Zuse Institute Berlin

Sikeler C., Kempter S., Sekulic I., Burger S., Liedl T.

Zhong X., Höfling F., John T.

AbstractGarnet has been widely used to decipher the pressure‐temperature‐time history of rocks, but its physical properties such as elasticity and diffusion are strongly affected by trace amounts of hydrogen. Experimental measurements of H diffusion in garnet are limited to room pressure. We use atomistic simulations to study H diffusion in perfect and defective garnet lattices, focusing on protonation defects at the Si and Mg sites, which are shown to be energetically favored. Transient trapping of H renders ab‐initio simulations of H diffusion computationally challenging, which is overcome with machine learning techniques by training a deep neural network that encodes the interatomic potential. Our results from such deep potential molecular dynamics (DeePMD) simulations show high mobility of hydrogen in defect‐free garnet lattices, whereas H diffusivity is significantly diminished in defective lattices. Tracer simulations focusing on H alone highlight the vital role of atomic vibrations of heavier atoms like Mg on the release of H atoms. Two regimes of H diffusion are identified: a diffuser‐dominated regime at high hydrogen content with low activation energies due to saturation of vacancies by hydrogen, and a vacancy‐dominated regime at low hydrogen content with high activation energies due to trapping of H atoms at vacancy sites. These regimes account for experimental observations, such as a H‐concentration dependent diffusivity and the discrepancy in activation energy between deprotonation and D‐H exchange experiments. This study underpins the crucial role of vacancies in H diffusion and demonstrates the utility of machine‐learned interatomic potentials in studying kinetic processes in the Earth's interior.

Shimada N., Sato M., Yaegashi Y., Koch T., Le T.T.

Nagel S., Heitzig J., Schöll E.

We present a stochastic dynamic model which can explain economic cycles. We show that the macroscopic description yields a complex dynamical landscape consisting of multiple stable fixed points, each corresponding to a split of the population into a large low and a small high income group. The stochastic fluctuations induce switching between the resulting metastable states and excitation oscillations just below a deterministic bifurcation. The shocks are caused by the decisions of a few agents who have a disproportionate influence over the macroscopic state of the economy due to the unequal distribution of wealth among the population. The fluctuations have a long-term effect on the growth of economic output and lead to business cycle oscillations exhibiting coherence resonance, where the correlation time is controlled by the population size which is inversely proportional to the noise intensity. Published by the American Physical Society 2025

Jain T., Singh U., Singh V., Boda V.K., Hotz I., Vadhiyar S.S., Vinayachandran P.N., Natarajan V.

AbstractOceanographers rely on visual analysis to interpret model simulations, identify events and phenomena, and track dynamic ocean processes. The ever increasing resolution and complexity of ocean data due to its dynamic nature and multivariate relationships demands a scalable and adaptable visualization tool for interactive exploration. We introduce pyParaOcean, a scalable and interactive visualization system designed specifically for ocean data analysis. pyParaOcean offers specialized modules for common oceanographic analysis tasks, including eddy identification and salinity movement tracking. These modules seamlessly integrate with ParaView as filters, ensuring a user‐friendly and easy‐to‐use system while leveraging the parallelization capabilities of ParaView and a plethora of inbuilt general‐purpose visualization functionalities. The creation of an auxiliary dataset stored as a Cinema database helps address I/O and network bandwidth bottlenecks while supporting the generation of quick overview visualizations. We present a case study on the Bay of Bengal to demonstrate the utility of the system and scaling studies to evaluate the efficiency of the system.

Mischau A., Weber M.

Trepczynski A., Kneifel P., Heyland M., Leskovar M., Moewis P., Damm P., Taylor W.R., Zachow S., Duda G.N.

IntroductionAnterior knee pain and other patello-femoral (PF) complications frequently limit the success of total knee arthroplasty as the final treatment of end stage osteoarthritis. However, knowledge about the in-vivo loading conditions at the PF joint remains limited, as no direct measurements are available. We hypothesised that the external knee flexion moment (EFM) is highly predictive of the PF contact forces during activities with substantial flexion of the loaded knee.Materials and methodsSix patients (65–80 years, 67–101 kg) with total knee arthroplasty (TKA) performed two activities of daily living: sit-stand-sit and squat. Tibio-femoral (TF) contact forces were measured in vivo using instrumented tibial components, while synchronously internal TF and PF kinematics were captured with mobile fluoroscopy. The measurements were used to compute PF contact forces using patient specific musculoskeletal models. The relationship between the EFM and the PF contact force was quantified using linear regression.ResultsMean peak TF contact forces of 1.97–3.24 times body weight (BW) were found while peak PF forces reached 1.75 to 3.29 times body weight (BW). The peak EFM ranged from 3.2 to 5.9 %BW times body height, and was a good predictor of the PF contact force (R2 = 0.95 and 0.88 for sit-stand-sit and squat, respectively).DiscussionThe novel combination of in vivo TF contact forces and internal patellar kinematics enabled a reliable assessment of PF contact forces. The results of the regression analysis suggest that PF forces can be estimated based solely on the EFM from quantitative gait analysis. Our study also demonstrates the relevance of PF contact forces, which reach magnitudes similar to TF forces during activities of daily living.

