Learning probabilistic models of hydrogen bond stability from molecular dynamics simulation trajectories
KAUST DepartmentApplied Mathematics and Computational Science Program
Computer, Electrical and Mathematical Sciences and Engineering (CEMSE) Division
Extensions of Dynamic Programming, Machine Learning and Discrete Optimization Research Group
Online Publication Date2011-02-18
Print Publication Date2011
Permanent link to this recordhttp://hdl.handle.net/10754/325468
MetadataShow full item record
AbstractBackground: Hydrogen bonds (H-bonds) play a key role in both the formation and stabilization of protein structures. They form and break while a protein deforms, for instance during the transition from a non-functional to a functional state. The intrinsic strength of an individual H-bond has been studied from an energetic viewpoint, but energy alone may not be a very good predictor.Methods: This paper describes inductive learning methods to train protein-independent probabilistic models of H-bond stability from molecular dynamics (MD) simulation trajectories of various proteins. The training data contains 32 input attributes (predictors) that describe an H-bond and its local environment in a conformation c and the output attribute is the probability that the H-bond will be present in an arbitrary conformation of this protein achievable from c within a time duration ?. We model dependence of the output variable on the predictors by a regression tree.Results: Several models are built using 6 MD simulation trajectories containing over 4000 distinct H-bonds (millions of occurrences). Experimental results demonstrate that such models can predict H-bond stability quite well. They perform roughly 20% better than models based on H-bond energy alone. In addition, they can accurately identify a large fraction of the least stable H-bonds in a conformation. In most tests, about 80% of the 10% H-bonds predicted as the least stable are actually among the 10% truly least stable. The important attributes identified during the tree construction are consistent with previous findings.Conclusions: We use inductive learning methods to build protein-independent probabilistic models to study H-bond stability, and demonstrate that the models perform better than H-bond energy alone. 2011 Chikalov et al; licensee BioMed Central Ltd.
CitationChikalov I, Yao P, Moshkov M, Latombe J-C (2011) Learning probabilistic models of hydrogen bond stability from molecular dynamics simulation trajectories. BMC Bioinformatics 12: S34. doi:10.1186/1471-2105-12-S1-S34.
PubMed Central IDPMC3044290
The following license files are associated with this item:
Except where otherwise noted, this item's license is described as This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
- Insights on hydrogen-bond lifetimes in liquid and supercooled water.
- Authors: Martiniano HF, Galamba N
- Issue date: 2013 Dec 19
- Unconventional N-H…N Hydrogen Bonds Involving Proline Backbone Nitrogen in Protein Structures.
- Authors: Deepak RN, Sankararamakrishnan R
- Issue date: 2016 May 10
- Case study of hydrogen bonding in a hydrophobic cavity.
- Authors: Chen YC, Cheng CS, Tjong SC, Yin HS, Sue SC
- Issue date: 2014 Dec 18
- Statistical and molecular dynamics studies of buried waters in globular proteins.
- Authors: Park S, Saven JG
- Issue date: 2005 Aug 15
- Intra-protein hydrogen bonding is dynamically stabilized by electronic polarization.
- Authors: Duan LL, Mei Y, Zhang QG, Zhang JZ
- Issue date: 2009 Mar 21
Showing items related by title, author, creator and subject.
All-Atom Molecular Dynamics Simulation of Protein Translocation through an α-Hemolysin NanoporeDi Marino, Daniele; Bonome, Emma Letizia; Tramontano, Anna; Chinappi, Mauro (The Journal of Physical Chemistry Letters, American Chemical Society (ACS), 2015-07-16) [Article]© 2015 American Chemical Society. Nanopore sensing is attracting the attention of a large and varied scientific community. One of the main issues in nanopore sensing is how to associate the measured current signals to specific features of the molecule under investigation. This is particularly relevant when the translocating molecule is a protein and the pore is sufficiently narrow to necessarily involve unfolding of the translocating protein. Recent experimental results characterized the cotranslocational unfolding of Thioredoxin (Trx) passing through an α-hemolisin pore, providing evidence for the existence of a multistep process. In this study we report the results of all-atom molecular dynamics simulations of the same system. Our data indicate that Trx translocation involves two main barriers. The first one is an unfolding barrier associated with a translocation intermediate where the N-terminal region of Trx is stuck at the pore entrance in a conformation that strongly resembles the native one. After the abrupt unfolding of the N-terminal region, the Trx enters the α-hemolisin vestibule. During this stage, the constriction is occupied not only by the translocating residue but also by a hairpin-like structure forming a tangle in the constriction. The second barrier is associated with the disentangling of this region.
MDcons: Intermolecular contact maps as a tool to analyze the interface of protein complexes from molecular dynamics trajectoriesAbdel-Azeim, Safwat; Chermak, Edrisse; Vangone, Anna; Oliva, Romina; Cavallo, Luigi (BMC Bioinformatics, Springer Nature, 2014-05-06) [Article]Background: Molecular Dynamics ( MD) simulations of protein complexes suffer from the lack of specific tools in the analysis step. Analyses of MD trajectories of protein complexes indeed generally rely on classical measures, such as the RMSD, RMSF and gyration radius, conceived and developed for single macromolecules. As a matter of fact, instead, researchers engaged in simulating the dynamics of a protein complex are mainly interested in characterizing the conservation/variation of its biological interface. Results: On these bases, herein we propose a novel approach to the analysis of MD trajectories or other conformational ensembles of protein complexes, MDcons, which uses the conservation of inter-residue contacts at the interface as a measure of the similarity between different snapshots. A "consensus contact map" is also provided, where the conservation of the different contacts is drawn in a grey scale. Finally, the interface area of the complex is monitored during the simulations. To show its utility, we used this novel approach to study two protein-protein complexes with interfaces of comparable size and both dominated by hydrophilic interactions, but having binding affinities at the extremes of the experimental range. MDcons is demonstrated to be extremely useful to analyse the MD trajectories of the investigated complexes, adding important insight into the dynamic behavior of their biological interface. Conclusions: MDcons specifically allows the user to highlight and characterize the dynamics of the interface in protein complexes and can thus be used as a complementary tool for the analysis of MD simulations of both experimental and predicted structures of protein complexes.
Analysis and Ranking of Protein-Protein Docking Models Using Inter-Residue Contacts and Inter-Molecular Contact MapsOliva, Romina; Chermak, Edrisse; Cavallo, Luigi (Molecules, MDPI AG, 2015-07-01) [Article]In view of the increasing interest both in inhibitors of protein-protein interactions and in protein drugs themselves, analysis of the three-dimensional structure of protein-protein complexes is assuming greater relevance in drug design. In the many cases where an experimental structure is not available, protein-protein docking becomes the method of choice for predicting the arrangement of the complex. However, reliably scoring protein-protein docking poses is still an unsolved problem. As a consequence, the screening of many docking models is usually required in the analysis step, to possibly single out the correct ones. Here, making use of exemplary cases, we review our recently introduced methods for the analysis of protein complex structures and for the scoring of protein docking poses, based on the use of inter-residue contacts and their visualization in inter-molecular contact maps. We also show that the ensemble of tools we developed can be used in the context of rational drug design targeting protein-protein interactions.