Molecular dynamics simulation of DNA translocation through solid-state nanopores

Abstract

This thesis explores the use of molecular dynamics (MD) simulations to model double-stranded DNA (dsDNA) translocation through nanopores, aiming to optimize nanopore-based DNA sequencing technologies. Using both LAMMPS and NAMD, the study investigates the potential for MD simulations to provide insights into nanopore sequencing mechanisms and system behaviors. Insights gained from LAMMPS tutorials by Simon Gravelle informed the understanding of force fields, system minimization, heating, and equilibrating processes necessary for accurate MD simulations. The atom pull method presented in these tutorials was foundational in understanding force application within molecular systems, analogous to grid-steered molecular dynamics (G-SMD) techniques later applied to DNA sequencing.Following these foundational steps, NAMD simulations based on protocols by Jeffrey R. Comer were performed to simulate dsDNA translocation through two types of nanopores: an alpha-hemolysin biological nanopore and a Silicon Nitride (Si₃N₄) solid-state nanopore. The alpha-hemolysin simulation allowed for an initial examination of DNA behavior in a biological nanopore, while the Si₃N₄ nanopore simulation provided detailed ionic current signatures critical for sequencing analysis. Protocols from Comer’s nanopore modeling guide were closely followed to accurately construct these systems, and no structural modifications were made. Results indicate that nanopore geometry, such as the hourglass shape in Si₃N₄, provides a stable pathway for DNA translocation and yields consistent current signals. The simulations utilized a higher-than-normal voltage to accelerate translocation, which provided faster insights but deviated from experimental conditions, suggesting areas for methodological refinement. The high than normal voltage is due to this work following steps described 15 years ago before GPU-computing which led to simplifications in the system. These findings were compared with both another Si₃N₄ simulation and real-world experimental data using Si₃N₄. The findings suggest that MD simulations can offer valuable insights into nanopore design and sequencing efficiency, providing a foundation for further optimization of real-world DNA sequencing applications. Future work could focus on refining computational methods to simulate translocation under more realistic voltages, ultimately bridging the gap between simulation and experimental results. This study contributes to the evolving field of nanopore sequencing, suggesting avenues for improvements in nanopore structure and operational conditions to enhance sequencing accuracy and reliability.