Comparison to Generalized Born
| Peptide sequence | This study | GB |
| …NNKKAAAAAAAAAAAAAAAAAAAKKNN… | 0.0 | 0.0 |
| …NNKKAAAAAAAAARAAAAAAAAAKKNN… | −4.7 | 26.3 |
| …NNKKAAAAAAAARRAAAAAAAAAKKNN… | 3.3 | 28.0 |
| …NNKKAAAAAAARRRAAAAAAAAAKKNN… | 0.7 | 22.9 |
| Peptide sequence | This study | GB |
| …NNKKAAAAAAAAAAAAAAAAAAAKKNN… | 0.0 | 0.0 |
| …NNKKAAAAAAAAARAAAAAAAAAKKNN… | −4.7 | 26.3 |
| …NNKKAAAAAAAARRAAAAAAAAAKKNN… | 3.3 | 28.0 |
| …NNKKAAAAAAARRRAAAAAAAAAKKNN… | 0.7 | 22.9 |
Comparison of insertion energy values between our method and a Generalized Born method (Im et al., 2003a). The electrostatic energy for the four sequences on the left was computed using both methods. Reported energy values are the difference between the current sequence and the R-1 containing helix, with the alanine sequence being set to zero. All energy values are in kcal/mol. While our method allows the membrane to bend to expose charged arginines, the Generalized Born method treats the membrane as a flat slab with a dielectric that smoothly switches from membrane to solvent. Thus, the Generalized Born method predicts an electrostatic cost that is nearly equal for each additional arginine, while our method does not exhibit additivity.