Comparative analysis of p38 MAP kinase dynamics from experiments (PCA) and theory (ANM).
See the PNAS
article
or
figure
for details.
Comparative analysis of dynamics of drug target proteins and model systems from experiments (PCA) and theory (ANM).
See the Protein Science
article
for details.
Comparative analysis of p38 MAP kinase dynamics from experiments (PCA), simulations (EDA), and theory (ANM).
See the Protein Science
article
for details.
Animation shows HIV-1 reverse transcriptase functional motions calculated
using anisotropic network model.
Arrows and animations are generated using
NMWiz
VMD plugin. See
NMWiz tutorial
for usage examples.
You can make a quick protein representation in interactive sessions using
showProtein()
function.
NMWiz is designed for picturing normal modes easy. Image shows arrows from slowest three ANM modes for p38 MAP kinase centered at the origin.
They indeed align with planes normal to each other.
NMWiz makes depicting elastic network models and protein motions predicted with
them easy. Image shows ANM model for p38 MAP kinase and three slow ANM modes (below).
NMWiz can be used to comparative dynamics inferred from experimental datasets and predicted using theory.
The movie shows a molecular dynamics simulation
for assessing the druggability of kinesin eg5.
NMWiz
VMD plugin. See
NMWiz tutorial
for usage examples.
Kinesin Eg5 druggable sites, including allosteric inhibitor binding site and
and tubulin binding site, identified by simulations are shown. See our
publication for details.
Sampling of the functional substates (inward-facing (IF) or outward-facing (OF), in closed (c) or open (o) forms) of LeuT using coMD simulations. See publication for details.
The movie illustrates a coMD trajectory for adenylate kinase. NMWiz VMD plugin. See NMWiz tutorial for usage examples.
Energy landscape in the space of principal coordinates.
Outward-facing (OF) and inward-facing (IF) structures of GltPh show a large displacement of the core domains.
See publication for details.
The second mode of the OF structure moves all three transport domains simultaneously through the membrane in a ‘lift-like’ motion.
See publication for details.
The second mode of the IF structure moves all three transport domains simultaneously through the membrane in a ‘lift-like’ motion.
See publication for details.
Deformability profile of ubiquitin (PDB code: 1UBI). Structure is automatically uploaded to VMD program where blue color shows regions which are mechanically more resistant to the external force.
Mean value of effective spring constant (calculated from mechanical stiffness matrix) with secondary structure of ubiquitin. Blue color indicates mechanically strong regions.
Mechanical Stiffness Map with effective force constant in a color bar (blue - strong regions, red - weak regions) for ubiquitin.
Workflow for GNM analysis of chromatin dynamics.
See publication for details.
Covariance matrix of chromosome 17 of human B cells. Structural domains and CCDDs are identified and outlined.
See publication for details.
3D Laplacian embedding of chromosome 17 loci using the first three principal modes.
See ChromD tutorial for details.
Perturbation response scanning of the Hsp70 chaperone reveals interdomain allostery.
See publication for details.
Perturbation response scanning of the AMPA-type glutamate receptor reveals sensors and effectors for allosteric signaling.
See publication for details.
A more in-depth analysis of the PRS matrix reveals interdomain signaling in the AMPA receptor.
See publication for details.
The left panels show the three softest GNM modes (blue lines) and their standard deviations (faint blue bands).
Red and blue regions in the corresponding ribbon diagrams show regions moving in opposite directions.
The right panel has the average cross-correlation matrix from the first 20 global modes (top) and its standard deviation (bottom).
Square fluctuations calculated from the top 5 global modes are shown for a number of LeuT fold family
members, revealing similarities and subfamily- or conformation-dependent differences.
Type-I periplasmic binding protein domains are mapped onto the first two signature ANM modes.
These domains, found in a range of proteins including bacterial solute carriers and eukaryotic receptors,
have two lobes that undergo well-characterised conserved motions that are evident from comparison of structures.
SignDy reveals such conserved dynamics.
SignDy analysis allows a comparison of the frequency dispersion of family members.
The distribution of inverse eigenvalues is shown for the softest five modes for TIM barrel fold family.
Pharmmaker (center) includes four main programs (steps 2 to 5 of the pipeline)
that bridge druggability simulations from DruGUI (left)
with pharmacophore-based virtual screening (right).
The names of the programs are given under each step in blue.
(Left) A snapshot extracted from druggability simulations for an AMPAR LBD dimer
using Pharmmaker including probe poses and target conformation.
Dominant binding interactions between probe and residues are shown.
(Right) A pharmacophore model built based on the snapshot.
One hydrogen acceptor, one donor, and two hydrophobic features were used to represent the probes.
Each residue is given a binding value for each probe type, based on an inverse square distance potential.
This is shown in the two graphs for the two subunits of an AMPAR LBD dimer.
The dotted lines indicate a cutoff of 500, above which residues are defined as high affinity residues
for a particular probe.
ESSA profile (A) gives a measure of the extent of frequency shift in the global modes induced by each residue.
Residues (red circles) interacting with the allosteric ligand (PDB id: 2jfn) correspond to essential sites.
