MOITF-EM

MOITF-EM identifies conserved structural motifs and non-conserved regions within cryogenic electron microscopy (cryoEM) volumetric maps to compare macromolecular assemblies and assist atomic model construction.

Key Features:

• Fully Automated Identification: Automatically detects conserved structural elements within cryoEM volumetric maps, pinpointing regions that remain conformationally stable across different functional states of macromolecular assemblies such as chaperonins and viruses.
• Detection of Non-conserved Regions: Highlights non-conserved areas within structures to indicate local molecular flexibility associated with specific biological activities or functional states.
• Rotationally Invariant Low-dimensional Representations: Constructs rotationally invariant, low-dimensional representations of local map regions inspired by advancements in 2D object recognition.
• Simple Metric Comparison: Compares the reduced representations across input maps using a simple metric to establish correspondences between local regions.
• Clustering and Graph Theory Application: Clusters correspondences using hash tables and applies graph theory to extract conserved structural domains or motifs.
• Atomic Model Construction: Enables construction of atomic models for certain cryoEM maps based on identified conserved regions.

Scientific Applications:

• Conserved domain extraction from large assemblies: Demonstrated on macromolecular maps including viruses P22 and epsilon 15, the Ribosome 70S complex, and the GroEL chaperonin.
• Comparative structural analysis of bacteriophages: Identified conserved folds among double-stranded DNA bacteriophages HK97, Epsilon 15, and ô29 despite low sequence similarity.

Methodology:

Constructs rotationally invariant, low-dimensional representations of local regions (inspired by 2D object recognition), compares them across maps with a simple metric, clusters correspondences using hash tables, and applies graph theory to extract conserved domains or motifs.