Search

Data-driven medical care delivery must always respect patient privacy-a requirement that is not easily met. This issue has impeded improvements to healthcare software and has delayed the long-predicted prevalence of artificial intelligence in healthcare. Until now, it has been very difficult to share data between healthcare organizations, resulting in poor statistical models due to unrepresentative patient cohorts. Synthetic data, i.e., artificial but realistic electronic health records, could overcome the drought that is troubling the healthcare sector. Deep neural network architectures, in particular, have shown an incredible ability to learn from complex data sets and generate large amounts of unseen data points with the same statistical properties as the training data. Here, we present a generative neural network model that can create synthetic health records with realistic timelines. These clinical trajectories are generated on a per-patient basis and are represented as linear-sequence graphs of clinical events over time. We use a variational graph autoencoder (VGAE) to generate synthetic samples from real-world electronic health records. Our approach generates health records not seen in the training data. We show that these artificial patient trajectories are realistic and preserve patient privacy and can therefore support the safe sharing of data across organizations., QC 20230526

Data-driven medical care delivery must always respect patient privacy—a requirement that is not easily met. This issue has impeded improvements to healthcare software and has delayed the long-predicted prevalence of artificial intelligence in healthcare. Until now, it has been very difficult to share data between healthcare organizations, resulting in poor statistical models due to unrepresentative patient cohorts. Synthetic data, i.e., artificial but realistic electronic health records, could overcome the drought that is troubling the healthcare sector. Deep neural network architectures, in particular, have shown an incredible ability to learn from complex data sets and generate large amounts of unseen data points with the same statistical properties as the training data. Here, we present a generative neural network model that can create synthetic health records with realistic timelines. These clinical trajectories are generated on a per-patient basis and are represented as linear-sequence graphs of clinical events over time. We use a variational graph autoencoder (VGAE) to generate synthetic samples from real-world electronic health records. Our approach generates health records not seen in the training data. We show that these artificial patient trajectories are realistic and preserve patient privacy and can therefore support the safe sharing of data across organizations. © 2023, The Author(s)., Funding: M.V. is partially supported by the “Wallenberg AI, Autonomous Systems and Software Program” (WASP). M.L. is partially supported by AIR Lund (Artificially Intelligent use of Registers at Lund University) research environment and received funding from the Swedish Research Council (VR; grant No. 2019-00198). G.N. is supported by the French National research agency via the AML-HELAS (ANR-19-CHIA-0020) project.

Nanomaterials possess a wide range of potential applications due to their novel properties and exceptionally high activity as a result of their large surface to volume ratios compared to bulk matter. The active surface may present both advantage and risk when the nanomaterials interact with living organisms. As the overall biological impact of nanomaterials is triggered and mediated by interactions at the bio-nano interface, an ability to predict those from the atomistic descriptors, especially before the material is produced, can present enormous advantage for the development of nanotechnology. Fast screening of nanomaterials and their variations for specific biological effects can be enabled using computational materials modelling. The challenge lies in the range of scales that needs to be crossed from the material-specific atomistic representation to the relevant length scales covering typical biomolecules (proteins and lipids). In this work, we present a systematic multiscale approach that allows one to evaluate crucial interactions at the bionano interface from the first principles without any prior information about the material and thus establish links between the details of the nanomaterials structure to protein-nanoparticle interactions. As an example, an advanced computational characterization of titanium dioxide nanoparticles (6 different surfaces of rutile and anatase polymorphs) has been performed. We computed characteristics of the titanium dioxide interface with water using density functional theory for electronic density, used these parameters to derive an atomistic force field, and calculated adsorption energies for essential biomolecules on the surface of titania nanoparticles via direct atomistic simulations and coarse-grained molecular dynamics. Hydration energies, as well as adsorption energies for a set of 40 blood proteins are reported.

The use of carbon-based nanomaterials is tremendously increasing in various areas of technological, bioengineering, and biomedical applications. The functionality of carbon-based nanomaterials can be further broadened via chemical functionalization of carbon nanomaterial surfaces. On the other hand, concern is rising on possible adverse effects when nanomaterials are taken up by biological organisms. In order to contribute into understanding of interactions of carbon-based nanomaterials with biological matter, we have investigated adsorption of small biomolecules on nanomaterials using enhanced sampling molecular dynamics. The biomolecules included amino acid side chain analogues, fragments of lipids, and sugar monomers. The adsorption behavior on unstructured amorphous carbon, pristine graphene and its derivatives (such as few-layer graphene, graphene oxide, and reduced graphene oxide) as well as pristine carbon nanotubes, and those functionalized with OH-, COOH-, COO-, NH2-, and NH3+ groups was investigated with respect to surface concentration. An adsorption profile, that is, the free energy as a function of distance from the nanomaterial surfaces, was determined for each molecule and surface using the Metadynamics approach. The results were analyzed in terms of chemical specificity, surface charge, and surface concentration. It was shown that although morphology of the nanomaterial has a limited effect on the adsorption properties, functionalization of the surface by various molecular groups can drastically change the adsorption behavior that can be used in the design of nanosurfaces with highly selective adsorption properties and safe for human health and environment.

