Intro to molecular dynamics simulation YuShan Lin YSL

Molecular dynamics simulations Simulate how molecules move v (0) = 100 km = 28

Molecular dynamics simulations Simulate how molecules move Bond stretch Angle bending Non-bonded electrostatic interaction

Molecular dynamics simulations Different “force fields” have different flavors… Example: There are many water

C-terminus Protein folding simulation of NTL 9 β 3 β 1 1. 5µs of

Molecular dynamics simulations Argon atoms Goal: (what do you want to do? ) To

Molecular dynamics simulations Argon atoms Method: (how are you going to do it? )

Molecular dynamics simulations Three basic ingredients Now what do you need? 1. Description of

Molecular dynamics simulations To learn more about running MD simulations, visit our website at

Slides: 22

Download presentation

Intro to molecular dynamics simulation Yu-Shan Lin (YSL) Department of Chemistry Tufts University

Molecular dynamics simulations Simulate how molecules move v (0) = 100 km = 28 m hr s x (0) = 0 m 200 ft m = 1000 kg F = − 7000 N m a = F = − 7 2 s m = 61 m v (t) = v(0)+at = 28− 7 t x ( t) = t 0 7 v(t’ )dt’ = 28 t − 2 t 2 Can predict the position and speed of the car at any time.

Molecular dynamics simulations Simulate how molecules move v (0) = 100 km = 28 m v (1 s) = 21 x (0) = 0 m x (1 s) = 24. 5 m hr m s s m = 1000 kg F = − 7000 N m a = F = − 7 2 s m = 61 m v (t) = v(0)+at = 28− 7 t x ( t) = t 0 200 ft 7 v(t’ )dt’ = 28 t − 2 t 2 Can predict the position and speed of the car at any time.

Molecular dynamics simulations Simulate how molecules move v (0) = 100 km = 28 m v (2 s) = 14 x (0) = 0 m x (2 s) = 42 m hr s m = 1000 kg F = − 7000 N m a = F = − 7 2 s m t 0 200 ft = 61 m v (t) = v(0)+at = 28− 7 t x ( t) = m s 7 v(t’ )dt’ = 28 t − 2 t 2 Can predict the position and speed of the car at any time.

Molecular dynamics simulations Simulate how molecules move v (0) = 100 km = 28 m v (3 s) = 7 m x (0) = 0 m x (3 s) = 52. 5 m hr s s m = 1000 kg F = − 7000 N m a = F = − 7 2 s m = 61 m v (t) = v(0)+at = 28− 7 t x ( t) = t 0 200 ft 7 v(t’ )dt’ = 28 t − 2 t 2 Can predict the position and speed of the car at any time.

Molecular dynamics simulations Simulate how molecules move v (0) = 100 km = 28 m v (4 s) = 0 m x (0) = 0 m x (4 s) = 56 m hr s s m = 1000 kg F = − 7000 N m a = F = − 7 2 s m = 61 m v (t) = v(0)+at = 28− 7 t x ( t) = t 0 200 ft 7 v(t’ )dt’ = 28 t − 2 t 2 Can predict the position and speed of the car at any time.

Molecular dynamics simulations Simulate how molecules move v (0) = 100 km = 28 m hr s x (0) = 0 m m = 1000 kg F = − 7000 N m a = F = − 7 2 s m v (t) = v(0)+at = 28− 7 t x ( t) = t 0 7 v(t’ )dt’ = 28 t − 2 t 2 Can predict the position and speed of the car at any time.

Molecular dynamics simulations Simulate how molecules move v (0) = 100 km = 28 m hr This is a car s x 2(0), v 2(0) Cl- x (0) = 0 m m = 1000 kg F = − 7000 N m a = F = − 7 2 s m Na+ v (t) = v(0)+at = 28− 7 t x ( t) = t 0 x 1(0), v 1(0) 7 v(t’ )dt’ = 28 t − 2 t 2 Can predict the position and speed of the car at any time. Clx 3(0), v 3(0) Your system looks more like this…

Molecular dynamics simulations Simulate how molecules move v (0) = 100 km = 28 m hr This is a car s Cl- x (0) = 0 m F 2 (0) m = 1000 kg F = − 7000 N m a = F = − 7 2 s m F 1 (0) Na+ v (t) = v(0)+at = 28− 7 t x ( t) = t 0 x 2(0), v 2(0) x 1(0), v 1(0) 7 v(t’ )dt’ = 28 t − 2 t 2 Can predict the position and speed of the car at any time. F 3 (0) Clx 3(0), v 3(0) Your system looks more like this…

