A5053642050- Nanopore and Nanochannel Transport Studies
- Electrostatics and Colloid Interactions
- Solar Radiation and Photovoltaics
- Solar Thermal and Photovoltaic Systems
- Advanced biosensing and bioanalysis techniques
- Spectroscopy and Quantum Chemical Studies
- Material Dynamics and Properties
- Computational Drug Discovery Methods
- Machine Learning in Materials Science
- Lattice Boltzmann Simulation Studies
- Photovoltaic System Optimization Techniques
- Higher Education Learning Practices
- Distributed and Parallel Computing Systems
- Scientific Computing and Data Management
- Magnetic confinement fusion research
- RNA modifications and cancer
- Metal Alloys Wear and Properties
- RNA Research and Splicing
- RNA and protein synthesis mechanisms
- Protein Structure and Dynamics
- Genetics, Bioinformatics, and Biomedical Research
- Rheology and Fluid Dynamics Studies
- Artificial Intelligence in Law
- RNA Interference and Gene Delivery
- Model Reduction and Neural Networks
Forschungszentrum Jülich
2022-2025
Deutsches Zentrum für Luft- und Raumfahrt e. V. (DLR)
2023-2024
Physical Sciences (United States)
2023
Karlsruhe Institute of Technology
2023
Helmholtz Institute Mainz
2023
University of Stuttgart
2010-2016
Stuttgart Observatory
2014
RWTH Aachen University
2010
Large language models have greatly enhanced our ability to understand biology and chemistry, yet robust methods for structure-based drug discovery, quantum chemistry structural are still sparse. Precise biomolecule-ligand interaction datasets urgently needed large models. To address this, we present MISATO, a dataset that combines mechanical properties of small molecules associated molecular dynamics simulations ~20,000 experimental protein-ligand complexes with extensive validation data....
We present a detailed investigation of the ionic current in cylindrical model nanopore absence and presence double stranded DNA homopolymer. Our atomistic simulations are capable reproducing almost exactly experimental data obtained by Smeets et al., including notably crossover salt concentration that yields equal measurements both situations. can rule out observed blockade is due to steric exclusion charge carriers from DNA, since for all investigated concentrations carrier density higher...
Coarse-grained models of soft matter are usually combined with implicit solvent that take the electrostatic polarizability into account via a dielectric background. In biophysical or nanoscale simulations include water, this constant can vary greatly within system. Performing molecular dynamics other need to compute exact interactions between charges in those systems is computationally demanding. We review here several algorithms developed by us perform exactly task. For planar surfaces...
Large language models (LLMs) have greatly enhanced our ability to understand biology and chemistry. Yet, relatively few robust methods been reported for structure-based drug discovery. Highly precise biomolecule-ligand interaction datasets are urgently needed in particular LLMs, that require extensive training data. We present MISATO, the first dataset combines quantum mechanics properties of small molecules associated molecular dynamics simulations about 20000 experimental protein-ligand...
Abstract Concentrating solar power plants are a clean energy source capable of competitive electricity generation even during night time, as well the production carbon-neutral fuels, offering complementary role alongside photovoltaic plants. In these plants, thousands mirrors (heliostats) redirect sunlight onto receiver, potentially generating temperatures exceeding 1000°C. Practically, such efficient never attained. Several unknown, yet operationally crucial parameters, e.g., misalignment...
This paper investigates scalable neural networks with learnable activation functions based on orthogonal function bases and tropical polynomials, targeting ImageNet-1K classification next token prediction OpenWebText. Traditional activations, such as ReLU, are static. In contrast, activations enable the network to adapt dynamically during training. However, stability issues, vanishing or exploding gradients, arise improper variance management in deeper networks. To remedy this, we propose an...
Earth observation (EO) yields large-scale, multimodal datasets collected from various satellite missions, DEMs, land-use data, and textual metadata. Foundation models like 4M (Massively Multimodal Masked Modeling) can learn a joint embedding space that bridges modality gaps, mitigates missing data issues, facilitates partial spatio-temporal alignment [1]. However, directly training such foundation on the vast, high-dimensional original EO is not only computationally intensive but also...
We performed MD simulations of heterogeneous crystallization soft colloids focusing on the effect hydrodynamic interactions. found that particle attachment is not purely diffusive as commonly assumed.