Atomistic simulations can complement the scarce experimental data on free energies of molecules at bio-inorganic interfaces. In molecular simulations, adsorption free energy landscapes are efficiently explored with advanced sampling methods, but classical dynamics is unable to capture charge transfer and polarization at the solid-liquid interface. Ab initio simulations do not suffer from this flaw, but only at the expense of an overwhelming computational cost. Here, we introduce a protocol for adsorption free energy calculations that improves sampling on the timescales relevant to ab initio simulations. As a case study, we calculate adsorption free energies of the charged amino acids Lysine and Aspartate on the fully hydrated anatase (101) TiO2 surface using tight-binding forces. We find that the first-principle description of the system significantly contributes to the adsorption free energies, which is overlooked by calculations with previous methods.

Atomistic simulations can complement the scarce experimental data on free energies of molecules at bio-inorganic interfaces. In molecular simulations, adsorption free energy landscapes are efficiently explored with advanced sampling methods, but classical dynamics is unable to capture charge transfer and polarization at the solid-liquid interface. Ab initio simulations do not suffer from this flaw, but only at the expense of an overwhelming computational cost. Here, we introduce a protocol for adsorption free energy calculations that improves sampling on the timescales relevant to ab initio simulations. As a case study, we calculate adsorption free energies of the charged amino acids Lysine and Aspartate on the fully hydrated anatase (101) TiO2 surface using tight-binding forces. We find that the first-principle description of the system significantly contributes to the adsorption free energies, which is overlooked by calculations with previous methods.

Atomistic simulations can complement the scarce experimental data on free energies of molecules at bio-inorganic interfaces. In molecular simulations, adsorption free energy landscapes are efficiently explored with advanced sampling methods, but classical dynamics is unable to capture charge transfer and polarization at the solid-liquid interface. Ab initio simulations do not suffer from this flaw, but only at the expense of an overwhelming computational cost. Here, we introduce a protocol for adsorption free energy calculations that improves sampling on the timescales relevant to ab initio simulations. As a case study, we calculate adsorption free energies of the charged amino acids Lysine and Aspartate on the fully hydrated anatase (101) TiO2 surface using tight-binding forces. We find that the first-principle description of the system significantly contributes to the adsorption free energies, which is overlooked by calculations with previous methods.

Atomistic simulations can complement the scarce experimental data on free energies of molecules at bio-inorganic interfaces. In molecular simulations, adsorption free energy landscapes are efficiently explored with advanced sampling methods, but classical dynamics is unable to capture charge transfer and polarization at the solid-liquid interface. Ab initio simulations do not suffer from this flaw, but only at the expense of an overwhelming computational cost. Here, we introduce a protocol for adsorption free energy calculations that improves sampling on the timescales relevant to ab initio simulations. As a case study, we calculate adsorption free energies of the charged amino acids Lysine and Aspartate on the fully hydrated anatase (101) TiO2 surface using tight-binding forces. We find that the first-principle description of the system significantly contributes to the adsorption free energies, which is overlooked by calculations with previous methods.

Atomistic simulations can complement the scarce experimental data on free energies of molecules at bio-inorganic interfaces. In molecular simulations, adsorption free energy landscapes are efficiently explored with advanced sampling methods, but classical dynamics is unable to capture charge transfer and polarization at the solid-liquid interface. Ab initio simulations do not suffer from this flaw, but only at the expense of an overwhelming computational cost. Here, we introduce a protocol for adsorption free energy calculations that improves sampling on the timescales relevant to ab initio simulations. As a case study, we calculate adsorption free energies of the charged amino acids Lysine and Aspartate on the fully hydrated anatase (101) TiO2 surface using tight-binding forces. We find that the first-principle description of the system significantly contributes to the adsorption free energies, which is overlooked by calculations with previous methods.

Atomistic simulations can complement the scarce experimental data on free energies of molecules at bio-inorganic interfaces. In molecular simulations, adsorption free energy landscapes are efficiently explored with advanced sampling methods, but classical dynamics is unable to capture charge transfer and polarization at the solid-liquid interface. Ab initio simulations do not suffer from this flaw, but only at the expense of an overwhelming computational cost. Here, we introduce a protocol for adsorption free energy calculations that improves sampling on the timescales relevant to ab initio simulations. As a case study, we calculate adsorption free energies of the charged amino acids Lysine and Aspartate on the fully hydrated anatase (101) TiO2 surface using tight-binding forces. We find that the first-principle description of the system significantly contributes to the adsorption free energies, which is overlooked by calculations with previous methods.

Cardiolipin is a non-bilayer phospholipid with a unique dimeric structure. It localizes to negative curvature regions in bacteria and is believed to stabilize respiratory chain complexes in the highly curved mitochondrial membrane. Cardiolipin's localization mechanism remains unresolved, because important aspects such as the structural basis and strength for lipid curvature preferences are difficult to determine, partly due to the lack of efficient simulation methods. Here, we report a computational approach to study curvature preferences of cardiolipin by simulated membrane buckling and quantitative modeling. We combine coarse-grained molecular dynamics with simulated buckling to determine the curvature preferences in three-component bilayer membranes with varying concentrations of cardiolipin, and extract curvature-dependent concentrations and lipid acyl chain order parameter profiles. Cardiolipin shows a strong preference for negative curvatures, with a highly asymmetric chain order parameter profile. The concentration profiles are consistent with an elastic model for lipid curvature sensing that relates lipid segregation to local curvature via the material constants of the bilayers. These computations constitute new steps to unravel the molecular mechanism by which cardiolipin senses curvature in lipid membranes, and the method can be generalized to other lipids and membrane components as well.

