combining atomic-level molecular dynamics with coarse-grained monte-carlo dynamics
DESCRIPTION
Combining atomic-level Molecular Dynamics with coarse-grained Monte-Carlo dynamics Andrzej Koliński Laboratory of Theory of Biopolymers, Faculty of Chemistry, University of Warsaw http://www.biocomp.chem.uw.edu.pl. Bioinformatics 2013 / BIT13, 26-29 June 2013, Toruń, Poland. - PowerPoint PPT PresentationTRANSCRIPT
Combining atomic-level Molecular Dynamics with Combining atomic-level Molecular Dynamics with coarse-grained Monte-Carlo dynamicscoarse-grained Monte-Carlo dynamics
Andrzej Koliński
Laboratory of Theory of Biopolymers, Faculty of Chemistry, University of Warsaw
http://www.biocomp.chem.uw.edu.pl
Bioinformatics 2013 / BIT13, 26-29 June 2013, Toruń, Poland
All-atom MD with explicit water
• Atomic-Level Characterization of the Structural Dynamics of Proteins, Science, 2010• How Fast-Folding Proteins Fold, Science, 2011
1 milisecond simulations
ANTON - David E. Shaw group
Different all-atom force-fields (explicit water) are: - able to fold a protein into its native tertiary structure - inconsistent in the description of a folding pathway
Simulations of near-native dynamics seem to be essentially force-field independent.
M. Rueda, C. Ferrer-Costa, T. Meyer, A. Perez, J. Camps, A. Hospital, J. L. Gelpi, M. Orozco, A consensus view of protein dynamics Proc. Natl. Acad. Sci. U.S.A. 104:796−801, 2007
Coarse-grained models
Coarse-grained models of moderate resolution (~102 faster than all-atom MD)
LatticeKolinski et al.
ContinuousBaker et al.Liwo et al.
CABS model
Force field
Short range conformational propensities
Context-dependent pairwise interactions of side groups
A model of main chain hydrogen bonds
Interaction parameters are modulated by the predicted secondary structure and account for complex multibody interactions and the averaged effect of solvent
Sampling – Monte Carlo dynamics
A. Kolinski, Protein modeling and structure prediction with a reduced representation Acta Biochimica Polonica 51:349-371, 2004
Reconstruction & optimization procedure
protein backbone reconstruction
side chain reconstruction
all-atom minimization step
All-atom MD (A – Amber, C – Charmm, G – Gromos and O – OPLS/AA force-fields) is consistent with CABS stochastic dynamics (after a proper renormalizations) at short time-scales (10 ns)
M. Jamroz, M. Orozco, A. Kolinski, S. Kmiecik, A Consistent View of Protein Fluctuations from All-atom Molecular Dynamics and Coarse-Grained Dynamics with Knowledge-based Force-field, J. Chem. Theory Comput. 9:19–125, 2013
J. Wabik, S. Kmiecik, D. Gront, M. Kouza, A. Kolinski, Combining Coarse-Grained Protein Models with Replica-Exchange All-Atom Molecular Dynamics, International Journal of Molecular Sciences 14:9893-9905, 2013.
Protein dynamics
CABS models
reconstructed all-atom models (AMBER)
Kmiecik, D. Gront, M. Kouza, A. Kolinski, From Coarse-Grained to Atomic-Level Characterization of Protein Dynamics: Transition State for the Folding of B Domain of Protein A, J. Phys. Chem. B 116:7026-7032, 2012
Dynamics: CABS and all-atoms MD
Example of residue fluctuation profiles
Benchmarks summary
Test set (10 ns trajectories)
Compared data Avg. Spearman’s corr. coeff. between residue
fluctuation profiles
22 proteins (each one by 4 different force fields)
MD vs. CABS 0.70(J Chem Theory Comput, 2013)
393 non-redundant proteins (Amber force field)
MD vs. CABS 0.70(Nucl Acid Res, 2013)
140 non-redundant and NMR solved proteins (Amber force field)
NMR vs. CABS 0.72 (yet unpublished )
NMR vs. MD 0.65
MD vs. CABS 0.69
http://biocomp.chem.uw.edu.pl/CABSflex
PDB: 1BSN, F1-ATPase subunit, 138 AA
CABS-flex
PDB: 1BSN, F1-ATPase subunit, 138 AA
CABS-flex
PDB: 1BHE, polygalacturonase, 376 AA
CABS-flex
CABS-fold: server for protein structure prediction
http://biocomp.chem.uw.edu.pl/CABSfold
CABS in structure prediction
M. Blaszczyk, M. Jamroz, S. Kmiecik, A. Kolinski, CABS-fold: server for the novo and consensus-based prediction of protein structure, Nucleic Acids Research, 2013
Structure prediction (de-novo)
The predicted models (colored in rainbow) are superimposed on native structures (colored in magenta)
Modeling accuracy could be highly improved when combined with compartive modeling.
A. Kolinski, J. M. Bujnicki, Generalized protein structure prediction based on combination of fold-recognition with de novo folding and evaluation of models, Proteins 61(S7):84-90, 2005
Structure prediction
(homology modeling)
CASP9 examples9th Community Wide
Experiment on theCritical Assessment of Techniques for Protein Structure Prediction
CABS – docking and interactionsSimulations of induced folding (binding) of intrisingly disordered protein pKIG
with KIX domain
CABS – docking and interactionsSimulations of induced folding (binding) of intrisingly disordered protein
pKIG with KIX domain
Summary:
CABS could be easily combined with all-atom Molecular Dynamics and used in studies of protein dynamics, interactions and structure prediction
LTB servers based on CABS tools:
URL: http://biocomp.chem.uw.edu.pl/CABSfold URL: http://biocomp.chem.uw.edu.pl/CABSflex
M. Jamroz, A. Kolinski & S. Kmiecik, CABS-flex: server for fast simulation of protein structure fluctuations, Nucleic Acids Research, 1-5, 2013
M. Blaszczyk, M. Jamroz, S. Kmiecik, A. Kolinski, CABS-fold: server for the novo and consensus-based prediction of protein structure, Nucleic Acids Research 1-6, 2013
Thank you!
Co-authors: Drs. Sebastian Kmiecik, Michał Jamróz,Dominik Gront, Maciej Błaszczyk, Mateusz Kurciński, Jacek Wabik and others ….