Two differentperspectives (B-C) display color-coded by z-scores from red (highest) to blue (lowest) together with bound ligands.
ESSA results for (A,D) muscarinic acetylcholine receptor and (B,C) free fatty acid receptor 1 GPR40.
GPCRs are color-coded by the ESSA profile. Various allosteric ligand binding sites,
as well as the G-protein (pink) binding site, correspond to essential or hot regions.
Complex structure (PDB id: 1pzo) resolved in the presence of two allosteric ligands
(spatially neighboring, both shown in magenta sticks) and the orthosteric ligand
(yellow sticks). The meshed surface displays the predicted allosteric pocket enclosing all ligands.
Superposition of the TRiC-AMP-PNP electron density map (EMD-1961; grey surface) and 8000 TRN nodes fitted to it (pink, spheres).
This number of nodes corresponds to 1 residue/node, but higher levels of coarse-graining such as 3000 nodes works too.
(A) Results from ANM analysis of the TRN (based on EMD-1961), displaying the architecture colour-coded by the MSFs of nodes (blue: most rigid; orange: most mobile) in the softest 20 modes.
(B) MSFs of the subunits as driven by the subsets of 5 (green), 10 (orange) and 20 (blue) softest modes. (C) Covariance between the global motions of the subunits based on the softest 20 modes.
(D) Orientational correlations between the global movements of the subunits.
The beads reconstructed from EMD-1961 are displaced along ANM mode 7 with an RMSD of 6 Å in both directions, revealing motions related to upper ring closure.
(A-C) Symmetric and anti-symmetric movements for circularly symmetric shapes. Gray and orange arrows indicate alternating motions between symmetric expansion and compression in A,
stretching and contraction along orthogonal directions in B, and opposite direction rotations in C. (D-F) Modes 6, 1 and 7 of CCT/TRiC (upper ring) approximate the above motions, respectively.
Population distribution of ClustENMD conformers shown on the angle space (LID-Core vs. NMP-Core angles) of adenylate kinase (AK), together with homologous experimental structures (black circles).
Independent 5-generation runs starting from open (4ake) and closed (1ake) states of AK highlight the major minima and the populated transition states.
Conformational surface of HIV-1 reverse transcriptase plotted along the first two principal components (PCs) obtained from experimental structures (black circles), onto which the ClustENMD conformers
(red circles, population levels in cyan) are projected.
This movie shows the population distribution for successive generations (gen-1 to gen-10) of conformers sampled starting from the open (blue levels; initial structure/black circle: 1tw7) and the closed
(red levels; initial structure/black diamond: 1bve) states of HIV-1 protease. As the distributions merge, they also cover the homologous experimental structures (gray circles).
ENM-Based Sampling of Essential Conformational Space at Full Atomic Resolution
ClustENMD is a hybrid method that performs unbiased conformational sampling of highly flexible and/or large biomolecular systems
(Kaynak et al., 2021; Kurkcuoglu et al., 2016)
with high computational efficiency. The major steps of the algorithm iteratively executed in each generation are:
(i) deforming each conformer (from previous generation) along random combinations of global ANM modes for generating new conformers,
(ii) clustering of new conformers to find cluster representatives, and
(iii) performing a short molecular dynamics (MD) simulation for each cluster representative using OpenMM software (Eastman et al., 2017).
Features
You can use ClustENMD for
- Unbiased sampling of conformers starting from a folded structure (experimental or modeled).
- Obtaining information on transition intermediates/paths connecting two states.
- Generating 2-D density distributions of conformers/energy landscapes.
- Comparison with experimental structures using principal component analysis.
- Generating conformers for ensemble docking purposes, as input for protein-ligand, protein-protein and protein-DNA docking
(Kurkcuoglu and Doruker, 2016;
Kurkcuoglu and Bonvin, 2020).
Reference
Kaynak BT, Zhang S, Bahar I, Doruker P
ClustENMD: Efficient sampling of biomolecular conformational space at atomic resolution
2021 Bionformatics (Accepted).
Kurkcuoglu Z, Bahar I, Doruker P
ClustENM: ENM-Based Sampling of Essential Conformational Space at Full Atomic Resolution
2016 J Chem Theory Comput 12(9):4549-4562.
Eastman P et al., OpenMM 7: Rapid development of high performance algorithms for molecular dynamics
2017 PLoS Comput Biol 13(7): e1005659.
Kurkcuoglu Z, Doruker P
Ligand Docking to Intermediate and Close-To-Bound Conformers Generated by an Elastic Network Model Based Algorithm for Highly Flexible Proteins
2016 PLoS One 11(6):e0158063.
Kurkcuoglu Z, Bonvin A
Pre‐ and post‐docking sampling of conformational changes using ClustENM and HADDOCK for protein‐protein and protein‐DNA systems
2020 Proteins 88(2):292-306.
Source Code
ProDy is open source
and you can contribute to its development in many ways.
See this guide for getting started.
Problems?
Let us know any problems you might have by opening an issue at the
tracker
so that we can make ProDy better.