Cardiolipin is a non-bilayer phospholipid with a unique dimeric structure. It localizes to negative curvature regions in bacteria and is believed to stabilize respiratory chain complexes in the highly curved mitochondrial membrane. Cardiolipin's localization mechanism remains unresolved, because important aspects such as the structural basis and strength for lipid curvature preferences are difficult to determine, partly due to the lack of efficient simulation methods. Here, we report a computational approach to study curvature preferences of cardiolipin by simulated membrane buckling and quantitative modeling. We combine coarse-grained molecular dynamics with simulated buckling to determine the curvature preferences in three-component bilayer membranes with varying concentrations of cardiolipin, and extract curvature-dependent concentrations and lipid acyl chain order parameter profiles. Cardiolipin shows a strong preference for negative curvatures, with a highly asymmetric chain order parameter profile. The concentration profiles are consistent with an elastic model for lipid curvature sensing that relates lipid segregation to local curvature via the material constants of the bilayers. These computations constitute new steps to unravel the molecular mechanism by which cardiolipin senses curvature in lipid membranes, and the method can be generalized to other lipids and membrane components as well.

Cardiolipin is a non-bilayer phospholipid with a unique dimeric structure. It localizes to negative curvature regions in bacteria and is believed to stabilize respiratory chain complexes in the highly curved mitochondrial membrane. Cardiolipin's localization mechanism remains unresolved, because important aspects such as the structural basis and strength for lipid curvature preferences are difficult to determine, partly due to the lack of efficient simulation methods. Here, we report a computational approach to study curvature preferences of cardiolipin by simulated membrane buckling and quantitative modeling. We combine coarse-grained molecular dynamics with simulated buckling to determine the curvature preferences in three-component bilayer membranes with varying concentrations of cardiolipin, and extract curvature-dependent concentrations and lipid acyl chain order parameter profiles. Cardiolipin shows a strong preference for negative curvatures, with a highly asymmetric chain order parameter profile. The concentration profiles are consistent with an elastic model for lipid curvature sensing that relates lipid segregation to local curvature via the material constants of the bilayers. These computations constitute new steps to unravel the molecular mechanism by which cardiolipin senses curvature in lipid membranes, and the method can be generalized to other lipids and membrane components as well.

Cardiolipin is a non-bilayer phospholipid with a unique dimeric structure. It localizes to negative curvature regions in bacteria and is believed to stabilize respiratory chain complexes in the highly curved mitochondrial membrane. Cardiolipin's localization mechanism remains unresolved, because important aspects such as the structural basis and strength for lipid curvature preferences are difficult to determine, partly due to the lack of efficient simulation methods. Here, we report a computational approach to study curvature preferences of cardiolipin by simulated membrane buckling and quantitative modeling. We combine coarse-grained molecular dynamics with simulated buckling to determine the curvature preferences in three-component bilayer membranes with varying concentrations of cardiolipin, and extract curvature-dependent concentrations and lipid acyl chain order parameter profiles. Cardiolipin shows a strong preference for negative curvatures, with a highly asymmetric chain order parameter profile. The concentration profiles are consistent with an elastic model for lipid curvature sensing that relates lipid segregation to local curvature via the material constants of the bilayers. These computations constitute new steps to unravel the molecular mechanism by which cardiolipin senses curvature in lipid membranes, and the method can be generalized to other lipids and membrane components as well.

Cardiolipin is a non-bilayer phospholipid with a unique dimeric structure. It localizes to negative curvature regions in bacteria and is believed to stabilize respiratory chain complexes in the highly curved mitochondrial membrane. Cardiolipin's localization mechanism remains unresolved, because important aspects such as the structural basis and strength for lipid curvature preferences are difficult to determine, partly due to the lack of efficient simulation methods. Here, we report a computational approach to study curvature preferences of cardiolipin by simulated membrane buckling and quantitative modeling. We combine coarse-grained molecular dynamics with simulated buckling to determine the curvature preferences in three-component bilayer membranes with varying concentrations of cardiolipin, and extract curvature-dependent concentrations and lipid acyl chain order parameter profiles. Cardiolipin shows a strong preference for negative curvatures, with a highly asymmetric chain order parameter profile. The concentration profiles are consistent with an elastic model for lipid curvature sensing that relates lipid segregation to local curvature via the material constants of the bilayers. These computations constitute new steps to unravel the molecular mechanism by which cardiolipin senses curvature in lipid membranes, and the method can be generalized to other lipids and membrane components as well.

Cardiolipin is a non-bilayer phospholipid with a unique dimeric structure. It localizes to negative curvature regions in bacteria and is believed to stabilize respiratory chain complexes in the highly curved mitochondrial membrane. Cardiolipin's localization mechanism remains unresolved, because important aspects such as the structural basis and strength for lipid curvature preferences are difficult to determine, partly due to the lack of efficient simulation methods. Here, we report a computational approach to study curvature preferences of cardiolipin by simulated membrane buckling and quantitative modeling. We combine coarse-grained molecular dynamics with simulated buckling to determine the curvature preferences in three-component bilayer membranes with varying concentrations of cardiolipin, and extract curvature-dependent concentrations and lipid acyl chain order parameter profiles. Cardiolipin shows a strong preference for negative curvatures, with a highly asymmetric chain order parameter profile. The concentration profiles are consistent with an elastic model for lipid curvature sensing that relates lipid segregation to local curvature via the material constants of the bilayers. These computations constitute new steps to unravel the molecular mechanism by which cardiolipin senses curvature in lipid membranes, and the method can be generalized to other lipids and membrane components as well.

