Introduction to Molecular Modeling

Molecular modeling/ computational chemistry is a branch of chemistry

concerned with Theoretically determining properties of molecules.
 “Computational chemistry simulates
chemical structures and reactions
numerically, based in full or in part
on the fundamental laws of physics.”
Foresman and Frisch
In “Exploring Chemistry with Electronic Structure Methods”, 1996

2
 How Molecular Modeling is Useful

 Molecular modeling/ Computational chemistry is used in a number of different ways.

 One particularly important way is to model a molecular system prior

to synthesizing that molecule in the laboratory.
This is very useful information because synthesizing a single
compound could require months of labor and raw materials, and
generate toxic waste.

 It is useful to understanding a problem more completely.

There are some properties of a molecule that can be obtained
computationally more easily than by experimental results.

 There are also insights into molecular bonding, which can be obtained from the
results of computations, that may not be obtained from any experimental method.

Thus, many experimental chemists and biologists are now using computational
modeling to gain additional understanding of the compounds being examined
in the laboratory.
3
Molecular modeling includes all theoretical methods and computational
techniques used to model or mimic the behavior of molecules.

Molecular modeling allows us to do and teach chemistry better by providing

better tools for investigating, interpreting, explaining and discovering new
phenomena.

These techniques are used in the fields of computational chemistry, drug

design, computational biology and material science for studying molecular
systems ranging from small chemical systems to large biological molecules
and material assemblies.

Molecular modeling is easy to perform with currently available software,

but the difficulty lies in getting the right model and proper interpretation.

The common feature of molecular modeling techniques is the atomistic level

description of the molecular systems..
What do we do with Molecular Modeling

 Molecular geometry – shapes of molecules

 Energies of molecules and transition states
-prediction of favorable isomer and reaction rates

 Reaction mechanisms and reaction path following studies

 Molecular Orbital Calculations

 Vibrational Frequencies

 UV, Fluorescence and NMR spectra

 The interaction of a substrate with an enzyme

 Electron and Charge distributions

 Chemical reactivity

 The physical properties of substances

5
Computational chemistry / Molecular Modeling

• Fairly cheap
• Fast compared to experiment
• Environmentally safe
• Does not replace experiment
• To make something – new drugs, new materials – one has to go into the lab
• Computation has become so reliable in some respects

6
I. Types of Calculations

 There are three basic types of calculations. From these

calculations, other information can be determined.

• Geometry Optimization
– Given starting structure
– Calculate lowest energy conformation
– Can also optimize to transition states

• Vibrational Frequencies
– vibrational specrta
– Characterize the shape of potential energy surface

• Single Point
– Given geometry
 Binding energies
– Calculate total energy and other properties
7
II. Computational Models

 A model is a system of equations, or computations used to

determine the energetics of a molecule

 Different models use different approximations (or levels of theory)

to produce results of varying levels of accuracy.

 There is a trade off between accuracy and computational time.

 There are two main types of models; those that use Schrödinger's
equation (or simplifications of it) and those that do not.

8
 Computational Models

Molecular mechanics Electronic structure methods

(Classical Mechanics) (Quantum Mechanics)

Law of classical physics to predict Laws of quantum mechanics rather

the structures and properties than classical physics.

• semi-empirical methods
Prog: MM3, Hyperchem…etc • ab initio methods
There are many different MM
methods.
Energy and other related properties
of a molecule may be obtained by
solving the Schrödinger equation:
H=E 9
Comparison

10
 III. Basis Sets
A basis set is a collection (set) of mathematical functions used to solve the
Schrödinger equation (mathematical equations represents the shapes of
orbitals occupied by the electrons and their energies).

Basis sets in common use have a simple mathematical form for representing
the radial distribution of electron density.

Linear combination of basis functions approximates total electronic

wavefunction

The accuracy of a calculation is dependent on both the model and the

type of basis set applied to it.

