Computational Chemistry in Drug Discovery 2

?
  • Created by: LBCW0502
  • Created on: 22-10-19 20:32
What is QSAR/QPSR? (1)
Quantitative Structure Activity/Property Relationship. A computational approach performing lead optimisation. Quantifying the relationship between physicochemical properties and biological activity
1 of 47
What is QSAR/QPSR? (2)
Relating molecular structure to pharmacological/biological activity (QSAR). Relating molecular structure to physical properties (QSPR). Compounds with biological activity/new compounds with improved biological activity
2 of 47
Why is QSAR used? (1)
There are too many possible combinations. The required number of compounds to synthesise in order to place 10 different groups in 4 positions of benzene ring (carbons) is 10^4
3 of 47
Why is QSAR used? (2)
Solution would be to synthesise and test all the combinations (possible but not rational). Use a small number of synthesised compounds against a certain target (train set)
4 of 47
Why is QSAR used? (3)
And deriving rules from data to predict biological activity of the rest of the compounds (test set) using QSAR approach
5 of 47
Why is QSAR used? (4)
QSAR gives the medicinal chemists some levels of prediction and enables them to predict the biological activity of a novel analogue in advance. Cut down the number of analogues/congeneric series which have to be made
6 of 47
What are the QSAR steps? (1)
Design diverse analogues by modification of lead to obtain the structures of active compounds. Calculate various physicochemical or structural parameters (descriptors). Quantify correlation of descriptors and biological activity by regression methods
7 of 47
What are the QSAR steps? (2)
MLR, PLS. Drive a mathematical equation relating structure and activity using statistical techniques (Hansch and Free-Wilson analysis). Use equation to predict biological activity of designed compounds. Synthesise favoured compounds
8 of 47
What is the goal of QSAR?
To find a good statistical correlation between an activity and a molecular property
9 of 47
What are molecular descriptors? (1)
One or more parameters that explain intrinsic properties such as chemical, physical and biological in a group of analogues (related or congeneric compounds). Descriptors are numerical values that characterise properties of molecules (e.g. mwt, log P)
10 of 47
What are molecular descriptors? (2)
For the definition of molecular descriptors, a knowledge of algebra, graph theory, information theory, computational chemistry, theories of organic reactivity and physical chemistry is usually required
11 of 47
What are the categories of molecular descriptors? (1)
Spatial (molecular volume), electronic (sum of formal charges), thermodynamic (heat formation), conformational (energy), topological (molecular flexibility index), information-content (graph-theoretical indices)
12 of 47
What are the categories of molecular descriptors? (2)
Quantum mechanical (lowest unoccupied molecular orbital). Structural (mwt). Descriptors on fragment constants. Descriptors based on molecular field analysis (MFA). Descriptors based on molecular shape analysis (MSA) - RMS to shape reference
13 of 47
What are the characteristics of good molecular descriptors? (1)
Structural interpretation. A good correlation with at least one property. No trivial correlation with other molecular descriptors. Gradual change in its values with gradual changes in molecular structure
14 of 47
What are the characteristics of good molecular descriptors? (2)
Not restricted to a too small class of molecules. Preferably - some discrimination power among isomers, not trivially including in the definition other molecular descriptors, allowing reversible decoding (back from descriptor value to structure)
15 of 47
What are the types of molecular descriptors?
0D, 1D, 2D, 3D, 4D
16 of 47
Describe features of 0D descriptors or count descriptors (1)
No information about molecular structure and atom connectivity is needed. Atom and bond counts. Sum or average of atomic properties are typical of this class of descriptors. Naturally interpreted. No need to optimise molecular structure
17 of 47
Describe features of 0D descriptors or count descriptors (2)
Independent of any conformational problem. Information content is low. Have equal values for several molecules such as isomers. E.g. mwt
18 of 47
Describe features of 1D descriptors or fingerprints
Calculated from sub-structural information. Counting of functional groups and sub-structure fragments. Often are presented as fingerprint - a binary vector where 1 indicates the presence of defined sub-structure and 0 its absence
19 of 47
Describe features of 2D or topological descriptors (1)
Based on graph representation of molecule. Molecular graph or chemical graph is a representation of structural formula of a chemical compound in terms of graph theory
20 of 47
Describe features of 2D or topological descriptors (2)
They can be sensitive to one or more structural features of the molecule (size, shape, symmetry, branching, cyclicity). Can encode chemical information concerning atom type and bond multiplicity
21 of 47
Describe features of 2D or topological descriptors (3)
Two categories - topo-structural indices, encode only information on adjacency and distance of atoms in molecular structure
22 of 47
Describe features of 2D or topological descriptors (4)
Topo-chemical indices quantify information on topology but also specific chemical properties of atoms such as their chemical identity and hybridisation state
23 of 47
Describe features of 2D or topological descriptors (5)
If 2D structure changes there are new properties sensitive to structural features e.g. change in atom type
24 of 47
Describe features of 3D or geometrical descriptors (1)
Derived from a geometrical representation of the molecule (x,y,z, Cartesian co-ordinates of atoms of the molecule). Usually provide more information/discrimination power also for similar molecular structures
25 of 47
