Академический Документы
Профессиональный Документы
Культура Документы
Molecular Modeling
Workshop
Laboratory for Molecular Simulation (LMS)
lms.chem.tamu.edu
mouse@chem.tamu.edu
Office: Rm. 2109 Chemistry (CHAN) phone: 845-9384
Please sign in and turn your cell phones to silent for the lecture
Microscopic Macroscopic
Time Grids
>min Continuum
Segments (FEA, CFD)
s
s Mesoscale
Atoms
ns Molecular
Electrons Dynamics
F=ma
ps Quantum
Mechanics
fs H=E
ngstroms nm m mm m Distance
February 19, 2016
Visualization
Building
Draw in 2-D
Convert to 3-D
Rotate
Rendering
Line
Stick
Ball and Stick
CPK Ball Maynard, D. and Vigh, G.
Dept. of Chemistry, Texas A&M University
Cartoon
Surfaces
Visualizing crystal structures
Visualization
Software Purchases
The Laboratory for Molecular Simulation currently provides
a wide variety of academic licenses for commercial and
academic molecular modeling software for students and
researchers at Texas A&M University.
Discovery Studio MM/MD & QM suite of software with a user-friendly Floating Licenses $$$
GUI for Life Sciences
GROMACS MM/MD specializing in speed and coarse-grained Site License* Free
simulations
GROMOS MM/MD Group License $
Materials Studio MM/MD & QM suite of software with a user-friendly Floating Licenses $$$
GUI for Materials Sciences
MOE MM/MD Drug discovery software Floating Licenses $$$
Avogadro Visualizer and GUI for many QM and MM codes Site License* Free
Chemmisian GUI for the analysis of electronic structure and spectra. Site License* $
Jimp2 GUI for Fenske-Hall (QM) and more Site License* Free
Maestro GUI for Schrdinger Suite of Software Token Based Licnese $$$
* Users need to register with the software provider at no charge. February 19, 2016
Schrdinger
Units Items
15 Tokens Glide (5 Tokens) QikProp (2 Tokens)
Liaison (4 Tokens) Canvas (1 Token)
Strike (1 Token) LigPrep (1 Token)
10 Licenses BioLuminate GUI
10 Tokens QSite (4 Tokens) pKa Predictor (3 Tokens)
MacroModel (2 Tokens) Epik (1 Token)
ConfGen (3 Tokens) SiteMap (1 Token)
Jaguar (2 Tokens) Prime (8 Tokens)
1 License PIPER
LMS holds a
3 Token License
Computational Chemistry
Quantum Mechanics
Ab initio - based on first principles
Hartree-Fock Theory (HF)
Mller-Plesset Perturbation Theory (MPn ; n = 2, 3, 4, )
Configuration Interaction (CI ; CIS, CISD, CISDT, )
Coupled-Cluster (CC ; CCD, CCSD, CCSD(T), CCSD(TQ), )
Complete Active Space Self Consitent Field (CASSCF)
Multi-Reference Configuration Interaction (MRCI)
and many more
Density Functional Theory
B3LYP, BP86, B3PW91, mPW1PW91, PBE, M06, TPSS, B97x-D
and many more
Semi-empirical
AM1
PM3, PM5, PM7
and many more
Quantum Mechanics
Ab initio methods
Schrdinger Equation
H=E ; time-independent Schrdinger equation.
Applicable to any system, in principle.
Can model bond breaking and formation
Used for benchmark values
Can only be used for small system ( normally < 200 atoms )
Computationally expensive
Scaling: Nn n=2, 3, 4, 5, 6,
Commonly used codes
Gaussian 09, GAMESS-US, Spartan, NWChem
Q-Chem, MOLPRO, Dalton, GAMESS-UK, CRYSTAL
and many more
Quantum Mechanics
Density Functional Theory (DFT)
Total energy of a system depends only on the electron density
Etot= F[(x,y,z,s)]
Applicable to any system, in principle.
