Application of graph theory in drug design

Reihaneh Safavi-Sohi, Jahan B Ghasemi
Drug Design in silico Lab
Chem Faculty, K. N. Toosi Univ of Tech
Tehran, Iran

History
Graph theory is a branch of mathematics which studies the
structure of graphs and networks.
Graph theory started in 1736, when Euler solved the
problem known as the Konigsberg bridges problem.
2

 Graph considers sets of objects, called nodes, and the
relationships, called edges, between pairs of these objects.
 A collection of dots that may or may not be connected to
each other by lines. It doesn't matter how big the dots are,
how long the lines are, or whether the lines are straight,
curved.
 Neighbor: The neighbors of a vertex are all the vertices
which are connected to that vertex by a single edge.
 The degree of a vertex in a graph is the number
of edges that touch it.
 The definition is completely general, allowing graphs to
be used in many different application domains as long as
an appropriate representation can be derived.
3

Seven Bridges of Königsberg
The problem was to find a walk through the city that would cross each bridge once and
only once. The islands could not be reached by any route other than the bridges, and
every bridge must have been crossed completely every time.
4

Isomorphism Algorithm
 Isomorphism procedures enable the comparison of pairs of
graphs .
 To detect whether two graphs, or parts of there, are
identical.
 Maximal common subgraph (MCS) isomorphism enables
the identification of the largest subgraph common to a pair
of graphs.
6

A subgraph isomorphism exists if all the nodes (atoms) of one graph (GQ)
can be mapped to a subset of the nodes of the other graph (GF) in such a
way that the edges (bonds) of GQ simultaneously map to a subset of the
edges in GF.
 The labels carried by the nodes and edges
(atom type and bond type, respectively) mustbe
identical if the nodes or edges are to be mapped to each other.
 Testing for subgraph isomorphism is an NP-complete problem.
 Trying every possible way of mapping each of the n nodes in GQ onto one
of the nF nodes in GF (nQ < nF);
Each of the nQ!/(nF - nQ)!nF! possible mappings must then be tested to see
if any of them obeys the adjacency condition.
 Even for very small graphs, the number of possible mappings rapidly
becomes unmanageable.
 The use of this approach has been limited due to its combinatorially
explosive nature and it has not been possible to apply the technique to
complex and large structural databases.
7

Molecular Similarity
Molecular similarity is a measure of the degree of overlap
between a pair of molecules in some property space.
 It can be calculated for a wide range of molecular properties.
 An important tool in the generation of quantitative structure-
activity relationships (QSAR's).
 QSAR aim to link systematically the chemical or biological
properties of molecules to their structures (rational drug
design)
9

Why similarity searching is important?
 Development of combinatorial chemistry and the ability to synthesise and to
assay vastly greater numbers of molecules than even a very few years ago
requires tools to rationalise the resulting structural and biological data.
 Simply increasing the throughput of the synthesis and test cycle in
itself does not necessarily lead to more high quality lead compounds.
 Many of the techniques that have been developed are based on the concept
of molecular similarity.
 Similarity methods have been used for many years and are typically applied
early in the drug discovery process when little is known about the biological
target.
10

The main aims and objectives
1. Similarity searching methods in chemical data
bases (Identification of new Cliques)
2. Exploring Binding Site Similarity (to find similarities
amongst various cavities)
3. Ligand Superposition (Alignment)
4. Clustering compounds
11

Similarity searching requirements
Definition of :
1. A chemistry space
2. Molecular descriptors 2D, 3D
3. A similarity index
12

2D descriptors (Physicochemical properties)
 Molecular weight
 Octanol - water partition coefficient
 Total energy
 Heat of formation
 Ionization potential
 Molar refractivity
 2D Fingerprints : record the presence or absence of molecular
fragments within a molecule.
13

2D Substructure Searching
Query
N
O
O
N N
O
O
O
O
N N
N
N
NN
N
N O
N
N
O
O
N
O
O
N
O
O
14