Membrane curvature sensing, where the binding free energies of membrane-associated molecules depend on the local membrane curvature, is a key factor to modulate and maintain the shape and organization of cell membranes. However, the microscopic mechanisms are not well understood, partly due to absence of efficient simulation methods. Here, we describe a method to compute the curvature dependence of the binding free energy of a membrane associated probe molecule that interacts with a buckled membrane, which has been created by lateral compression of a flat bilayer patch. This buckling approach samples a wide range of curvatures in a single simulation, and anisotropic effects can be extracted from the orientation statistics. We develop an efficient and robust algorithm to extract the motion of the probe along the buckled membrane surface, and evaluate its numerical properties by extensive sampling of three coarse-grained model systems: local lipid density in a curved environment for single-component bilayers, curvature preferences of individual lipids in two-component membranes, and curvature sensing by a homotrimeric transmembrane protein. The method can be used to complement experimental data from curvature partition assays and provides additional insight into mesoscopic theories and molecular mechanisms for curvature sensing.

Postsynthetic reactions of metal-organic frameworks (MOFs) are versatile tools for producing functional materials, but the methods of evaluating these reactions are cumbersome and destructive. Here we demonstrate and validate the use of in situ NMR spectroscopy of species in the liquid state to examine solvent-assisted ligand exchange (SALE) and postsynthetic modification (PSM) reactions of metal-organic frameworks. This technique allows functionalization to be monitored over time without decomposing the product for analysis, which simplifies reaction screening. In the case of SALE, both the added ligand and the ligand leaving the framework can be observed. We demonstrate this in situ method by examining SALE and PSM reactions of the robust zirconium MOF UiO-67 as well as SALE with the aluminum MOF DUT-5. In situ NMR spectroscopy provided insights into the reactions studied, and we expect that future studies using this method will permit the examination of a variety of MOF-solute reactions.

Membrane curvature sensing, where the binding free energies of membrane-associated molecules depend on the local membrane curvature, is a key factor to modulate and maintain the shape and organization of cell membranes. However, the microscopic mechanisms are not well understood, partly due to absence of efficient simulation methods. Here, we describe a method to compute the curvature dependence of the binding free energy of a membrane associated probe molecule that interacts with a buckled membrane, which has been created by lateral compression of a flat bilayer patch. This buckling approach samples a wide range of curvatures in a single simulation, and anisotropic effects can be extracted from the orientation statistics. We develop an efficient and robust algorithm to extract the motion of the probe along the buckled membrane surface, and evaluate its numerical properties by extensive sampling of three coarse-grained model systems: local lipid density in a curved environment for single-component bilayers, curvature preferences of individual lipids in two-component membranes, and curvature sensing by a homotrimeric transmembrane protein. The method can be used to complement experimental data from curvature partition assays and provides additional insight into mesoscopic theories and molecular mechanisms for curvature sensing.

Dissociative and molecular water adsorption on the anatase (001) surface is studied in the context of state-of-the-art density functional theory in large supercells suited for adsorption studies at various water coverage ratios. At low coverage values below 1/4 ML, water adsorption remains dissociative and a network of hydrogen bonds between the so formed hydroxyl groups favors the formation of a ridge surface structure. The hydroxyl patterned (4 X 4) surface thus undergoes a (2 X 4) reconstruction that causes the relief of the large tensile stress measured in the unreconstructed surface along the direction orthogonal to the ridge. This phenomenology is accompanied by the loss of reactivity of the reconstructed surface with respect to the dissociative water adsorption that becomes molecular above 1/4 ML. We also show that the molecular adsorption on the terrace is weaker than the one on the ridge. The present water reconstruction model is discussed and compared to the well-known ADM model of the reconstructed anatase (001) surface in dry environment.

Ab initio molecular dynamics simulations are reported forwater-embedded TiO2 surfaces to determine the diffusive and reactive behavior at full hydration. A three-domain model is developed for six surfaces [rutile (110), (100), and (001), and anatase (101), (100), and (001)] which describes waters as hard (irreversibly bound to the surface), soft (with reduced mobility but orientation freedom near the surface), or bulk. The model explains previous experimental data and provides a detailed picture of water diffusion near TiO2 surfaces. Water reactivity is analyzed with a graph-theoretic approach that reveals a number of reaction pathways on TiO2 which occur at full hydration, in addition to direct water splitting. Hydronium (H3O+) is identified to be a key intermediate state, which facilitates water dissociation by proton hopping between intact and dissociated waters near the surfaces. These discoveries significantly improve the understanding of nanoscale water dynamics and reactivity at TiO2 interfaces under ambient conditions.

Small-sized wet TiO2 nanoparticles have been investigated by ab initio molecular dynamics simulations. Chemical and physical adsorption of water on the TiO2-water interface was studied as a function of water content, ranging from dry nanoparticles to wet nanoparticles with monolayer coverage of water. The surface reactivity was shown to be a concave function of water content and driven by surface defects. The local coordination number at the defect was identified as the key factor to decide whether water adsorption proceeds through dissociation or physisorption on the surface. A consistent picture of TiO2 nanoparticle wetting at the microscopic level emerges, which corroborates existing experimental data and gives further insight into the molecular mechanisms behind nanoparticle wetting. These calculations will facilitate the engineering of metal oxide nanoparticles with a controlled catalytic water activity.