Can model bond breaking and formation
Includes electron correlation with little cost compared to ab initio
methods
Exact functional is not known
Commonly used software for DFT
Gaussian 09, Jaguar, DMol3, Turbomole, Amsterdam Density Functional
(ADF), GAMESS-US, NWChem, MOLPRO, Spartan, GAMESS-UK,
CRYSTAL, and many more
Cunxiang Zhao, T. Andrew Mitchell, Ravikrishna Vallakati, Lisa M. Prez, and Daniel Romo
J. Am. Chem. Soc., 2012, 134 (6), pp 30843094
DOI: 10.1021/ja209163w
2+
3
Ru c2v Ru
2+
5
Ru c2v Ru
N
N
N N N N
2+
4
Ru cs Ru
N
N N
NMR Calculations
Original
assignment for
spectra a)
Upfield
Downfield
(E)-2a is
lower in
energy than
(E)-2a-ZnCl2
Thomson (Prez), L.M.; Hall, M.B. J. Am. Chem. Soc.2001, 123, 3995.
2+
4
2+ Ru Ru
cs
N
N N
3
Ru c2v Ru
2+
5
Ru c2v Ru
N
N
N N N N
Wavefunction Analysis
Atoms in Molecules (AIM)
A) Contour plot of the electron density of [1-F]+ showing the short and long C-F bonds. The
plane was selected to contain the C02, F, and C01 atoms. B) Contour plot of 1/4 2 (r) for
[1-F]+ illustrating the covalent nature of the short C-F bond and the dative nature of the long
C-F bond. Positive and negative values are shown with blue solid and red dashed lines,
respectively.
February 19, 2016
Quantum Mechanics
Semi-empirical Methods
AM1, PM3, PM5, PM7, PM7-TM, SAM1, etc.
Approximate solution to the Schrdinger equation
Replaces the expensive integrals with parameters
Applicability is limited by available parameterization
Mostly used for 1st row main group elements
Limited applicability to transition metals
Molecular Mechanics
Newtons equations
The potential is approximated by an empirical function
force field that is fitted to approximately reproduce
known interactions
Applicability is limited by the availability of
parameterization
Generally, the connectivity of atoms cannot change during
the simulation
Generally, not suitable for reaction mechanisms
Can predict relative energies of different conformational states of
material
And much more
Molecular Mechanics
The molecule is considered to be a collection of atoms held
together by simple elastic or harmonic forces.
Force Field
Terms - Torsion,
Inversion &
Coulombic
Terms
Force
Field
Terms
Van der
Waals
Terms
Force Field
Terms
Cross Terms
(CFF91)
Torsion Out-of-plane
Molecular Mechanics
Force Fields differ in their parameters, terms and the method of
development
Class I - simple functional form with data fitted to quantum mechanical
calculations and/or experiment (AMBER, CHARMM, etc)
Class II - more complicated functional form using cross terms and data
fitted to quantum mechanical calculations and/or experiment (CFF, PCFF,
etc)
Class III - new generation force fields that incorporate polarizability
(AMOEBA, AMBER ff02, CHARMM Drude, etc)
Rules Based - covers most of the periodic table UFF, DREIDING, etc
Fundamental quantities are derived for each atom type: electronegativity, hardness,
atomic radius, etc.
Forcefield parameters are derived at runtime using a series of theoretically or
empirically derived rules
Specialist - developed for a particular family of compounds flourinated
polymers, zeolites, etc.
Reaction Forcefields - ReaxFF
A. K. Rappe; C. J. Casewit; K. S. Colwell; W. A. Goddard III; W. M. Skiff J. Am. Chem. Soc. 114, 10024-10035
(1992).
Atom Types o=
hn
Sf
Cp
nb
s1
PCFF force field - PCFF was developed based on CFF91 and is intended for
application to polymers and organic materials. It is useful for polycarbonates, melamine
resins, polysaccharides, other polymers, organic and inorganic materials, about 20
inorganic metals, as well as for carbohydrates, lipids, and nucleic acids and also cohesive
energies, mechanical properties, compressibility's, heat capacities, elastic constants. It
handles electron delocalization in aromatic rings by means of a charge library rather than
bond increments.