Problem
• Structurally similar molecules will tend to have the
same properties.
• Less effective at identifying compounds that have
similar activities but that do not share the same
structural skeleton.
 Similar Property Principle: Structurally similar compounds
will exhibit
similar physicochemical and biological properties
 For lead discovery want a diverse space to locate all possible
hits (actives) – called a diverse library
15

3D Similarity
 Distance-based and angle-based descriptors (e.g. inter-atomic distance)
 Field similarity
 Comparative Molecular Field Analysis (CoMFA), CoMSIA
 Electrostatic potential
 Shape
 Electron density
 Any grid-based structural property
 Shape descriptors
 van der Waals volume and surface (reflect the size of substituents)
 Molecular Shape Analysis
 WHIM descriptors (Weighted Holistic Invariant Molecular Descriptors)
 Receptor binding
16

O
O N
a
b
c
a = 8.62 0.58 Angstroms
b = 7.08 0.56 Angstroms
c = 3.35 0.65 Angstroms+-
+-
+-
O
O
O
O
OO
N
O
O
O
N
N
N O
O
O
O
O
O
N
N
N
N
S
O
O
O O P O
O
O P O
O P
O
O
O
O
N
N
N
N
N
O
O
O O
O
N
N
N
O
N
O
O
O
17

3D descriptors
 3D descriptors are highly selective than 2D descriptors
 Handling of conformational flexibility is a major problem !
18

Creativity is not about inventing something
totally new, it is about making new connections!
(Albert Einstein)
219

History
• A back-tracking graph algorithm for establishing a
mapping between two structures was first described
by Ray and Kirsch in 1957.
• The use of reduced graph representations of
individual structures and the creation of
hyperstructures encompassing all
the structures in a database
 was first described by Martin,Y.C.,PeterWillett
 University of Sheffield, UK,1990.
20

The HierarchicalTree Substructure Search (HTSS)
Each level in the hierarchical fragment tree is effectively part of an
hierarchical classification of all the atoms in the database as a
whole, initially by number of neighbors and atom type, and then
by bonding pattern and atom type of neighbors.
21

Similarity Searching using Reduced Graphs
Reduced graphs are topological graphs that
attempt to summarize the features of molecules
that can result in drug-receptor binding while
retaining the connections between the features.
potential to find structures that have the same binding
characteristics with different carbon skeletons and may
belong to different lead series.
22

 R ring systems
 Ac acyclic comp.
 C corbon
 H heteroatom
 Ar aromatic rings
 F functional group
23

 Many ways of computing the similarity between two
molecules
 Different representations
 Different similarity coefficients
26

2D Similarity Searching
OH
N
N
OH
N
H
N NH2
O
N
N
H
N
NH2
N
H
N
NH2
N
H
N
N
N
H
N
N
H
OH
Query
27

Query
N
O
O
N N
O
O
O
O
N N
N
N
NN
N
N O
N
N
O
O
N
O
O
N
O
O
28

Enrichment Factor
EF =
na : Number of actives, in the top of 10% of the ranked
hit list
Na : Total number of actives
33

New Cliques Detection
 Discovery of novel bioactive molecules
34

Graph-Based Molecular Alignment
(GMA) was first carried out by J.Apostolakis
35

RMSD = 1.29
Comparison between aligned pose and crystal structure
RMSD= 0.59
Template is kept rigid, query molecule is flexible
By using torsion space optimization
36

Searching 3D Protein Structures
 Extensive collaboration between Information Studies and
Molecular Biology and Biotechnology to develop graph
representations of proteins that can be searched with
isomorphism algorithms analogous to those used for
chemical structures
 Focus here on folding motifs (secondary structure elements)
in proteins and protein amino acid side chains
37

Representation Of Protein Folding Motifs
 The helix and strand secondary structure elements (SSE)
are both approximately linear, repeating structures, which
can be represented by vectors drawn along their major axes
 The nodes of the graph are these vectors (individual amino
acid side-chains) and the edges comprise:
 The angle between a pair of vectors
 The distance of closest approach of the two vectors
 The distance between the vectors’ mid-points
 PROTEP compares such representation using a maximal
common subgraph isomorphism algorithm to identify
common folds
38