Adsorption profiles and adsorption free energies were determined for the side chain analogues of the 20 naturally occurring amino acids and a titanium binding peptide on the TiO2 (100) surface. Microsecond simulations with umbrella sampling and metadynamics were used to sample the free energy barriers associated with desolvation of strongly bound water molecules at the TiO2 surface. Polar and aromatic side chain analogues that hydrogen bond either to surface waters or directly to the metal oxide surface were found to be the strongest binders. Further, adsorption simulations of a 6 residue titanium binding peptide identified two binding modes on TiO2 (100). The peptide structure with lowest free energy was shown to be stabilized by a salt bridge between the end termini. A comparison between the free energies of the side chain analogues of the peptide sequence and the peptide itself shows that the free energy contributions are not additive. The simulations emphasize that tightly bound surface waters play a key role for peptide and protein structures when bound to inorganic surfaces in biological environments.

Atomistic force field parameters were developed for the TiO2-water interface by systematic optimization with respect to experimentally determined crystal structures (lattice parameters) and surface thermodynamics (water adsorption enthalpy). Optimized force field parameters were determined for the two cases where TiO2 was modeled with or without covalent bonding. The nonbonded TiO2 model can be used to simulate different TiO2 phases, while the bonded TiO2 model is particularly useful for simulations of nanosized TiO2 and biomatter, including protein-surface and nanoparticle-biomembrane simulations. The procedure is easily generalized to parametrize interactions between other inorganic surfaces and biomolecules.

The traditional method for extracting electron density and other transmembrane profiles from molecular dynamics simulations of lipid bilayers fails for large bilayer systems, because it assumes a flat reference surface that does not take into account long wavelength undulations. We have developed what we believe to be a novel set of methods to characterize these undulations and extract the underlying profiles in the large systems. Our approach first obtains an undulation reference surface for each frame in the simulation and subsequently isolates the long-wavelength undulations by filtering out the intrinsic short wavelength modes. We then describe two methods to obtain the appropriate profiles from the undulating reference surface. Most combinations of methods give similar results for the electron density profiles of our simulations of 1024 DMPC lipids. From simulations of smaller systems, we also characterize the finite size effect related to the boundary conditions of the simulation box. In addition, we have developed a set of methods that use the undulation reference surface to determine the true area per lipid which, due to undulations, is larger than the projected area commonly reported from simulations., QC 20110527

Atomic resolution and coarse-grained simulations of dimyristoylphosphatidylcholine lipid bilayers were analyzed for fluctuations perpendicular to the bilayer using a completely Fourier-based method. We find that the fluctuation spectrum of motions perpendicular to the bilayer can be decomposed into just two parts: 1), a pure undulation spectrum proportional to q(-4) that dominates in the small-q regime; and 2), a molecular density structure factor contribution that dominates in the large-q regime. There is no need for a term proportional to q(-2) that has been postulated for protrusion fluctuations and that appeared to have been necessary to fit the spectrum for intermediate q. We suggest that earlier reports of such a term were due to the artifact of binning and smoothing in real space before obtaining the Fourier spectrum. The observability of an intermediate protrusion regime from the fluctuation spectrum is discussed based on measured and calculated material constants., QC 20110530

Lipid molecules form thin biological membranes that envelop all living cells, and behave as two-dimensional liquid sheets immersed in bulk water. The interactions of such biomembranes with their environment lay the foundation of a plethora of biological processes rooted in the mesoscopic domain - length scales of 1-1000 nm and time scales of 1-1000 ns. Research in this intermediate regime has for a long time been out of reach for conventional experiments, but breakthroughs in computer simulation methods and scattering experimental techniques have made it possible to directly probe static and dynamic properties of biomembranes on these scales. Biomembranes are soft, with a relatively low energy cost of bending, and are thereby influenced by random, thermal fluctuations of individual molecules. Molecular dynamics simulations show how in-plane (density fluctuations) and out-of-plane (undulations) motions are intertwined in the bilayer in the mesoscopic domain. By novel methods, the fluctuation spectra of lipid bilayers can be calculated withdirect Fourier analysis. The interpretation of the fluctuation spectra reveals a picture where density fluctuations and undulations are most pronounced on different length scales, but coalesce in the mesoscopic regime. This analysis has significant consequences for comparison of simulation data to experiments. These new methods merge the molecular fluctuations on small wavelengths, with continuum fluctuations of the elastic membrane sheet on large wavelengths, allowing electron density profiles (EDP) and area per lipid to be extracted from simulations with high accuracy. Molecular dynamics simulations also provide insight on the small-wavelength dynamics of lipid membranes. Rapidly decaying density fluctuations can be described as propagating sound waves in the framework of linearized hydrodynamics, but there is a slow, dispersive, contribution that needs to be described by a stretched exponential over a broad range of length- and t, QC 20111014, Modelling of biological membranes

