| 1. |
- Brodmerkel, Maxim N.
(författare)
-
Theoretical and Biochemical : Advancing Protein Structure Investigations with Complementing Computations
- 2023
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Life as we know it today would not exist without proteins. The functions of proteins for us and other organisms are linked to their three-dimensional structures. As such, protein structure investigations are a crucial contribution for understanding proteins and the molecular basis of life. Some methods probe the structure of proteins in the gas phase, which brings various advantages as well as complications. Amongst them is mass spectrometry, a powerful method that provides a multitude of information on gaseous protein structures. Whilst mass spectrometry shines in obtaining data of the higher-order structures, atomistic details are out of reach. Molecular dynamics simulations on the other hand allow the interrogation of proteins in high-resolution, which makes it an ideal method for their structural research, be it in or out of solution.This thesis aims to advance the understanding of protein structures and the methods for their study utilising classic molecular dynamics simulations. The research presented in this thesis can be divided into two themes, comprising the rehydration of vacuum-exposed structures and the interrogation of the induced unfolding process of proteins. Out of their native environment, proteins undergo structural changes when exposed to vacuum. Investigating the ability to revert those potential vacuum-induced structural changes by means of computational rehydration provided detailed information on the underlying protein dynamics and how much of the structure revert back to their solution norm. We have further shown through rehydration simulations that applying an external electric field for dipole-orientation purposes does not induce irreversible changes to the protein structures. Our investigations on the induced unfolding of protein structures allowed a detailed look into the process of unfolding, accurately pinpointing areas within the proteins that unfolded first. The details provided by our simulations enabled us to describe potential mechanisms of the unfolding processes of different proteins on an atomistic level. The obtained results thus provide a potent theoretical basis for current and future experiments, where it will be very interesting to see MD compared with or complemented to experiments.
|
|
| 2. |
- Kutzner, Carsten, et al.
(författare)
-
More bang for your buck : Improved use of GPU nodes for GROMACS 2018
- 2019
-
Ingår i: Journal of Computational Chemistry. - : Wiley. - 0192-8651 .- 1096-987X. ; 40:27, s. 2418-2431
-
Tidskriftsartikel (refereegranskat)abstract
- We identify hardware that is optimal to produce molecular dynamics (MD) trajectories on Linux compute clusters with the GROMACS 2018 simulation package. Therefore, we benchmark the GROMACS performance on a diverse set of compute nodes and relate it to the costs of the nodes, which may include their lifetime costs for energy and cooling. In agreement with our earlier investigation using GROMACS 4.6 on hardware of 2014, the performance to price ratio of consumer GPU nodes is considerably higher than that of CPU nodes. However, with GROMACS 2018, the optimal CPU to GPU processing power balance has shifted even more toward the GPU. Hence, nodes optimized for GROMACS 2018 and later versions enable a significantly higher performance to price ratio than nodes optimized for older GROMACS versions. Moreover, the shift toward GPU processing allows to cheaply upgrade old nodes with recent GPUs, yielding essentially the same performance as comparable brand-new hardware.
|
|
| 3. |
- Kutzner, Carsten, et al.
(författare)
-
Software news and update : Speeding up parallel GROMACS on high-latency networks
- 2007
-
Ingår i: Journal of Computational Chemistry. - : Wiley. - 0192-8651 .- 1096-987X. ; 28:12, s. 2075-84
-
Tidskriftsartikel (refereegranskat)abstract
- We investigate the parallel scaling of the GROMACS molecular dynamics code on Ethernet Beowulf clusters and what prerequisites are necessary for decent scaling even on such clusters with only limited bandwidth and high latency. GROMACS 3.3 scales well on supercomputers like the IBM p690 (Regatta) and on Linux clusters with a special interconnect like Myrinet or Infiniband. Because of the high single-node performance of GROMACS, however, on the widely used Ethernet switched clusters, the scaling typically breaks down when more than two computer nodes are involved, limiting the absolute speedup that can be gained to about 3 relative to a single-CPU run. With the LAM MPI implementation, the main scaling bottleneck is here identified to be the all-to-all communication which is required every time step. During such an all-to-all communication step, a huge amount of messages floods the network, and as a result many TCP packets are lost. We show that Ethernet flow control prevents network congestion and leads to substantial scaling improvements. For 16 CPUs, e.g., a speedup of 11 has been achieved. However, for more nodes this mechanism also fails. Having optimized an all-to-all routine, which sends the data in an ordered fashion, we show that it is possible to completely prevent packet loss for any number of multi-CPU nodes. Thus, the GROMACS scaling dramatically improves, even for switches that lack flow control. In addition, for the common HP ProCurve 2848 switch we find that for optimum all-to-all performance it is essential how the nodes are connected to the switch's ports. This is also demonstrated for the example of the Car-Parinello MD code.
