Introduction to Pharmacophore Modeling and Feature Trees (2.5D)

Definition : IUPAC defines "pharmacophore is the ensemble of steric and electronic features that is necessary to ensure the optimal supramolecular interactions with a specific biological target structure and to trigger (or to block) its biological response".A pharmacophore does not represent a real molecule or a real association of functional groups, but a purely abstract concept that accounts for the common molecular interaction capacities of a group of compounds towards their target structure.

In drug design, the term 'pharmacophore‘ refers to a set of features that is common to a series of active molecules.

Monty Kier, using Roald Hoffmann’s extended–Hückel–theory quantum mechanics package in a study with muscarinic agonists, was the first to calculate a pharmacophore; he called it a ‘proposed receptor pattern’.You can read about the origin of pharmacophore concept from a paper by John H. Van Drie .

Pharmacophore is defined in terms of atoms or centers which can interact with the receptor with different interaction centers as follows:
  • Hydrogen bond donors
  • Hydrogen bond acceptors
  • Positive charge centers
  • Aromatic ring centers
  • Hydrophobic centers

Atom is acceptor if it can attract an hydrogen (nitrogen, oxygen or sulfur and not an amide nitrogen, aniline nitrogen and sulfonyl sulfur and nitro group nitrogen), and donor if it can give an hydrogen.



3D Pharmacophores
• A three-dimensional pharmacophore specifies the spatial relation-ships between the groups.
• Expressed as distance ranges,angles and planes
• A commonly used 3D pharmacophore for antihistamines contains two aromatic rings and a tertiary nitrogen given below .
Image1.jpeg Taken from Laak etal. J Med Chem 1995,38(17)

The the workflow of Ligand based Pharmacophore design in virtual screening is presented below.
Pharmacophore modeling and Virtual Screening Workflow

To be recognized as a useful tool,a pharmacophore model has to provide a valid information for the medicinal chemist exploring structure-activity relationships

1. It has to highlight the functional groups involved in the interaction with the target.Pharmacophore-based queries allow to find novel drug candidates with different scaffolds('scaffold Hopping') and functional groups than the original ligands used for the modeling of the pharmacophores.The objective is to identify compounds with a different core group scaffold but essential functional group points are preserved.
2. The second criterion for a valid pharmacophore model is that it should discriminate stereoisomers. Stereospecificity is one of the principal attributes of pharmacological receptors and a perfect stereochemical complementarity between the ligand and the binding-site protein is an essential criterion for high affinity and selectivity.

Validation of the pharmacophore model:

The generated pharmacophore model should be statistically significant, and should identify active compound from a database. Therefore, the derived pharmacophore was validated using statistical methods like sensitivity ,specificity,Enrichment score and Receiver Operating Characteristic (ROC) analysis
enrichment factor (EF) and Güner-Henry (GH) Scoring method.The details are given below:

1) %yieldofactives(Ya)=TP/ (TP+FP)×100
2) %Actives=TP/A×100
3) Sensitivity(Se)=TP/TP+FN
4) Specificity(Sp)=TN/TN+FP
5) GH Score =(3/4.Ya+1/4.Se)Sp
6) EnrichmentFactor(EF)=Ya/(A/D)

TP is the number of true positives returned after screening the database, TN, is the number of true negatives, FP, is the number of false positives ,FN, is the number of false negatives and A , is the total number of actives in the database D, is the total number of database compounds

For more details in pharmacophore modeling you can read Pharmacophores and Pharmacophore Searches and Pharmacophore Perception, Development, and Use in Drug Design.

Challenges in Pharmacophore modelling:
  1. The first challenging problem is the modeling of ligand flexibility. Currently, two strategies have been used to deal with this problem: the first is the pre-enumerating method, in which multiple conformations for each molecule are precomputed and saved in a database.The second is the on-the-fly method, in which the conformation analysis is carried out in the pharmacophore modeling process. The first requires less computing cost and the second requires less storage space but high CPU time.
  2. The second challenging issue in ligand based pharmacophore modeling is alignment of Molecules.The alignment methods can be classified into two categories in terms of their fundamental nature point-based and property-based approaches.The points (in the point-based method) can be further differentiated as atoms, fragments or chemical features. In point-based algorithms,pairs of atoms, fragments or chemical feature points are usually superimposed using a least-squares fitting. The biggest limitation of these approaches is the need for predefined anchor points because the generation of these points can become problematic in the case of dissimilar ligands.The property-based algorithms make use of molecular field descriptors, usually represented by sets of Gaussian functions, to generate alignments.
  3. The third challenging problem lies in selection of training set compounds.The size of the dataset and its chemical diversity affect the final generated pharmacophore model noticeably.In some cases, completely different pharmacophore models of ligands interacting with the same macromolecular target could be generated from the same algorithm and program that uses different training sets.An example discussed in the here .

Feature Trees

Molecular descriptors are used for retrieval of compounds and also for clustering and and property prediction.Most descriptors use today are in linear format such as the properties are stored in the form of a vector.
The alignment free approach of comparison is extremely fast but it has disadvantages i.e the relative arragnment of funtional groups on the molecular surface cannot be determined and its weakly described in linear descriptors.On the other hand 3D model can be itself considered as descriptor and they are aligned in 3D space , but its is difficult due to conformational flexibility and it might miss the right alignment.

Feature Trees (FTrees) is a mix of 2D and 3D ligand-based approach .Alignment based but conformation independent descriptor. A feature tree represents a molecule by a tree such that the tree should capture the major Building blocks of the molecule in addition to the overall alignment.
Reduced graph descriptors, like feature trees, are frequently applied in cases where the relative arrangement of functional groups is more important than exact substructure matches. The Feature Tree captures the pharmacophore information in a sufficiently fuzzy way to enable the identification of structurally dissimilar actives (i.e., scaffold hopping).


The three molecules above represent the same reduced graphs,in the figure above tetrazole is a bioisostere of carboxylic acid and squaric acid.

FTrees descriptors are as follows:
1) The shape descriptor has two components
- the number of atoms
- the approximated van der Waal’s volume

2) Chemical features are used to describe an interaction pattern, FTrees uses the FlexX interaction profile.

Ftrees pattern.png

More details on FTrees is here


Tools for Pharmacophore modelling and 3D searching
1) Catalyst (Accelrys)
2) Phase (Schrodinger)
3) LigandScout (Inte:Ligand)
4) PharmaGist
5) Pharmer

You can get a list of all the exciting cheminformatics tools links here

3D Shape based method ROCS (Rapid overlay of Chemical Structures)

We will be doing one assignment on pharmacophore modeling using LigandScout.

Using LigandScout to generate ligand pharmacophore model and perform virtual screening of database.