Describe features of 3D or geometrical descriptors (2)
And molecule conformations than topological descriptors. High information content. Require geometry optimisation. For flexible molecules, several molecule conformations are available. Identify different isomers, may have more available conformers
26 of 47
Describe features of 4D or grid-based descriptors (1)
Quantitatively identify and characterise interactions between a molecule and receptor's active site. Placing molecules in a 3D lattice and use a probe (steric, electrostatic, hydrophilic etc.)
27 of 47
Describe features of 4D or grid-based descriptors (2)
to map the surface of the molecule on the basis of molecule interactions within the probe. Generating new approach of 3D QSAR 0/1/2/3/4D), 2D QSAR (0/1/2/3)
28 of 47
Describe features of statistical techniques on QSAR equation (1)
Free energy models (Hansch analysis - physicochemical properties and biological activities of molecules, considers descriptors)
29 of 47
Describe features of statistical techniques on QSAR equation (2)
Mathematical models (Free-Wilson analysis - biological activities of molecules, doesn't consider descriptors, considers certain functional groups)
30 of 47
Describe features of the Hansch analysis (1)
Equation to predict activity of compound based on descriptor. Able to apply more than one descriptor at a time. Biological activity, minimum effective dose. Descriptors - electronic characteristics, hydrophobicity
31 of 47
Describe features of the Hansch analysis (2)
log (1/C) = k1 log P - k2 (log P_)^2 + k3 x sigma + k4. Hansch equation relates biological activity to common physicochemical properties and considers more than one physicochemical property at the time
32 of 47
Describe features of the Hansch analysis (3)
Minimum effective dose (C). Octanol-water partition coefficient (P). Hammett substituent constant (sigma). Constants derived from regression (Kx)
33 of 47
Describe features of Free and Wilson analysis (1)
Additive model - numerical method which directly relates structural features with biological properties. Considers structural features not physicochemical features of the compound. log BAi = sum of aXmu
34 of 47
Describe features of Free and Wilson analysis (2)
BA (biological activity values). Number of different positions (j). Number of different substituents in each position (i). Biological activity of substituent R in each position j (a). Overall mean of biological activity values (mu)
35 of 47
Describe features of Free and Wilson analysis (3)
1 when substituent R is present in position (X)
36 of 47
Describe features of Free-Wilson vs Hansch analysis (1)
Quite different. Free-Wilson only interested in attributing incremental values to all different groups and substituents, Hansch model interprets activity contributions in physicochemical terms
37 of 47
Describe features of Free-Wilson vs Hansch analysis (2)
They are fundamental related - start from additivity concept of group contributions to biological activity
38 of 47
Describe features of regression methods and correlation coefficient (1)
Experimental (observed) and computational (predicted) data are needed to calibrate the regression line. Computation y = co + C1X + C2X2 + C3X3 etc. Diverse approaches (Multiple Linear Regression, MLR, Partial Least Squares, PLS)
39 of 47
Describe features of regression methods and correlation coefficient (2)
Correlation coefficient (r squared) indicator to show fitness among the parameters. MLR and PLS determined by number of descriptors applied in calculations
40 of 47
Describe features of regression methods and correlation coefficient (3)
If there are too many descriptors in the calculations use PLS. If there are too few different descriptors, predict activity using MLR. Correlation coefficient (closer to 1, better value, not close to 1, due to errors)
41 of 47
Describe features of the QSAR chart (1)
In vitro, synthesised compounds against certain targets (experimental/observed activity). In silico (database 3D structures). Train set (80% compounds, generate descriptors pool, reduce descriptors to few numbers, (1 for first 20, 1 for every 5)
42 of 47
Describe features of the QSAR chart (2)
If you consider many descriptors with few compounds – leads to false correlation/results
43 of 47
Describe features of the QSAR chart (3)
PLS and MLR apply regression method, generate mathematical model (local/global QSAR, predict activities with calibration correlation >0.55). Test set (20% compounds) - predict activities of test and train sets using model, prediction correlation >0.7
44 of 47
Describe features of the QSAR chart (4)
Contour map (design new novel compounds with same or diverse substructure). Predict activities of designed compounds using model. If error in values, change train/test set or change descriptors
45 of 47
Describe features of a contour map (1)
Visualise mathematical model of 3D QSAR. Define fields. Steric (bulky group increases activity, small group increases activity). Electrostatic - activity increased with electropositive/negative groups or substituents). Colour coding
46 of 47
Describe features of a contour map (2)
Determine where to put functional groups in order to increase the activity of the compound e.g. bulky groups, small groups, electropositive groups, electronegative groups
47 of 47

Other cards in this set

Card 2

Front

What is QSAR/QPSR? (2)

Back

Relating molecular structure to pharmacological/biological activity (QSAR). Relating molecular structure to physical properties (QSPR). Compounds with biological activity/new compounds with improved biological activity

Card 3

Front

Why is QSAR used? (1)

Back

Preview of the front of card 3

Card 4

Front

Why is QSAR used? (2)

Back

Preview of the front of card 4

Card 5

Front

Why is QSAR used? (3)

Back

Preview of the front of card 5
View more cards

Comments

No comments have yet been made

Similar Pharmacy resources:

See all Pharmacy resources »See all Computational Chemistry in Drug Discovery 2 resources »