|
|
| 4. |
|
|
| 5. |
- Kutzner, Carsten, et al.
(författare)
-
Speeding up parallel GROMACS on high-latency networks
- 2007
-
Ingår i: Journal of Computational Chemistry. - : Wiley. - 0192-8651 .- 1096-987X. ; 28:12, s. 2075-2084
-
Tidskriftsartikel (refereegranskat)abstract
- We investigate the parallel scaling of the GROMACS molecular dynamics code on Ethernet Beowulf clusters and what prerequisites are necessary for decent scaling even on such clusters with only limited bandwidth and high latency. GROMACS 3.3 scales well on supercomputers like the IBM p690 (Regatta) and on Linux clusters with a special interconnect like Myrinet or Infiniband. Because of the high single-node performance of GROMACS, however, on the widely used Ethernet switched clusters, the scaling typically breaks down when more than two computer nodes are involved, limiting the absolute speedup that can be gained to about 3 relative to a single-CPU run. With the LAM MPI implementation, the main scaling bottleneck is here identified to be the all-to-all communication which is required every time step. During such an all-to-all communication step, a huge amount of messages floods the network, and as a result many TCP packets are lost. We show that Ethernet flow control prevents network congestion and leads to substantial scaling improvements. For 16 CPUs, e.g., a speedup of 11 has been achieved. However, for more nodes this mechanism also fails. Having optimized an all-to-all routine, which sends the data in an ordered fashion, we show that it is possible to completely prevent packet loss for any number of multi-CPU nodes. Thus, the GROMACS scaling dramatically improves, even for switches that lack flow control. In addition, for the common HP ProCurve 2848 switch we find that for optimum all-to-all performance it is essential how the nodes are connected to the switch's ports. This is also demonstrated for the example of the Car-Parinello MD code.
|
|
| 6. |
- Liu, Jing
(författare)
-
Towards Fast and Robust Algorithms in Flash X-ray single-particle Imaging
- 2020
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Modern X-ray Free Electron Laser (XFEL) technology provides the possibility to acquire a large number of diffraction patterns from individual biological nano-particles, including proteins, viruses, and DNA. Ideally, the collected data frames are high-quality single-particle diffraction patterns. However, unfortunately, the raw dataset is noisy and also contains patterns with scatterings from multiple particles, contaminated particles, etc. The data complexity and the massive volumes of raw data make pattern selection a time-consuming and challenge task. Further, X-rays interact with particles at random and the captured patterns are the 2D intensities of the scattered waves, i.e. we cannot observe the particle orientations and the phase information from the 2D diffraction patterns. To reconstruct 2D diffraction patterns into 3D structures of the desired particle, we need a sufficiently large single-particle-pattern dataset. The computational methodology for this reconstruction task is still under development and in need of an improved understanding of the algorithmic uncertainties.In this thesis, we tackle some of the challenges to obtain 3D structures of sample molecules from single-particle diffraction patterns. First, we have developed two classification methods to select single-particle diffraction patterns that are similar to provided templates. Second, we have accelerated the 3D reconstruction procedures by distributing the computations among Graphics Processing Units (GPUs) and by proposing an adaptive discretization of 3D space. Third, to better understand the uncertainties of the 3D reconstruction procedure, we have evaluated the impact of the different sources of resolution-limiting factors and introduced a practically applicable computational methodology in the form of bootstrap procedures for assessing the reconstruction uncertainty. These technologies form a data-analysis pipeline for recovering 3D structures from the raw X-ray single-particle data, which also analyzes the uncertainties. With the experimental developments of the X-ray single-particle technology, we expect that the data volumes will be increasing sharply, and hence, we believe such a computational pipeline will be critical to retrieve particle structures in the achievable resolution.