Molecular dynamics simulations Simulate how molecules move v (0) = 100 km = 28 m hr This is a car s x 2(dt), v 2(dt) Cl- x (0) = 0 m m = 1000 kg F = − 7000 N m a = F = − 7 2 s m Na+ x 1(dt), v 1(dt) v (t) = v(0)+at = 28− 7 t x ( t) = t 0 7 v(t’ )dt’ = 28 t − 2 t 2 Can predict the position and speed of the car at any time. Clx 3(dt), v 3(dt) Your system looks more like this…

Molecular dynamics simulations Simulate how molecules move v (0) = 100 km = 28 m hr This is a car s Cl- x (0) = 0 m F 2 (dt) m = 1000 kg F = − 7000 N m a = F = − 7 2 s m Na+ t 0 F 1 (dt) x 1(dt), v 1(dt) v (t) = v(0)+at = 28− 7 t x ( t) = x 2(dt), v 2(dt) 7 v(t’ )dt’ = 28 t − 2 t 2 Can predict the position and speed of the car at any time. F 3 (dt) Clx 3(dt), v 3(dt) Your system looks more like this…

Molecular dynamics simulations Simulate how molecules move v (0) = 100 km = 28 m hr This is a car s x 2(2*dt), v 2(2*dt) Cl- x (0) = 0 m m = 1000 kg F = − 7000 N m a = F = − 7 2 s m Na+ x 1(2*dt), v 1(2*dt) v (t) = v(0)+at = 28− 7 t x ( t) = t 0 7 v(t’ )dt’ = 28 t − 2 t 2 Can predict the position and speed of the car at any time. Clx 3(2*dt), v 3(2*dt) Your system looks more like this…

Molecular dynamics simulations Simulate how molecules move Bond stretch Angle bending Non-bonded electrostatic interaction − + These parameters form a “force field” Torsion Non-bonded Van der Waals interaction

Molecular dynamics simulations Different “force fields” have different flavors… Example: There are many water models… − 0. 8476 − 0. 8340 1. 0000 Å 0. 9572 Å 109. 47° +0. 4238 104. 52° +0. 4238 “spc/e” +0. 4170 “tip 3 p” simple point charge/extended transferable intermolecular potential 3 -point 2. 5 5. 5 Diffusion coefficient at 25°C D 25°C (10 -5 cm 2/s) Exp: 2. 3 Example: There are many peptide force fields… “Helix-friendly” “β-Sheet-friendly”

C-terminus Protein folding simulation of NTL 9 β 3 β 1 1. 5µs of simulation = [ 600 x 106 steps ]x[ 2. 5 fs/step ] Movie: configuration at every 1 ns, 1500 snapshots α N-terminus “Primary sequence” β 2 α-Helix β-sheet 310 -Helix β-bridge + +- + ++ - + + + - + MKVIFLKDVKGMGKKGEIKNVADGYANNFLFKQGLAIEATPANLKALEAQKQ K. Lindorff-Larsen, S. Piana, R. O. Dror, D. E. Shaw, Science 334, 517 (2011)

Molecular dynamics simulations Argon atoms Goal: (what do you want to do? ) To study the structure and dynamics of Ar You meant…? 3 Ar atoms? Ar liquid/gas interface? A droplet of Ar? Liquid Ar? Temperature=? Volume/Pressure=? Phase diagram of Ar

Molecular dynamics simulations Argon atoms Method: (how are you going to do it? ) Classical molecular dynamics simulations Is this a suitable method? Are there situations where classical molecular dynamics simulations won’t work? MD simulations of liquid Periodic Boundary Condition

Molecular dynamics simulations Three basic ingredients Now what do you need? 1. Description of initial positions 1. Description of initial velocities Maxwell-Boltzmann Distribution Equipartition Theorem 2. Description of interaction potentials (aka “force field”) Lennard-Jones Potential/Force (and others) Cut-off? Hard/Shifted/Shift-force 3. An integrator Leap-frog Algorithm

Molecular dynamics simulations To learn more about running MD simulations, visit our website at http: //ase. tufts. edu/chemistry/lin/outreach. html