 Larger basis sets will describe the orbitals more accurately but take
longer to solve.

11
Molecular Mechanics Assumptions:

 Nuclei and electrons are together treated as unified atom-like particles

 Atom-like particles are treated as spherical balls
 Bonds between particles are viewed as springs
 Interactions between these particles are treated using potential functions
derived from classical mechanics
 Individual potential functions are used to describe different types of interactions
 Potential energy functions rely on empirically derived parameters that describe
the interactions between sets of toms
 Potential functions and the parameters used for evaluating
interactions are termed a force field
 The sum of interactions determines the conformation of atom-like particles

12
Molecular Mechanics

Concepts:
 Macroscopic property
 Uses laws of classical physics
 Atoms are considered to be spheres
that are connected to other atoms by a spring
 Individual method is characterized by its particular Force Filed (FF)

 Also known as Force Field Methods

 MM methods are available in many computer programmes

eg., Gaussian, Hyperchem, Sybyl, Quanta, Alchemy etc,.
 Calculations are very fast, hence can be applied to large molecular systems

13
Geometrical Parameters

Definition of basic parameters in force fields. Bondlengths (l), angles (), torsion angles
() and nonbonded distances (r) of n-propanol

 E
bonds
stretch  E
angles
bend  
dihedrals
Etorsions   Enonbond
pairs
14
Applications:

Molecular Mechanics

 Energy minimization / geometry optimization

 Calculation of binding constants, Protein folding kinetics etc.,

 Docking, QSAR studies (Receptor-ligand interactions - drug design)

 Molecular dynamics

 Molecular mechanics energies can be very accurate for families of

compounds for which the force field has been parameterized.

15
Electronic structure methods
Based on Schrödinger’s Equation: Hψ = Eψ
 The Schrödinger equation is the basis of quantum mechanics and gives a
complete description of the electronic structure of a molecule.

Hψ = Eψ Ψ is f(x,y,z)
A set of spatial distributions
H is energy operator which describes
describing the probability of
the kinetic and potential energy of an
finding electrons (orbitals).
electron in field of nuclei and other
electrons.
H= Ek + Ep A single atom has AOs.

P2 Molecules have MOs.

Ek 
2m
p is the momentum and m is the mass of the electron

1
E p  kx2
2
k is constant and x is the direction of electron E = sum of the energies of the orbitals, which
may contain 1, or at most 2, electrons.

 Schrodinger wave equation is a differential equation

 Solutions of this equation give the wave function

16
Schrödinger’s Equation: Hψ = Eψ …….

 The Schrödinger equation is the basis of quantum mechanics and gives a

complete description of the electronic structure of a molecule. If the
equation could be fully solved, all information pertaining to a molecule
could be determined.

 Complex mathematical equation that completely describes the chemistry of a

molecular system.

 Not solvable for systems with many atoms.

 Since Schrödinger's equation cannot be completely solved for molecules with more
than a few atoms, computers are used to solve approximations of the equation

17
 Approximations to solve the Schrödinger equation

Born-Oppenheimer:
Compared to electrons, nuclei are stationary
Electrons move in field of fixed nuclei

Hartree-Fock:

Separate Ψ (many electron wavefunction) into series of one

electron spin orbitals

LCAO (Linear Combination of Atomic Orbitals):

MO’s expressed as linear combinations of single electron

atomic orbitals, represented by basis functions

18
 Ab intio Methods

It means ‘from first principles’ or ‘from the beginning’

Based on approximated quantum mechanical calculations.

It’s purely from theoretical principles with no inclusion of experimental data

Most accurate, more difficult to perform and time-consuming processes.