Secondary Structure Elements (SSE)
39

The majority of PDB files include lines labeled HELIX and
SHEET, detailing the residues involved in α-helices and β-
sheets within the protein.
 This information is used to label each residue in the vector
file with
 ‘h’ to indicate α-helix
 ‘s’ to indicate β-sheet
 ‘x’ for residues that are in neither of these.
 e.g. ASPh
 A minority of PDB files (about 7.5% in the search files)
do not have secondary structure information
40

Data Fusion
 Improved performance can be obtained in many
classification tasks by combing evidence from several
different sources
 Calculate the similarity between a user query and each of the
documents in a database
 Rank the documents in order of decreasing similarity
 Repeat using several different representations, coefficients,
etc.
 Add the rank positions for a given document to give an overall
fused rank position
 The resulting fused ranking is the output from the search
 Small, but consistent, improvements in performance over use
of a single ranking
41

M2
M41
M23
M3
M9
M18
.
.
.
.
M6
M89
M234
M13
M90
M180
.
.
.
.
M8
M130
M257
M16
M99
M198
.
.
.
.
sort
M8
M16
M99
M130
M198
M257
.
.
.
.
sum
Ar/F R/F Fused Data
42

Fusion Of Chemical Similarity Coefficients
 Searches were carried out using different
similarity coefficients, and the resulting rankings
fused to give rankings corresponding to all
combinations of 1, 2, 3…. 20, 21 coefficients
 The effectiveness of a combination was evaluated
by the number of actives in the 10% top positions
of the fused ranking
43

Advantages Data Fusion:
 Fusion of rankings can provide a small, but consistent,
improvement in the effectiveness of searching if an
appropriate combination of coefficients is chosen
 Data fusion of this sort provides a simple, but highly
cost-effective, way of enhancing existing systems for
chemical similarity searching
44

Advantages of using Reduced Graph
 Highly effective at finding compounds that share the same
activity but that are based on different lead series.
 High-throughput screening
 Method is capable of analysing large data sets and can
deal with noisy data.(High speed )
45

Structure is not the sole factor for biological
activity
 Interactions with
environment
 Solvation effects
 Metabolism
 Time dependence
 More...
46

References
 Searching Databases of Three-Dimensional Structures , Chapter 6, pp 213-263 Martin Y.C.,
Willett P.
 Applications of Graph Theory in Chemistry (J. Chem. InJ Compur. Sci. 1985, 25, 334-343)
 Substructure Searching Methods: Old and New (J. Chem. Inf. Comput. Sci. 1993, 33, 532-538)
 Similarity Searching Using Reduced Graphs(J. Chem. Inf. Comput. Sci. 2003, 43, 338-345)
 Searching for Patterns of Amino Acids in 3D Protein Structures (J. Chem. Inf. Comput. Sci.
2003, 43, 412-421)
 Graph-Based Molecular Alignment (GMA),( J. Chem. Inf. Model. 2007, 47, 591-601)
 Chemical Similarity Searching (J. Chem. Inf. Comput. Sci. 1998, 38, 983-996)
 Scaffold Hopping Using Clique Detection Applied to Reduced Graphs (J. Chem. Inf. Model.
2006, 46, 503-511)
 Further Development of Reduced Graphs for Identifying Bioactive Compounds (J. Chem. Inf.
Comput. Sci. 2003, 43, 346-356)
 Maximum common subgraph isomorphism algorithms for the matching of chemical structures
(J. Comp. Aided. Mol. Des, 16: 521–533, 2002)
 A Robust Clustering Method for Chemical Structures (J. Med. Chem. 2005, 48, 4358-4366)
 Automatic Identification of Molecular Similarity Using Reduced-Graph Representation of
Chemical Structure (J. Chem. Inf: Comput. Sci. 1992, 32, 639-643)
47

Application of graph theory in drug design

More Related Content

What's hot (20)

Similar to Application of graph theory in drug design (20)

Recently uploaded (20)

Application of graph theory in drug design