FELIX is dynamic, super fast software for searching and exploring non-covalent interactions in protein-ligand data.
FELIX matches query fingerprints against data derived from protein-ligand crystal structure data stored in the Proasis database.
FELIX searches are very fast, typically taking less than a few seconds.
A unique feature of FELIX is the sorting of hits by atom-based binding affinities. This is achieved by storing affinity results from DesertSci’s proprietary Scorpion software, for computing atom based ligand affinity scores.
For searching interactions, DesertSci’s proprietary View Contacts software is used to find and classify all close-contacts between ligands and proteins, which are stored as vectors to facilitate fast matching. FELIX also stores information about the kinase ATP binding site into its vectors, enabling kinase specific queries. Additionally, all pair-wise distances between residues around a ligand are computed and stored, enabling residue-residue distance constraints to be included into queries.
FELIX offers a broad range of powerful search possibilities, including:
- Search all sites with a specific interaction type
- Search all sites with a set of interactions, with or without distance constraints between the residues
- Search all sites with grouped interaction types, for example, search all favourable polar contacts to any sidechain
- Search all sites and sort hits according to Scorpion scores
- Search all kinases for interactions involving specific residues in the ATP site
- Search all sites with a specific interaction type and create a histogram showing binned distances
- Similarity search all sites based on all residues around a ligand
- Similarity search all sites based on residues with a favourable interaction with ligand, with a similarity search cutoff as an additional constraint
- Similarity search all sites based on interaction types and residue types
- Similarity search all sites using a sub-pocket of residues, with interaction types as additional constraints
