AI Insight
This study reveals that the energy components commonly used in MM-GBSA binding free energy calculations are not independent but highly correlated with each other. Through mathematical analysis and simulations of 139 protein-protein complexes, researchers found that Coulomb interaction energy and polar solvation terms are nearly perfectly collinear (R²≥0.99), while van der Waals and nonpolar solvation terms also show strong correlation. These correlations persist across different calculation protocols and at multiple timescales, indicating fundamental redundancy in how these energy terms are typically interpreted and used.
Why it matters
The findings challenge current practices in computational drug design where these energy terms are treated as independent variables in predictive models. This has important implications for improving the accuracy and interpretability of binding affinity predictions used in drug discovery, suggesting researchers should instead combine correlated terms into grouped polar, nonpolar, and entropic contributions.
Understand the Science
⚠️ Preprint – Noch nicht peer-reviewed
Dieser Artikel wurde noch nicht von unabhängigen Experten begutachtet. Die Ergebnisse sind vorläufig und sollten mit Vorsicht interpretiert werden.
The molecular mechanics-generalized Born surface area method (MMGBSA) is one of the most commonly used end state approaches used for the calculation of the binding free energy towards computational drug design and screening studies. It is customary to break up the free energy into van der Waals, electrostatic, polar solvation (GB), and nonpolar solvation (SA) terms and then either correlate these terms with experiment or assign physical meaning to each term. Here, we demonstrate that this assumption of independent fitting coefficients for decomposed energy terms could be invalid. Through analytic derivation and large-scale molecular dynamics simulations, we show that (i) the protein and ligand Coulomb interaction energy and the GB solvation correction are almost perfectly collinear (R2[≥]0.99) reflecting their designed role as vacuum electrostatics plus solvent screening, and (ii) the van der Waals interaction and SA term likewise exhibit strong correlation, as both depend primarily on buried surface area. Interaction entropy and C2 entropy corrections are also found to be strongly dependent on underlying electrostatic fluctuations, further reinforcing redundancy. These findings hold both at the level of instantaneous trajectory fluctuations and when averaged across a diverse set of 139 protein-protein complexes and persist in both single-trajectory and three trajectory MMGBSA protocols. Our results caution against using decomposed MMGBSA terms as independent predictors in regression models and suggest instead combining correlated terms into effective polar, nonpolar, and entropic contributions. Our study provides a systematic diagnosis of collinearity in MMGBSA and highlights pathways toward more interpretable and statistically robust predictive modeling.
Source: Collinearity of Decomposed Energy Terms in MM-GBSA Binding Free Energy Calculations