This method is an iterative procedure based on SCF method

Energy gives in Hartrees units

Types of Ab intio Methods

Ab Initio

HF Post HF DFT

Configuration Interaction

Møller-Plesset Perturbation theory

Coupled Cluster theory

MultiConfigurational SCF

20
 Hartree--Fock method (HF)
Hartree
Hψ = Eψ
To solve the Schrodinger equation, we need to make some approximations.
These will lead to the Hartree-Fock method (which is the simplest ab initio
method)
The first approximation is Born-Oppenheimer approximation, it separates the
nuclear and electronic motions .
The mass of the nuclei is much greater than the electrons,
hence the electrons can respond almost instantaneously to any change
in the nuclear positions
Only the electron motions need to be considered because the nuclei are
relatively motionless, i.e. Electrons are moving in the field of fixed nuclei
K.E. of nuclear can be neglected;
Nuclear-nuclear repulsion term can be considered as constant
It helps to separate two terms (nuclear and electron)

Total  electronic   nucleus

H  k (r )  k ( R)  p(r)  p( R)  p( R, r )
H (R, r )(R, r )  E ( R, r ) 21
 The next approximation is Hartree-fock approximation, decompose ψ
into a combination of molecular orbitals. MO: one-electron wave
function (n)
In 1927 D.R Hartree introduced a procedure, which he called the self-
consistent field method, to calculate approximate wave functions

Hartree product    (1) (2) (3)..... (n)

Here  is a function of the coordinates of all the electrons in the atom, (1) is a function of the coordinates
of
electron 1, (2) is a function of the coordinates of electron 2, etc.

 0   0 (1)0 (2)0 (3)..... 0 (n)

1(1)   c (1) 0 (2) 0 (3).....0 (n)

1(2)  1(1) c (2) 0 (3)..... 0 ( n)

1(3)  1 (1)1 (2) c (3)..... 0 ( n)

This completes the first cycle

of calculations and gives

1  1(1)1(2)1(3).....1 (n)
22
 Repetition of the cycle gives

 2   2 (1) 2 (2) 2 (3)..... 2 (n)

The process is continued for k cycles till we have a wavefunction k

The energy calculated from k that are essentially same as the

wavefunction and/or energy from the previous cycle.

At this stage the field of cycle k is essentially same as that of cycle k-1,
i.e. it is “consistent with” this previous field, and so the procedure is called
the self-consistent-field (SCF) procedure

However, this is not a good wave function, there are two problems with
Hartree product
1) Electrons have a property called spin (which is not
considered)
2) Any rigorous attempt to approximate the wavefunction  should use an
antisymmetric function of the coordinates of the electrons 1, 2, . . . n, but
the Hartree product is symmetric rather than antisymmetric
To approximate a helium atom wavefunction (as the product of two
hydrogen atom 1s orbitals), then

   a b

 a  1s( x1, y1, z1)1s( x2 , y2 , z2 )

 b  1s( x2 , y2 , z2 )1s( x1, y1, z1)

 a  b symmetric

as wave functions need to be antisymmetric

These two problems in hartree product have been corrected by Fock and Slater
by constructing the wavefunction as a Slater Determinant:

Slater total wavefunction  is from one-electron wavefunctions (with spin

and spatial orbitals) and  will be antisymmetric wave function
24
Consider 4 electron system

   (1) (2) (3) (4)

Slater determinant for a four-electron system

25
The next approximation is expressing the molecular orbitals as a linear
combination of a pre defined set of one electron functions known as basis
function or AOs

Basis functions or AOs

MO N
i 

1
ci 

Molecular orbital expansion coefficient

These approximations leads to the Hartree-Fock equations, which can be

solved by the Self-Consistent Field (SCF) method

Use the self-consistent field (SCF) method:

 Begin with guess set of orbitals.

 Solve HF equations to yield new, improved set.
 Use new set in HF equations to yield further improved set.
 Iterate until system meets convergence criterion.

26
 Input 3D
coordinates

Initial guess
MOs

Solve HF Improved
equation Set of MOs

Calculation Yes SCF No

End Converged?