|
|
| 7. |
- Seibert, Mark Marvin, 1980-
(författare)
-
Protein Folding and DNA Origami
- 2010
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- In this thesis, the folding process of the de novo designed polypeptide chignolin was elucidated through atomic-scale Molecular Dynamics (MD) computer simulations. In a series of long timescale and replica exchange MD simulations, chignolin’s folding and unfolding was observed numerous times and the native state was identified from the computed Gibbs free-energy landscape. The rate of the self-assembly process was predicted from the replica exchange data through a novel algorithm and the structural fluctuations of an enzyme, lysozyme, were analyzed. DNA’s structural flexibility was investigated through experimental structure determination methods in the liquid and gas phase. DNA nanostructures could be maintained in a flat geometry when attached to an electrostatically charged, atomically flat surface and imaged in solution with an Atomic Force Microscope. Free in solution under otherwise identical conditions, the origami exhibited substantial compaction, as revealed by small angle X-ray scattering. This condensation was even more extensive in the gas phase. Protein folding is highly reproducible. It can rapidly lead to a stable state, which undergoes moderate fluctuations, at least for small structures. DNA maintains extensive structural flexibility, even when folded into large DNA origami. One may reflect upon the functional roles of proteins and DNA as a consequence of their atomic-level structural flexibility. DNA, biology’s information carrier, is very flexible and malleable, adopting to ever new conformations. Proteins, nature’s machines, faithfully adopt highly reproducible shapes to perform life’s functions robotically.
|
|
| 8. |
- Östlin, Christofer, 1988-
(författare)
-
Simulations of Biomolecular Fragmentation and Diffraction with Ultrafast X-ray Lasers
- 2019
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Studies of biomolecules have recently seen substantial developments. New X-ray lasers allow for high-resolution imaging of protein crystals too small for conventional X-ray crystallography. Even structures of single particles have been determined at lower resolutions with these new sources. The secret lies in the ultrashort high-intensity pulses, which allow for diffraction and retrieval of structural information before the sample gets fragmented. However, the attainable resolution is still limited, in particular when imaging non-crystalline samples, making further advancements highly desired. In this thesis, some of the resolution-limiting obstacles facing single particle imaging (SPI) of proteins are studied in silico.As the X-ray pulse interacts with injected single molecules, their spatial orientation is generally unknown. Recovering the orientation is essential to the structure determination process, and currently nontrivial. Molecular dynamics simulations show that the Coulomb explosion due to intense X-ray ionization could provide information pertaining to the original orientation. Used in conjunction with current methods, this would lead to an enhanced three-dimensional reconstruction of the protein.Radiation damage and sample heterogeneity constitute considerable sources of noise in SPI. Pulse durations are presently not brief enough to circumvent damage, causing the sample to deteriorate during imaging, and the accuracy of the averaged diffraction pattern is impaired by structural variations. The extent of these effects were studied by molecular dynamics. Our findings suggest that radiation damage in terms of ionization and atomic displacement promotes a gating mechanism, benefiting imaging with longer pulses. Because of this, sample heterogeneity poses a greater challenge and efforts should be made to minimize its impact.X-ray lasers generate pulses with a stochastic temporal distribution of photons, affecting the achievable resolution on a pulse-to-pulse basis. Plasma simulations were performed to investigate how these fluctuations influence the damage dynamics and the diffraction signal. The results reveal that structural information is particularly well-preserved if the temporal distribution is skewed such that most photons are concentrated at the beginning.While many obstacles remain, the prospect of atomic-resolution SPI is drawing ever closer. This thesis is but one of the stepping stones necessary to get us there. Once we do, the possibilities are limitless.
|
|
