Introduction to Ligand-Based Drug Design Chimica Farmaceutica
Identification of Active Ligands Identification of Suitable Descriptors (molecular fingerprint) Establish Mathematical Expression Relating Descriptors to Activity Construction and Validation of the QSAR model Introduction to Ligand-Based Drug Design Chimica Farmaceutica 3
Introduction to Ligand-Based Drug Design Chimica Farmaceutica 4
Introduction to Ligand-Based Drug Design Chimica Farmaceutica 5
Chimica Farmaceutica Introduction to Ligand-Based Drug Design 6 Squared Correlation Coefficient R or r Cross-Validated R Q or q = = = N i i N i i calc i Y Y Y Y r 1 exp, 1, exp, ) ( ) ( 1 = = = N i i N i i pred i Y Y Y Y q 1 exp, 1, exp, ) ( ) ( 1
( Y Squared Correlation i= 1 r = 1 Coefficient R or r N ( Y 0 Fitting r 1 N i= 1 exp, i exp, i Y calc, i Y ) ) r ESS TSS 1 RSS TSS i= 1 = N 1 N ( Y i= 1 exp, i ( Y exp, i Y calc, i Y ) ) Introduction to Ligand-Based Drug Design Chimica Farmaceutica 7
CV (Cross-Validation) Cross-validated R Q or q q N i= 1 = 1 N ( Y i= 1 exp, i ( Y exp, i Y pred, i Y ) ) SDEP N i= 1 = ( Yexp, Y, ) i N pred i q 1 The predictive ability of a model is estimated using a reduced set of structural data Introduction to Ligand-Based Drug Design Chimica Farmaceutica 8
Introduction to Ligand-Based Drug Design Chimica Farmaceutica 9
Y-Scrambling A statistical test of prediction tools, in which models are fitted for randomly reordered property/activity values and compared with the model obtained for the actual property/activity values. Anew model newis obtained model for such permuted isdata, Robtained and Q are then recalculated. for such This step is repeated for a sufficient number of times (iterations): permuted data, R and Q are then a good number being 50 to 100. recalculated. Values obtained in the above fashion are compared with the true values obtained for the model that was fitted on the real data. This step is repeated for a sufficient number of times (iterations): a good number being 50 to 100. Original X block Scrambled Y vector Scrambling Original Y vector Values obtained in the above fashion are compared with the true values obtained for the model that was fitted on the real data. Introduction to Ligand-Based Drug Design Chimica Farmaceutica 10
External Test-Set SDEP (Standard Deviation Error of Prediction) SDEP N = i= 1 ( Y Y ) exp, i pred, i N N of predicted compounds Introduction to Ligand-Based Drug Design Chimica Farmaceutica 11
Description of Molecules with Molecular Interaction Fields (MIF) Introduction to Ligand-Based Drug Design Chimica Farmaceutica 1
Reduction of Dimensionality into Few New Highly Informative Entities ----- Principal Components ----- Introduction to Ligand-Based Drug Design Chimica Farmaceutica 13
PLS + MIF CoMFA! Introduction to Ligand-Based Drug Design Chimica Farmaceutica 14
Physical properties are measured for the molecule as a whole Properties are calculated using computer software No experimental constants or measurements are involved Properties are known as Fields Steric field - defines the size and shape of the molecule Electrostatic field - defines electron rich/poor regions of molecule Hydrophobic properties are relatively unimportant Advantages over classical QSAR No reliance on experimental values (i.e. logp) Can be applied to molecules with unusual substituents Not restricted to molecules of the same structural class Improved predictive capability Introduction to Ligand-Based Drug Design Chimica Farmaceutica 15
Training Set Molecular Alignment Molecular Interaction Fields (MIF) Model Generation Internal and External Validation Introduction to Ligand-Based Drug Design Chimica Farmaceutica 16
Training Set Molecular lignment Molecular Interaction Fields (MIF) Model Generation Internal and External Validation Introduction to Ligand-Based Drug Design Chimica Farmaceutica 17
Training Set Molecular Alignment Molecular Interaction Fields (MIF) Model Generation Internal and External Validation Introduction to Ligand-Based Drug Design Chimica Farmaceutica 18
Training Set Molecular Alignment Molecular Interaction Fields (MIF) Model Generation Internal and External Validation Introduction to Ligand-Based Drug Design Chimica Farmaceutica 19
Training Set Molecular lignment Molecular Interaction Fields (MIF) Model Generation Internal and External Validation Introduction to Ligand-Based Drug Design Chimica Farmaceutica 0
Training Set Molecular Alignment Molecular Interaction Fields (MIF) Model Generation Internal and External Validation SDEP = ( y y ) exp / n 1 pred q ( ) y ( ) exp y pred / yexp = 1 y Introduction to Ligand-Based Drug Design Chimica Farmaceutica 1
Introduction to Ligand-Based Drug Design Chimica Farmaceutica
Introduction to Ligand-Based Drug Design Chimica Farmaceutica 3
Introduction to Ligand-Based Drug Design Chimica Farmaceutica 4
The Lattice Model: A General Paradigm for Shape-Related Structure/Activity Correlation Cramer, R.D., and Milne, M., Abstracts ACS Meeting, Honolulu, 1979, COMP 44. Baroni, M.; Costantino, G.; Cruciani, G.; Riganelli, D.; Valigi, R.; Clementi, S., Generating Optimal Linear Pls Estimations (Golpe) an Advanced Chemometric Tool for Handling 3D QSAR Problems. Quant Struct Act Rel 1993, 1, (1), 9 0. Introduction to Ligand-Based Drug Design Chimica Farmaceutica 5
Introduction to Ligand-Based Drug Design Chimica Farmaceutica 6
Introduction to Ligand-Based Drug Design Chimica Farmaceutica 7
Introduction to Ligand-Based Drug Design Chimica Farmaceutica 8
Introduction to Ligand-Based Drug Design Chimica Farmaceutica 9
Introduction to Ligand-Based Drug Design Chimica Farmaceutica 30
Introduction to Ligand-Based Drug Design Chimica Farmaceutica 31
Introduction to Ligand-Based Drug Design Chimica Farmaceutica 3
CoMFA 3-D QSAutoGrid/R Introduction to Ligand-Based Drug Design Chimica Farmaceutica 33
GRS GRS GRS Introduction to Ligand-Based Drug Design Chimica Farmaceutica 34
Introduction to Ligand-Based Drug Design Chimica Farmaceutica 35
Introduction to Ligand-Based Drug Design Chimica Farmaceutica 36
Introduction to Ligand-Based Drug Design Chimica Farmaceutica 37