27
The main weakness of Hartree Fock is that it neglects electron correlation

Electron correlation: ignored the interaction between electrons in the

electronic structure (electron-electron repulsions)

Post-HF methods include electron correlation

Correlation Methods (Post-HF Methods)

 Configuration Interaction (CI)

 Møller-Plesset Perturbation theory (MP)
 Coupled Cluster theory (CC)
 MultiConfigurational SCF (MCSCF)

28
 Density Functional Theory.
Background:
 1964: Hohenberg-Kohn paper proving existence of exact DF.
 1965: Kohn-Sham scheme introduced
 1970s and early 80s: DFT becomes useful.
 1988: Becke and LYP functionals. DFT useful for some chemistry.
 1998: Nobel prize awarded to Walter Kohn in chemistry for
development of DFT (Walter Kohn, John Pople).

The Nobel Prize in Chemistry 1998 was divided

equally between Walter Kohn "for his development
of the density-functional theory" and John A. Pople
"for his development of computational methods in
quantum chemistry".

29
 DFT
Wavefunction based methods in QC to get correlation energy are quite time-
consuming (MP2 and CCSD etc....)

DFT techniques provide a means for recovering correlation energy at a

fraction of the computational cost

DFT is the most popular QC method used today.

DFT describes a molecular system directly via its density (another strategy
for solving SE using functional).

Includes some part of electron correlation.

Considered an ab initio method, but different from other ab initio methods

because the wavefunction is not used to describe a molecule, instead the
electron density is used.

30
DFT

Theoretical framework for utilizing the electron density as the primary

variable to determining the properties of a molecular system

Electron Density
– The measure of the probability of finding an electron at a point in space.

In DFT, the energy of a molecular system is derived as a functional of the

electron density

DFT method considers the electronic energy in to several terms

E is the electronic energy
Ej is the electron-electron repulsion term and also describes coulomb self-
E  E  E p  E j  Exc interaction of the electron density.
k
Exc is the exchange-correlation term

31
DFT: Functionals
• Many different advanced functionals have been developed.
• Generally described by two parts:
– Exchange functional
– Correlation functional

For a given basis set, the difference between the

exact energy (ɛ0) and the HF energy (EHF) is the
The two or more
correlation energy. (~ 85 kJ/mol correlation
electrons with the same
energy per electron pair)
spin in the degenerate
orbitals sub shell, tend
to exchange their
positions and the
energy released due to
the exchange is called Electron correlation is mainly caused by the
exchange energy. instantaneous repulsion of the electrons
The No. of exchanges
that can take place is
maximum when the sub
shell is either half filled
or completely filled.

•For example, BLYP functional is made up of B (Becke exchange) and LYP

(Lee, Yang and Parr correlation).
32
DFT

COMBINATION FORMS ……

33
DFT: Functionals
Hybrid Functionals

• Try to overcome difficulties of pure exchange functionals by mixing

in component of exact exchange from HF theory.

• Most common hybrid functional is B3LYP.

Method HF %

B3LYP 20
PBE0 25
M06 27
BHHLYP 50
M062X 54
CAM-B3LYP 65
M06HF 100

34
DFT

Hybrid functionals

35
DFT
• Advantages:
– Treat larger molecules than possible using post-HF ab initio.

– Particularly good for ground state and excited states.

• Disadvantages:
– Not systematically improvable.

– Bewildering choice of functionals

Programs available with DFT

Gaussian, Q-Chem,
NWChem, MPQC,
GAMESS, Turbomol,
Molpro, Jaguar,
Spartan....