February 19, 2016
Minimization
Minimize the potential energy
E = Ebond + Eangle + Etorsion + Eoop + Enonbond + Eother
Local
Energy
Minima
Local
Minima
Local
Global Minima
Minimum
February 19, 2016
Minimization Strategies
Cascade Approach - Smart Minimizer
Steepest decent
Max gradient > 100
Conjugate Gradient
Max gradient < 100
Newton
iterative (pure) Newton-Raphson.
BFGS (Broyden-Fletcher-Goldfarb-Shanno)
DFP (Davidon-Fletcher-Powell)
truncated Newton-Raphson
Final Convergence
Native
Molecular Dynamics
Minimization methods will only optimize your molecule to
the closest local minimum
Methods to find the global minimum
Systematic conformational search
Very time consuming and essentially impossible for anything but the
smallest of molecules
CH3(CH2)n+1CH3
n=1 3 possible configurations <1 sec
n=2 243 1 min
n=10 59,049 2 hour
n=15 14,348,907 100 days
Molecular Dynamics, Random Sampling, Monte Carlo, Hybrid
Monte Carlo/Dynamics methods
Samples the potential energy surface by perturbing the geometry
Molecular Dynamics
Methods to find the global minimum
High temperature dynamics
Simulated annealing
Quench dynamics
Energy supplied to the minimized
structure at the start of the simulation
T(K)
Energy
minimize
Time ps Time ps
Molecular Dynamics
Molecular Dynamics Variations
Constant Volume - Constant Temperature (NVT)
Constant Volume - Constant Energy (NVE)
Constant Pressure Constant Temperature (NPT)
1
Calculate the new position r (t + t ) = r (t )+ t v(t )+ t 2 a (t )
2
Repeat for as many time steps as desired
Molecular Dynamics
Choosing a time step
Your time step should be a factor of 10 smaller that the fastest process
in your system.
Molecular motions such as rotations and vibrations are on the order of 10-11 -
10-14 s
Therefore, a time step of 1 fs (10-15 s) or less must be used for most
systems.
You can increase your time step by restricting the fastest processes
SHAKE or RATTLE algorithms restrict the vibrational motion of the
molecule of interest
Therefore, a time step of 2-3 fs can be used with the SHAKE or RATTLE
algorithm
There are some modified shake algorithms that claim they are stable up to
time steps of 8 fs
Most simulations are on the order of picoseconds (10-12 s) or
nanoseconds (10-9 s)
Protein folding tripzip2 (12-residue protein) folds on the order of
2.5 s (10-6 s)
http://youtu.be/gFcp2Xpd29I
February 19, 2016
Conformational Searches
Ion mobility-mass spectrometry
peptide map of bovine
hemoglobin. Two low-energy
MD calculated structures are
assigned to peptide signals within
the plot: (1) LLGNVLVVVLAR
and (2) LLVVYPWTQR. The
two peptide projections shown
are 15 (top) and 10 (bottom).
Brandon T. Ruotolo, Guido F. Verbeck, Lisa M.
Prez (Thomson), Kent J. Gillig, and David H.
Russell J. Am. Chem. Soc., 124, 4214, 2002.
Ion mobility-mass
spectrometry peptide map of
horse heart myoglobin. Two
low-energy MD calculated
structures are assignaed to
peptide signals within the plot:
(1) HGTVVLTALGGILK and
(2) VEADIAGHGQEVLIR.
The two peptide projections
shown are 10 (top) and 15
(bottom).
February 19, 2016
Solvation
Explicit Solvation
Very expensive
Solvent molecules tend to boil off Non-Periodic water simulation
Informatics
Storage and retrieval of information
Databases
Structures
Properties
Activities
Combinatorial Chemistry
Protein Bioinformatics
Drug Design
Catalysis
QSAR Quantitative Structure Activity Relationships
Homology
Utilizes structure and sequence similarities for predicting unknown
protein structures.
Database comparison
Molecular mechanics/
dynamics
NMR constraints
Drug Design
Generate a pharmacophore based on a set of known
biologically active molecules.
Use the pharmacophore to search a database for other
potentially active molecules.
Drug Design
de Novo drug design analog based drug design
Design ligands to interact with a know receptor