Lipid molecules form thin biological membranes that envelop all living cells, and behave as two-dimensional liquid sheets immersed in bulk water. The interactions of such biomembranes with their environment lay the foundation of a plethora of biological processes rooted in the mesoscopic domain - length scales of 1-1000 nm and time scales of 1-1000 ns. Research in this intermediate regime has for a long time been out of reach for conventional experiments, but breakthroughs in computer simulation methods and scattering experimental techniques have made it possible to directly probe static and dynamic properties of biomembranes on these scales. Biomembranes are soft, with a relatively low energy cost of bending, and are thereby influenced by random, thermal fluctuations of individual molecules. Molecular dynamics simulations show how in-plane (density fluctuations) and out-of-plane (undulations) motions are intertwined in the bilayer in the mesoscopic domain. By novel methods, the fluctuation spectra of lipid bilayers can be calculated withdirect Fourier analysis. The interpretation of the fluctuation spectra reveals a picture where density fluctuations and undulations are most pronounced on different length scales, but coalesce in the mesoscopic regime. This analysis has significant consequences for comparison of simulation data to experiments. These new methods merge the molecular fluctuations on small wavelengths, with continuum fluctuations of the elastic membrane sheet on large wavelengths, allowing electron density profiles (EDP) and area per lipid to be extracted from simulations with high accuracy. Molecular dynamics simulations also provide insight on the small-wavelength dynamics of lipid membranes. Rapidly decaying density fluctuations can be described as propagating sound waves in the framework of linearized hydrodynamics, but there is a slow, dispersive, contribution that needs to be described by a stretched exponential over a broad range of length- and t, QC 20111014, Modelling of biological membranes

Lipid molecules form thin biological membranes that envelop all living cells, and behave as two-dimensional liquid sheets immersed in bulk water. The interactions of such biomembranes with their environment lay the foundation of a plethora of biological processes rooted in the mesoscopic domain - length scales of 1-1000 nm and time scales of 1-1000 ns. Research in this intermediate regime has for a long time been out of reach for conventional experiments, but breakthroughs in computer simulation methods and scattering experimental techniques have made it possible to directly probe static and dynamic properties of biomembranes on these scales. Biomembranes are soft, with a relatively low energy cost of bending, and are thereby influenced by random, thermal fluctuations of individual molecules. Molecular dynamics simulations show how in-plane (density fluctuations) and out-of-plane (undulations) motions are intertwined in the bilayer in the mesoscopic domain. By novel methods, the fluctuation spectra of lipid bilayers can be calculated withdirect Fourier analysis. The interpretation of the fluctuation spectra reveals a picture where density fluctuations and undulations are most pronounced on different length scales, but coalesce in the mesoscopic regime. This analysis has significant consequences for comparison of simulation data to experiments. These new methods merge the molecular fluctuations on small wavelengths, with continuum fluctuations of the elastic membrane sheet on large wavelengths, allowing electron density profiles (EDP) and area per lipid to be extracted from simulations with high accuracy. Molecular dynamics simulations also provide insight on the small-wavelength dynamics of lipid membranes. Rapidly decaying density fluctuations can be described as propagating sound waves in the framework of linearized hydrodynamics, but there is a slow, dispersive, contribution that needs to be described by a stretched exponential over a broad range of length- and t, QC 20111014, Modelling of biological membranes

Lipid molecules form thin biological membranes that envelop all living cells, and behave as two-dimensional liquid sheets immersed in bulk water. The interactions of such biomembranes with their environment lay the foundation of a plethora of biological processes rooted in the mesoscopic domain - length scales of 1-1000 nm and time scales of 1-1000 ns. Research in this intermediate regime has for a long time been out of reach for conventional experiments, but breakthroughs in computer simulation methods and scattering experimental techniques have made it possible to directly probe static and dynamic properties of biomembranes on these scales. Biomembranes are soft, with a relatively low energy cost of bending, and are thereby influenced by random, thermal fluctuations of individual molecules. Molecular dynamics simulations show how in-plane (density fluctuations) and out-of-plane (undulations) motions are intertwined in the bilayer in the mesoscopic domain. By novel methods, the fluctuation spectra of lipid bilayers can be calculated withdirect Fourier analysis. The interpretation of the fluctuation spectra reveals a picture where density fluctuations and undulations are most pronounced on different length scales, but coalesce in the mesoscopic regime. This analysis has significant consequences for comparison of simulation data to experiments. These new methods merge the molecular fluctuations on small wavelengths, with continuum fluctuations of the elastic membrane sheet on large wavelengths, allowing electron density profiles (EDP) and area per lipid to be extracted from simulations with high accuracy. Molecular dynamics simulations also provide insight on the small-wavelength dynamics of lipid membranes. Rapidly decaying density fluctuations can be described as propagating sound waves in the framework of linearized hydrodynamics, but there is a slow, dispersive, contribution that needs to be described by a stretched exponential over a broad range of length- and t, QC 20111014, Modelling of biological membranes

QC 20140924

The decay of fluctuations in fluid biomembranes is strongly stretched and nonexponential on nanometer lengthscales. We report on calculations of structural correlation functions for lipid bilayer membranes from atomistic and coarse-grained molecular dynamics simulations. The time scales extend up to microseconds, whereas the linear size of the largest systems is around 50 nm. Thus, we can cover the equilibrium dynamics of wave vectors over two orders of magnitude (0.2-20 nm(-1)). The time correlations observed in the simulations are best described by stretched exponential functions, with exponents of 0.45 for the atomistic and 0.60 for the coarse-grained model. Area number density fluctuations, thickness fluctuations, and undulations behave dynamically in a similar way and have almost exactly the same dynamics for wavelengths below 3 nm, indicating that in this regime undulations and thickness fluctuations are governed by in-plane density fluctuations. The out-of-plane height fluctuations are apparent only at the longest wavelengths accessible in the simulations (above 6 nm). The effective correlation times of the stretched exponentials vary strongly with the wave vector. The variation fits inverse power-laws that change with wavelength. The exponent is 3 for wavelengths smaller than about 1.25 nm and switches to 1 above this. There are indications for a switch to still another exponent, 2, for wavelengths above 20 nm. Compared to neutron spin-echo (NSE) experiments, the simulation data indicate a faster relaxation in the hydrodynamic limit, although an extrapolation of NSE data, as well as inelastic neutron scattering data, is in agreement with our data at larger wave vectors., QC 20101208