36
Ab Initio Applications
 Equilibrium structures
 Transition State structures
 IR, NMR spectra
 Reaction energies
 Reaction barriers
 Dissociation energies
 Charge distributions
 Reaction Rates
 Reaction Free Energies
 Circular Dichroism (optical, magnetic, vibrational)
 Spin-orbit couplings
 Excited States (vertical)
 Solvent Effects
 pKa’s
 Density matrix methods/geminals
 Accurate enzyme-substrate interactions
 Protein folding
 Full reaction dynamics
 Excited States (adiabatic)…
Energy minimization

To find a route from an initial conformation to the nearest minimum energy

conformation using smallest number if calculations possible

Stable states of molecular systems correspond to minima on their potential

energy surface

38
Energy minimization
When a molecule is built in a computational chemistry software package, the
initial geometry does not necessarily correspond to one of the stable
conformers.
Therefore, energy minimization is usually carried out to determine a stable
conformer; this same process also is commonly referred to as geometry
optimization.
To find a route from an initial conformation to the nearest minimum energy conformation using
smallest number if calculations possible called energy minimization
(OR)
Energy minimization is a numerical procedure for
finding a minimum on the potential energy
surface starting from a higher energy initial
structure

During energy minimization, the geometry is changed in

a stepwise fashion so that the energy of the molecule is
reduced, from steps 2 to 3 to 4 as shown in Figure.

The process of energy minimization changes the

After a number of steps, a local or global minimum on geometry of the molecule in a step-wise fashion
the potential energy surface is reached. until a minimum is reached.
Energy minimization
The energy of the molecule changes with its structure/geometry
Understanding the geometry optimization is the major requirements for energy
minimization

Minimization methods:
Molecular Mechanics
Semi empirical: AM1, PM3
Ab initio: HF, MP2, MP4, CIS CISD, CASSCF, CCSD, CCSDT and more

Energy Minimization Procedures

There are numerous procedures for actually varying the geometry to find the
minimum. Many of the methods used to find a minimum on the potential
energy surface of a molecule use an iterative formula and proceed in a step-
wise fashion.

These are all based on formulas of the type:

In the equation, xnew refers to the value of the geometry at the next step (for example, moving from step 1 to 2 in the figure),
xold refers to the geometry at the current step, and correction is some adjustment made to the geometry.
Energy Minimization Procedures

 Steepest Descent Minimiser

 Conjugate Gradient Method (CGM)
 The Powell Method
 Newton Rapson Method
Geometry optimization

It is used to find minima on the potential energy surface

In this course of minimization, molecular structure will be relaxed

It is applied to model-built structures and those derived from coordinate data
banks

Molecular Mechanics and/or Quantum chemical methods can be applied

Bond lengths, Bond angles and Dihedral or Torsional angles are considered

Geometry optimization can be used to:

Characterize a Potential Energy Surface (PES)

Obtain a structure for a single point quantum chemical calculation

(provides structural and electronic properties)

Prepare a structure for Molecular Dynamics Simulations

 Potential Energy Surfaces

A potential energy surface is a plot of energy vs. reaction coordinate. It

connects reactants to products via a transition state.

Energy minima correspond to equilibrium structures.

The energy maximum corresponds to a transition state structure.

44
The relative energies of equilibrium structures give the relative stabilities of
the reactants and products (thermodynamics).

The energy of the transition state relative to reactants gives information

about the rate of reaction (kinetics).

Molecular modeling is primarily a tool for calculating the energy of a given

molecular structure.

The first step in designing a molecular modeling investigation is to define the

problem as one involving a structure-energy relationship.
45
 Potential Energy Surfaces

Mathematical relationship linking the molecular structure and the resultant energy

For Diatomic systems, Potential energy illustration is 2-Dimensional

For larger systems, it is 3-Dimensional

Minima: is point at the bottom of valley, from which motion in any direction leads
to a higher energy

Local Minima: Lowest point in limited region of Potential surface

Global Minima: Lowest point anywhere on the potential surface

Saddle Point: 1 or more imaginary frequencies

Different minima correspond to different conformations or structural isomers of the

molecule under investigation

In order to distinguish a local minimum from the global minimum, it is

necessary to perform a conformational search.
Potential Energy surfaces
Using Gaussian09W
Transition State:

