computer aided drug design (CADD) 1

the steps to discover new drugs

• target

• hit- confirm compound showing reproducible activity after orthogonal testing and interference controls

• lead series; chemically related compound with tractable SAR, relevant activity and early developability

• candiate; compone with an integrated pharmacology, exposure safety and pharmaceuticals profile

CADD contribution

• identify virtual hits

• explain SAR

• props analogues

• predict binding modes

How small molecules drugs work

• most drugs exert their therapeutic effect by interacting with specific macromolecules in the body. these molecular targets include the nucleic acids DNA and RNA but in most cases they are proteins

Right shape and right functional group

• in addition to conformational changes which optimise shape complementarity, drugs usually make a series of interaction with their targets

• right functional groups to make good interactions with our target

amino acids in binding pocket

• interaction occurs between the function groups of our drugs, and the amino acids forming the binding pocket

• amino acids: interact with ligands/ inhibitors with their R groups and with the connecting amide bond

gefitinib binding example

• binding affinity reflects the balance between favourable protein ligand interactions and energetic penalties associated with desolation, confiormational restriction and entropy

Quantitative structure activity relationships (QSAR)

• mathematical method which put in relation the biological activity of a drug with its structural physical and chemical properties

• activity is commonly transformed to a logarithmic scale

• using a training set of compound for which the biological activity is measured together with some measured description

A simple QSAR workflow

• collect consistent chemical structure and activity data

• calculate molecular descriptiors

• divide data into training and test sets

• build the model using the training set

• tune and validate without using the final test set

• evaluate predictive performance (test)

• define the applicability domain

• predict new, suitable compounds

• test predictions experimentally

pharmacophore modelling

• three dimensional arrangement of steric and electronic features required for optimal interactions with specific biological target and for producing or blocking its biological response

virtual screening

• high performance computing used to analyse large databases of chemical compounds to identify possible drug candidates

• virtual databases is using a pharmacophoric model: only those compounds with the right functionalities to match the pharmacophore will be identified as virtual hits

virtual screening can

Prioritise compounds for experimental testing

• Reduce the number tested

• Explore larger chemical libraries

• Identify alternative chemical scaffolds

• Enrich a library in likely actives

Shape based similarity search

• compares the shape and properties of a known ligand of a biological target with the shape and properties of a series of small molecules: evaluating

  • 2D similarities

  • 3d similarities

shape vs pharmacophore

Pharmacophore: focuses on selected essential interaction features.

 Shape similarity: considers overall molecular volume and sometimes

electrostatics.

 Combined search: requires compatible shape and interaction features.