QC 20200310

The binding of zinc (Zn) ions to proteins is important for many cellular events. The theoretical and computational description of this binding (as well as that of other transition metals) is a challenging task. In this paper the binding of the Zn ion to four cysteine residues in the structural site of horse liver alcohol dehydrogenase (HLADH) is studied using a synthetic peptide mimic of this site. The study includes experimental measurements of binding constants, classical free energy calculations from molecular dynamics (MD) simulations and quantum mechanical (QM) electron structure calculations. The classical MD results account for interactions at the molecular level and reproduce the absolute binding energy and the hydration free energy of the Zn ion with an accuracy of about 10%. This is insufficient to obtain correct free energy differences. QM correction terms were calculated from density functional theory (DFT) on small clusters of atoms to include electronic polarisation of the closest waters and covalent contributions to the Zn-S coordination bond. This results in reasonably good agreement with the experimentally measured binding constants and Zn ion hydration free energies in agreement with published experimental values. The study also includes the replacement of one cysteine residue to an alanine. Simulations as well as experiments showed only a small effect of this upon the binding free energy. A detailed analysis indicate that the sulfur is replaced by three water molecules, thereby changing the coordination number of Zn from four (as in the original peptide) to six (as in water)., QC 20100923

Dynamic structure factors for a lipid bilayer have been calculated from molecular dynamics simulations. From trajectories of a system containing 1024 lipids we obtain wave vectors down to 0.34 nm(-1), which enables us to directly resolve the Rayleigh and Brillouin lines of the spectrum. The results confirm the validity of a model based on generalized hydrodynamics, but also improves the line widths and the position of the Brillouin lines. The improved resolution shows that the Rayleigh line is narrower than in earlier studies, which corresponds to a smaller thermal diffusivity. From a detailed analysis of the power spectrum, we can, in fact, distinguish two dispersive contributions to the elastic scattering. These translate to two exponential relaxation processes in separate time domains. Further, by including a first correction to the wave-vector-dependent position of the Brillouin lines, the results agree favorably to generalized hydrodynamics even up to intermediate wave vectors, and also yields a 20% higher adiabatic sound velocity. The width of the Brillouin lines shows a linear, not quadratic, dependence to low wave vectors., QC 20101006

Computer simulations of simplified model systems provide understanding of how complex biological systems behave. The simulations give detailed information about the systems, with atomistic resolution, that can be used in combination with experimental knowledge to shed light on underlying physical principles. The thesis presents background information about the studies of two important model systems in biological physics. First, metal ion-binding to proteins is investigated in a computational study on a zinc-binding synthetic peptide, to elucidate the binding details. The major scientific contributions from the study are the identification and mapping of the detailed contributions to the binding. A novel correction scheme is worked out, where classic free energy calculations are combined with density functional theory to adjust for quantum mechanical effects. The sensitivity of the zinc-binding to a specific amino acid segment can be explained in terms of the zinc coordination. Second, equilibrium density fluctuations in biological membranes are studied using computer simulations of the lipid bilayer. The fluctuations are linked to several processes; pore formation, membrane permeability and transport of small molecules across themembrane. Because the lipid bilayer behaves similar to a 2D fluid the density fluctuations can be described in the framework of generalized hydrodynamics. The major scientific contributions from the study are the direct calculation of the density-density autocorrelation function from raw data and the observation that the diffusive contribution to the power spectrum (the Rayleigh line) is not single-exponential. In addition, the accuracy of the approximate hydrodynamic solutions is questionable for the propagation of sound waves in the membrane.

The binding of zinc (Zn) ions to proteins is important for many cellular events. The theoretical and computational description of this binding (as well as that of other transition metals) is a challenging task. In this paper the binding of the Zn ion to four cysteine residues in the structural site of horse liver alcohol dehydrogenase (HLADH) is studied using a synthetic peptide mimic of this site. The study includes experimental measurements of binding constants, classical free energy calculations from molecular dynamics (MD) simulations and quantum mechanical (QM) electron structure calculations. The classical MD results account for interactions at the molecular level and reproduce the absolute binding energy and the hydration free energy of the Zn ion with an accuracy of about 10%. This is insufficient to obtain correct free energy differences. QM correction terms were calculated from density functional theory (DFT) on small clusters of atoms to include electronic polarisation of the closest waters and covalent contributions to the Zn–S coordination bond. This results in reasonably good agreement with the experimentally measured binding constants and Zn ion hydration free energies in agreement with published experimental values. The study also includes the replacement of one cysteine residue to an alanine. Simulations as well as experiments showed only a small effect of this upon the binding free energy. A detailed analysis indicate that the sulfur is replaced by three water molecules, thereby changing the coordination number of Zn from four (as in the original peptide) to six (as in water).

