Computational Drug Discovery and Design

Table 2
List of various enhanced sampling methods in molecular dynamics (MD) simulation

Enhanced MD technique Methodology Applications

Accelerated MD Adds a positive boost potential to the

system’s potential energy (PE) when

it drops below a certain threshold,

thus decreasing the energy barriers

and accelerating the transitions

between low energy conformational

states

Protein–ligand binding and protein

folding mechanisms

Adaptive biasing force

(ABF) simulation

Adds a continuously updated biasing

force to the equations of motion,

such that in the long-time limit a

Hamiltonian devoid of an average

force acting along the transition

coordinate of interest is added

Free-energy calculations along a

reaction coordinate, potentials of

mean force (PMF) profile

estimation

Biased/Steered

molecular dynamics

(SMD)

A dummy atom is attached to the

center of mass (COM) of the ligand

through a virtual spring with a

spring constant (k), which is later

pulled with a constant velocity (v)

along a predefined direction for the

unbinding of ligand

Free energy calculations, binding/

unbinding mechanisms of

biomolecular systems, PMF profile

estimation

Conformational flooding Adds a flooding potential to the

systems’s PE that destabilizes the

initial conformation, thus lowering

free-energy barriers of structural

transitions and accelerating the

transitions

Protein structure prediction/

conformational search, check

protein models stability, predict

functional motions, improve

thermodynamic quantities like free

energies and entropies for proteins

Dynamic importance

sampling (DIMS)/

self-guided Langevin

dynamics

Employs a bias with correction

approach to improve the sampling

efficiency of rare events

Transition pathways

Local elevation (LE) Adds a bias term (i.e., sum of repulsive

functions) to the PE function

Low-energy structure search

Metadynamics The location of the system is calculated

based on the collective variables

A positive Gaussian potential is

subsequently added to the energy

landscape to drive the simulation

back to its previous location

Folding of small proteins, protein

switches, ligand dissociation

pathways, and ion-induced

diffusion of small molecules in

channels or cavities

(continued)

Molecular Dynamics Approach for Investigation of GPCR Allostery 467