Atom-Based Binding Site Clustering

DesertSci has developed a significant new tool that generates detailed reports of the structural similarities and differences within sets of closely related protein structures.

The atom-based binding site clustering tool was developed to provide medicinal chemists, modellers, crystallographers and other drug design experts with overviews that can deliver new insights into the relationship between ligand candidates and protein targets.  This  complex analysis and reporting module provides researchers with new insights for structure based drug design.

In 2019, our tool for atom-based binding site clustering will become a standard component of DesertSci’s Proasis4 software system.

For each query, two types of clustering reports are created:

• a summary report that includes colour-coding of residues for each property, and
• a full report that provides detailed information for every binding site residue in each overlaid structure.   
• visualisation of residue flexibility, based specifically on the selected set of overlaid binding sites, is also provided.

Specifically, our atom-based binding site clustering tool analyses sets of closely related protein structures loaded into Proasis4, and then generates comprehensive reports that include:

1. the overall similarity order for the set of binding sites, with a list starting with the most similar pair of structures, and ending with the most dissimilar outlier
2. structures grouped according to overall binding site similarity, with different groupings at different levels of similarity cut-offs, based on UPGMA hierarchic clustering
3. structures grouped according to ligand similarity, based on ligand fingerprints, Tanimoto similarity, and WARDS hierarchic clustering
4. residue number and sequence labelling
5. sequence variation
6. residue flexibility according to RMSD about each protein atom, separately for mainchain and sidechain atoms
7. residue locations with alternative conformers
8. residue locations with atoms with the highest relative b factors
9. atom-based cluster membership details, with separate reporting for mainchain and sidechain atoms based on density reachability and connectivity clustering

Reports can be created listing:

• only residues in close proximity to the ligand (suited to medicinal chemists) and
•  more distal residues (suited to modellers/crystallographers).

Our new Proasis4 atom-based binding site clustering module is ideal for a wide range of applications, including:

• researchers wanting a quick understanding of sidechain flexibility
• chemists wanting to see how the protein target adapts when different ligand scaffolds are introduced
• modellers wanting to know if a residue location is disordered or is adopting a small number of specific configurations
• structural biologists looking for loop movements
• crystallographers checking for high b factors in unexpected locations in their data sets
• drug designers wanting to know if they can add larger substituents to their current lead candidate.

Please contact us for more information on DesertSci’s atom-based binding site clustering module.