uQlust

uQlust performs ultrafast ranking and clustering of macromolecular structures to enable efficient analysis and comparison of protein and RNA structural models.

Key Features:

• Profile Hashing: Utilizes structural profiles to enable efficient clustering with a reduced memory footprint.
• Linear-Time Algorithm: Implements a linear-time comparison algorithm to perform rapid implicit comparisons across model pairs.
• Structural Profiles: Employs structural profiles for proteins and nucleic acids to represent models for comparison.
• Implicit All-Pairs Comparison: Facilitates implicit comparison across all pairs of models within large datasets without explicit pairwise computations.
• Fragment-Based Profiles: Supports clustering of structures of arbitrary length by integrating fragment-based profiles.
• Length Versatility: Handles ranking and clustering for model sets of identical or varying lengths.
• Reduced Complexity and Memory Usage: Lowers computational complexity and memory requirements compared to traditional pairwise methods.
• Hierarchical Clustering Capability: Enables hierarchical clustering of extensive structural datasets.

Scientific Applications:

• Structural classification: Comparison and classification of protein and RNA macromolecular structures.
• Model ensemble analysis: Ranking and clustering of large sets of predicted models and molecular simulation ensembles.
• Large-scale PDB analysis: Hierarchical clustering and exploration of entire databases such as the Protein Data Bank (PDB).

Methodology:

Combines profile hashing and structural and fragment-based profiles with a linear-time algorithm for implicit all-pairs comparison of models.