Accurate prediction of DNA-Intercalator binding energies: Ensemble of short molecular dynamics simulations vs. long simulations.

Despite the wide use of molecular dynamics (MD) simulations for binding energy predictions in biomolecular systems, results from single MD simulations are non-reproducible and often deviate from experimental values, even for longer simulations. This study addresses these limitations using ensemble MD simulations for the complex formation of three DNA-intercalator systems. Twenty-five replicas of short (10 ns) and long (100 ns) MD simulations were performed on the intercalation of Doxorubicin into DNA. The MM/PBSA and MM/GBSA binding energies, including entropy and deformation energy corrections, are -7.3 ± 2.0 kcal/mol and - 8.9 ± 1.6 kcal/mol, using 25 replicas of 100 ns. These values were closely reproduced even with shorter simulations of 10 ns, where the energies averaged over 25 replicas are -7.6 ± 2.4 kcal/mol (MM/PBSA) and - 8.3 ± 2.9 kcal/mol (MM/GBSA). In both cases, the energies align well with the experimental range, -7.7 ± 0.3 to -9.9 ± 0.1 kcal/mol. This shows that reproducibility and accuracy of the binding energies depend more on the number of replicas than simulation length. Bootstrap analysis revealed that 6 replicas of 100 ns or even 8 replicas of 10 ns provide a good balance between computational efficiency and accuracy within 1.0 kcal/mol from experimental values.