The camera target method is the most commonly used calibration for heliostats at solar tower power plants to minimize their sun tracking errors. In this method, individual are moved a white surface and deviation from targeted position measured. A regression calculate errors in geometry model tabular data obtained way. For modern aim point strategies, or simply rearmost end of field, extremely high accuracies needed, which can only be achieved by many degrees freedom model. problem here that...
Based on a coarse-grained model, we carry out molecular dynamics simulations to analyze the diffusion of small tracer particle inside cylindrical channel whose inner wall is covered with randomly grafted short polymeric chains. We observe an interesting transient subdiffusive behavior along axis at high attraction between and chains, however, long-time always normal. This process found be enhanced for case that immobilize i.e., sets in earlier time spans over longer period before becoming...
Abstract The resolution of chiral compound‐forming systems using hybrid processes was discussed recently. concept is large relevance as these form the majority substances. In this study, a novel process presented, which combines pertraction and subsequent preferential crystallization applicable for such systems. A supported liquid membrane applied in provides enantiomeric enrichment. This contains solution compound acting selective carrier one enantiomers. Screening number membranes...
On the path to full understanding of structure-function relationship or even design RNA, structure prediction would offer an intriguing complement experimental efforts. Any deep learning on RNA structure, however, is hampered by sparsity labeled training data. Utilizing limited data available, we here focus predicting spatial adjacencies ("contact maps") as a proxy for 3D structure. Our model, BARNACLE, combines utilization unlabeled through self-supervised pre-training and efficient use...
We investigate the impact of dielectric boundary forces on translocation process stiff charged macromolecules through solid-state nanopores by means coarse grained molecular dynamics simulations. find that low permittivity typical pore materials (compared to surrounding water) is able create a high free energy barrier more than 10kBT under salt conditions for radii comparable Debye length. Interestingly significantly enhanced flexible macromolecules. To achieve description profile we...
We present an implicit solvent coarse-grained double-stranded DNA (dsDNA) model confined to infinite cylindrical pore that reproduces the experimentally observed current modulations of a KaCl solution at various concentrations. Our extends previous and mean-field approaches by incorporating position dependent friction term on ions, which Kesselheim et al. [Phys. Rev. Lett. 112, 018101 (2014)] identified as essential ingredient correctly reproduce experimental data Smeets [Nano 6, 89 (2006)]....
Abstract The ability to differentiate between viable and dead microorganisms in metagenomic samples is crucial for various microbial inferences, ranging from assessing ecosystem functions of environmental microbiomes inferring the virulence potential pathogens. While established viability-resolved approaches are labor-intensive as well biased lacking sensitivity, we here introduce a new fully computational framework that leverages nanopore sequencing technology assess viability directly...
Conducting a current through nanopore allows for the analysis of molecules inside pore because modulation caused by electrostatic properties passing can be measured. This mechanism shows great potential DNA sequencing, as four different nucleotide bases induce modulations. We present visualisation approach to investigate this phenomenon in our simulations within combining state-of-the-art molecular with vector field illustration. By spatial and temporal aggregation ions transported pore, we...
<title>Abstract</title> Solar tower power plants play a key role to facilitate the ongoing energy transition as they deliver climate neutral electricity and direct heat for chemical processes. These generate temperatures over 1000 °C by reflecting sunlight with thousands of mirrors (heliostats) receiver. The temper- ature achievable in practice is limited due system’s susceptibility small surface defects misalignments individual mirrors, hindering plant’s full efficiency. We present an...
The Derjaguin-Landau-Verwey-Overbeek (DLVO) theory, introduced more than 70 years ago, is a hallmark of colloidal particle modeling. For highly charged particles in the dilute regime, it often supplemented by Alexander's prescription (Alexander et al. J Chem Phys 80:5776, 1984) for using renormalized charge. Here, we solve problem interaction between two colloids at finite ionic strength, including dielectric mismatch effects, an efficient numerical scheme to nonlinear Poisson-Boltzmann...
The recent success of Large Language Models (LLMs) has been predominantly driven by curating the training dataset composition, scaling model architectures and sizes advancements in pretraining objectives, leaving tokenizer influence as a blind spot. Shedding light on this underexplored area, we conduct comprehensive study choice LLM downstream performance 24 mono- multilingual LLMs at 2.6B parameter scale, ablating different algorithms parameterizations. Our studies highlight that can...
We employ a fast and accurate algorithm to treat dielectric interfaces within molecular dynamics simulations demonstrate the importance of boundary forces (DBFs) in two systems interests soft-condensed matter science. investigate salt solution confined slit pore, model DNA fragment translocating thorugh narrow pore.