Atomistic simulations are powerful for probing molecules at bioinorganic interfaces and excellent complements to scarcely available experimental techniques. The free energy controls the adsorption behavior of molecules on nanosurfaces, and is therefore a quantity of particular importance. Advanced sampling techniques can efficiently explore the adsorption free energy landscape, but molecular simulations with classical (Newtownian) dynamics fail to capture charge transfer and polarization at the solid-liquid interface. First principle simulations do not suffer from this limitation but come with a heavy computational load. Here, we introduce an efficient protocol to explore the free energy of adsorption in the ab initio framework. This approach accurately models the complex phenomena at bio-inorganic surfaces on the nanoscale and properly samples the relevant thermodynamic properties. We present a case study of adsorption of the Lysine and Aspartate amino acids on the anatase (101) TiO2 surface with the tight binding method. The high values of the calculated adsorption free energies highlight the importance of a proper description of the electronic state for surface binding processes.

Atomistic simulations are powerful for probing molecules at bioinorganic interfaces and excellent complements to scarcely available experimental techniques. The free energy controls the adsorption behavior of molecules on nanosurfaces, and is therefore a quantity of particular importance. Advanced sampling techniques can efficiently explore the adsorption free energy landscape, but molecular simulations with classical (Newtownian) dynamics fail to capture charge transfer and polarization at the solid-liquid interface. First principle simulations do not suffer from this limitation but come with a heavy computational load. Here, we introduce an efficient protocol to explore the free energy of adsorption in the ab initio framework. This approach accurately models the complex phenomena at bio-inorganic surfaces on the nanoscale and properly samples the relevant thermodynamic properties. We present a case study of adsorption of the Lysine and Aspartate amino acids on the anatase (101) TiO2 surface with the tight binding method. The high values of the calculated adsorption free energies highlight the importance of a proper description of the electronic state for surface binding processes.

Atomistic simulations are powerful for probing molecules at bioinorganic interfaces and excellent complements to scarcely available experimental techniques. The free energy controls the adsorption behavior of molecules on nanosurfaces, and is therefore a quantity of particular importance. Advanced sampling techniques can efficiently explore the adsorption free energy landscape, but molecular simulations with classical (Newtownian) dynamics fail to capture charge transfer and polarization at the solid-liquid interface. First principle simulations do not suffer from this limitation but come with a heavy computational load. Here, we introduce an efficient protocol to explore the free energy of adsorption in the ab initio framework. This approach accurately models the complex phenomena at bio-inorganic surfaces on the nanoscale and properly samples the relevant thermodynamic properties. We present a case study of adsorption of the Lysine and Aspartate amino acids on the anatase (101) TiO2 surface with the tight binding method. The high values of the calculated adsorption free energies highlight the importance of a proper description of the electronic state for surface binding processes.

Atomistic simulations are powerful for probing molecules at bioinorganic interfaces and excellent complements to scarcely available experimental techniques. The free energy controls the adsorption behavior of molecules on nanosurfaces, and is therefore a quantity of particular importance. Advanced sampling techniques can efficiently explore the adsorption free energy landscape, but molecular simulations with classical (Newtownian) dynamics fail to capture charge transfer and polarization at the solid-liquid interface. First principle simulations do not suffer from this limitation but come with a heavy computational load. Here, we introduce an efficient protocol to explore the free energy of adsorption in the ab initio framework. This approach accurately models the complex phenomena at bio-inorganic surfaces on the nanoscale and properly samples the relevant thermodynamic properties. We present a case study of adsorption of the Lysine and Aspartate amino acids on the anatase (101) TiO2 surface with the tight binding method. The high values of the calculated adsorption free energies highlight the importance of a proper description of the electronic state for surface binding processes.

Atomistic simulations are powerful for probing molecules at bioinorganic interfaces and excellent complements to scarcely available experimental techniques. The free energy controls the adsorption behavior of molecules on nanosurfaces, and is therefore a quantity of particular importance. Advanced sampling techniques can efficiently explore the adsorption free energy landscape, but molecular simulations with classical (Newtownian) dynamics fail to capture charge transfer and polarization at the solid-liquid interface. First principle simulations do not suffer from this limitation but come with a heavy computational load. Here, we introduce an efficient protocol to explore the free energy of adsorption in the ab initio framework. This approach accurately models the complex phenomena at bio-inorganic surfaces on the nanoscale and properly samples the relevant thermodynamic properties. We present a case study of adsorption of the Lysine and Aspartate amino acids on the anatase (101) TiO2 surface with the tight binding method. The high values of the calculated adsorption free energies highlight the importance of a proper description of the electronic state for surface binding processes.

Atomistic simulations are powerful for probing molecules at bioinorganic interfaces and excellent complements to scarcely available experimental techniques. The free energy controls the adsorption behavior of molecules on nanosurfaces, and is therefore a quantity of particular importance. Advanced sampling techniques can efficiently explore the adsorption free energy landscape, but molecular simulations with classical (Newtownian) dynamics fail to capture charge transfer and polarization at the solid-liquid interface. First principle simulations do not suffer from this limitation but come with a heavy computational load. Here, we introduce an efficient protocol to explore the free energy of adsorption in the ab initio framework. This approach accurately models the complex phenomena at bio-inorganic surfaces on the nanoscale and properly samples the relevant thermodynamic properties. We present a case study of adsorption of the Lysine and Aspartate amino acids on the anatase (101) TiO2 surface with the tight binding method. The high values of the calculated adsorption free energies highlight the importance of a proper description of the electronic state for surface binding processes.