third one, but they overall remained within a reasonable range of one another, meaning that my predicted binders may bind better than the known binder.
Evaluate properties of generated peptides in the PeptiVerse
| Binder | Solubility | Hemolysis | Binding Affinity | MW | Net Charge (pH 7) |
|---|---|---|---|---|---|
| WRYPVAGLAHWE | Soluble | Non-hemolytic | Weak binding (5.888 pKd/pKi) | 1484.7 Da | -0.15 |
| WRYPVAAVAHKE | Soluble | Non-hemolytic | Weak binding (5.298 pKd/pKi) | 1426.6 Da | 0.85 |
| WRYPAVALRHKK | Soluble | Non-hemolytic | Weak binding (5.522 pKd/pKi) | 1524.8 Da | 3.84 |
| WRYGPAALAWGE | Soluble | Non-hemolytic | Weak binding (6.051 pKd/pKi) | 1376.5 Da | -0.23 |
| FLYRWLPSRRGG | Soluble | Non-hemolytic | Weak binding (5.968 pKd/pKi) | 1507.7 Da | 2.76 |
I was surprised to see that besides the fourth generated binder, the ipTM score had an inverse relationship with the binding affinity observed with PeptiVerse. All of my binderse were once again fairly similar. For these reasons, I chose to further explore the fourth peptide.
Generate optimized peptides with moPPIt
Using moPPIt, I generated four new binders with the following properties.
| Binder | Hemolysis | Solubility | Affinity |
|---|---|---|---|
| KRDKQKKKTCYV | 0.983 | 0.917 | 7.452 |
| GGHTRTRSHTYI | 0.967 | 0.917 | 6.118 |
| KYDEKEETCKQL | 0.826 | 0.917 | 6.885 |
| KRRGRKRKKTSE | 0.966 | 1.0 | 7.122 |
These generated proteins have a higher affinity than all of the previously generated ones.