It corresponds to Saddle Point

1st order saddle point is TS

 Conformational analysis
Conformational isomers: The isomers, which can be interconverted
exclusively by rotations about formally single bonds

Most of the molecules can adopt more than one conformation or

molecular geometry simply by rotating around the rotatable bonds

Different conformations of a molecule can be regarded as different spatial

arrangements of the atoms, these are the inter convertible and most of them
cannot be isolated

25 kJ mol-1 14 kJ mol-1 14 kJ mol-1 25 kJ mol-1

(6.00 k cal mol-1) (3.35 k cal mol-1) (3.35 k cal mol-1) (6.00 k cal mol-1)

3 kJ mol-1
3 kJ mol-1 (0.72 k cal mol-1)
(0.72 k cal mol-1)
0.00 kJ mol-1
(0.00 k cal mol-1)

53
Conformational Analysis

Two distinct computational approaches to calculate the molecular geometry

1) Quantum Chemical Methods (Ab initio and Semiempirical)

2) Molecular Mechanics Methods (Force Field)

Conformational Search

Randomly or systematically
generated starting geometries

Energy minimization

Duplicates elimination

Gives coverage of potential surface

Conformational Search Methods

1. Grid searches

2. Monte Carlo algorithms

3. Simulated annealing

4. Genetic algorithms

5. Rule-based systems

6. Chain growth algorithms where applicable

7. Fragment-based algorithms

8. Distance-geometry algorithms

9. Homology-based starting structures

Conformational analysis
Cyclohexane
Bioactive conformation:

The conformation adopted by ligands when bound to a

biological target

The conformation responsible for the biological activity

Bioactive conformations are target-dependent

2.3 kcal

Bio-active conformations reside within well defined energy windows

relative to global energy minima
Methods for finding Bioactive conformation from low-energy conformations:

1) Keep very diverse set of conformers

2) Retain a set of low energy conformers that are with in a certain conformational energy
threshold of global minimum
3) Based on assessment of the RMS differences in conformations (Duplicate Removal)

Duplicate Removal: The extent to which conformers are discarded based on how
similar their RMS values
Conformational Restriction

Ligand binding to a biological target is generally associated with an entropic

penalty arising from the loss of conformational degrees of freedom of a molecule

Conformational restrictions are introduced into molecules

to increase the binding affinities
to mimic the known inhibitors

Conformational locks can be introduced into molecules by adding substituent to a

molecule to create constraints that favor a particular conformation, or by
introducing ring closures
 Conformational Restriction
The introduction of substituents in the ortho position was favored the orthogonal
conformation, reversing the orthogonal/coplanar ring relationship

Protein
Tyrosin
Phospatase
-1B
Conformational Restriction
 Applications (case study)

p-ABDI (p-aminobenzylideneimidazolinone)

p-ABDI in DMSO-d6 (a) before and

(b) after 20 min of irradiation with 350
nm light and (c) after 24 h at rt in the
dark.
1H NMR spectra of (a) the Z and (b) E isomer of p-ABDI in
DMSO-d6 and HPLC chromatograms of p-ABDI in THF (c)
before and (d) after irradiation with 350 nm light
Applications
Stable conformation
p-ABDI- E p-ABDI -Z

-704.60027031 (2.39)
-704.60408918 (0.00)

Vibrational modes
Applications
Absorption spectra

expt max nm  Transition %Ci

p-ABDI- Z 370 nm 361 0.823 HL 98

Mos and electron densities

HOMO LUMO
Applications O
H
O

Absorption and emission spectra

OH Curcumin O
OH

AM1 Computed Measured

Abs. Fl.  Abs. Fl. 
Cyclohexane 422 489 67 409 444 35
Methanol 468 608 140 428 546 118
Possible computations:

Atomic Charges

Bond orders

Ionisation Poteentials

Electron Affinities

UV Spectra (absorption properties)

NMR spectra

Polarisabilities

Molecular orbitals generation and visualization