JEDi

JEDi analyzes protein conformational dynamics by applying principal component analysis to atomic trajectories from molecular dynamics and geometric simulations to quantify essential motions and correlated residue dynamics.

Key Features:

• Dual PCA modes: Implements Cartesian-based PCA (cPCA) and internal distance pair PCA (dpPCA) for analysis of atomic trajectories.
• Matrix construction: Constructs covariance (Q), correlation (R), and partial correlation (P) matrices from cPCA and dpPCA coordinates.
• Statistical enhancements: Applies shrinkage, outlier thresholding, and sparsity thresholds to improve covariance estimation and identify latent correlated motions.
• Hierarchical PCA: Performs local cPCA per residue to derive eigenresidues followed by global PCA on eigenresidues for large-scale motion description.
• Residue–residue coupling mapping: Generates residue–residue dynamical coupling maps by performing local cPCA on residue pairs.
• Kernel PCA: Provides kernel PCA for nonlinear dimensionality reduction of conformational data.
• Subspace comparisons: Quantifies similarity or overlap of eigenvector subspaces using various statistical metrics.
• Free energy landscapes: Computes free energy landscapes for both cPCA and dpPCA to assess energetic profiles of conformational states.
• Outputs and formats: Produces PNG images, text files with aligned coordinates, mobility metrics, PCA modes with eigenvalues, displacement vector projections, PyMOL scripts, and PDB files for mode visualization.
• Spatial and subregion analysis: Supports analysis at multiple spatial resolutions and within specific subregions, including multi-chain proteins.

Scientific Applications:

• Comparative essential dynamics: Comparative studies of essential dynamics across related biopolymers using multivariate statistical metrics.
• Functional mechanism identification: Identification and quantification of mobility and dynamic correlations to inform hypotheses about molecular function.
• Dynamical coupling analysis: Mapping residue–residue dynamical couplings to link correlated motions with structural features.
• Energetic interpretation: Use of free energy landscapes from cPCA and dpPCA to interpret energetic profiles associated with conformational changes.

Methodology:

Applies cPCA (Cartesian) and dpPCA (distance pair) to atomic trajectories from molecular dynamics or geometric simulations; constructs covariance (Q), correlation (R), and partial correlation (P) matrices; applies shrinkage, outlier thresholding, and sparsity thresholds; performs local cPCA per residue and per residue pair to derive eigenresidues and residue–residue coupling maps, performs global PCA on eigenresidues, implements kernel PCA, conducts subspace comparisons using statistical metrics, and computes free energy landscapes, with outputs including projections, eigenvalues, aligned coordinates, PyMOL scripts, and PDB files.