Algorithm¶

Pocketeer detects pockets in proteins using a simple, fast approach based on geometry.

How It Works¶

Pocketeer finds empty spaces (pockets) by looking for "alpha-spheres": spheres that fit between protein atoms without touching any atoms except the ones that define them.
It uses a mathematical method called Delaunay tessellation to divide atom coordinates into groups, making it easy to look for these spheres.
Only spheres of certain sizes (set by the user) and in specific locations are considered pockets.
Detected spheres are grouped into clusters (pockets) if they are close together.

Radius range (r_min, r_max): Sets the smallest and largest spheres to try; usually around 3-6 Å for most proteins.
Cluster size (min_spheres): Minimum number of spheres to call something a pocket.
Merge distance: How close spheres need to be to be grouped into a pocket.
Polarity probe radius: Determines if a sphere is inside the protein (buried) or at the surface.

Use Pocketeer for fast, simple pocket detection, easy integration in Python workflows, or algorithm customization.
For more advanced analysis or high-throughput work, tools like fpocket may be better.

Does not consider chemical features of pockets.
Estimates pocket volume simply, may miss complex cavities (although this is unlikely).
Analyzes one structure at a time.

Pocketeer aims for simplicity and useful results, making it easy to understand and adapt. Feel free to submit an issue or PR!