Berns M.M., Yildiz M., Winkelmann S., Walter A.M.

AbstractSynaptic vesicle (SV) trafficking toward the plasma membrane (PM) and subsequent SV maturation are essential for neurotransmitter release. These processes, including SV docking and priming, are co‐ordinated by various proteins, such as SNAREs, Munc13 and synaptotagmin (Syt), which connect (tether) the SV to the PM. Here, we investigated how tethers of varying lengths mediate SV docking using a simplified mathematical model. The heights of the three tether types, as estimated from the structures of the SNARE complex, Munc13 and Syt, defined the SV–PM distance ranges for tether formation. Geometric considerations linked SV–PM distances to the probability and rate of tether formation. We assumed that SV tethering constrains SV motility and that multiple tethers are associated by independent interactions. The model predicted that forming multiple tethers favours shorter SV–PM distances. Although tethers acted independently in the model, their geometrical properties often caused their sequential assembly, from longer ones (Munc13/Syt), which accelerated SV movement towards the PM, to shorter ones (SNAREs), which stabilized PM‐proximal SVs. Modifying tether lengths or numbers affected SV trafficking. The independent implementation of tethering proteins enabled their selective removal to mimic gene knockout (KO) situations. This showed that simulated SV–PM distance distributions qualitatively aligned with published electron microscopy studies upon removal of SNARE and Syt tethers, whereas Munc13 KO data were best approximated when assuming additional disruption of SNARE tethers. Thus, although salient features of SV docking can be accounted for by independent tethering alone, our results suggest that functional tether interactions not yet featured in our model are crucial for biological function. imageKey points A mathematical model describing the role of synaptic protein tethers to localize transmitter‐containing vesicles is developed based on geometrical considerations and structural information of synaptotagmin, Munc13 and SNARE proteins. Vesicle movement, along with tether association and dissociation, are modelled as stochastic processes, with tethers functioning independently of each other. Multiple tethers cooperate to recruit vesicles to the plasma membrane and keep them there: Munc13 and Syt as the longer tethers accelerate the movement towards the membrane, whereas short SNARE tethers stabilize them there. Model predictions for situations in which individual tethers are removed agree with the results from experimental studies upon gene knockout. Changing tether length or copy numbers affects vesicle trafficking and steady‐state distributions.

Djurdjevac Conrad N., Tonello E., Zonker J., Siebert H.

AbstractTemporal networks are a powerful tool for studying the dynamic nature of a wide range of real-world complex systems, including social, biological and physical systems. In particular, detection of dynamic communities within these networks can help identify important cohesive structures and fundamental mechanisms driving systems behaviour. However, when working with real-world systems, available data is often limited and sparse, due to missing data on systems entities, their evolution and interactions, as well as uncertainty regarding temporal resolution. This can hinder accurate representation of the system over time and result in incomplete or biased community dynamics. In this paper, we consider established methods for community detection and, using synthetic data experiments and real-world case studies, we evaluate the impact of data sparsity on the quality of identified dynamic communities. Our results give valuable insights on the evolution of systems with sparse data, which are less studied in existing literature, but are frequently encountered in real-world applications.

del Razo M., Delle Site L.

A varying number of particles is one of the most relevant characteristics of systems of interest in nature and technology, ranging from the exchange of energy and matter with the surrounding environment to the change of particle number through internal dynamics such as reactions. The physico-mathematical modeling of these systems is extremely challenging, with the major difficulty being the time dependence of the number of degrees of freedom and the additional constraint that the increment or reduction of the number and species of particles must not violate basic physical laws. Theoretical models, in such a case, represent the key tool for the design of computational strategies for numerical studies that deliver trustful results. In this manuscript, we review complementary physico-mathematical approaches of varying number of particles inspired by rather different specific numerical goals. As a result of the analysis on the underlying common structure of these models, we propose a unifying master equation for general dynamical systems with varying number of particles. This equation embeds all the previous models and can potentially model a much larger range of complex systems, ranging from molecular to social agent-based dynamics.

Surapaneni V.A., Flaum B., Schindler M., Hayat K., Wölfer J., Baum D., Hu R., Kong T.F., Doube M., Dean M.N.

AbstractAmong hornbill birds, the critically endangered helmeted hornbill (Rhinoplax vigil) is notable for its casque (a bulbous beak protrusion) being filled with trabeculae and fronted by a very thick keratin layer. Casque function is debated but appears central to aerial jousting, where birds (typically males) collide casques at high speeds in a mid‐flight display that is audible for more than 100 m. We characterized the structural relationship between the skull and casque anatomy using X‐ray microtomography and quantitative trabecular network analysis to examine how the casque sustains extreme impact. The casque comprises a keratin veneer (rhamphotheca, ∼8× thicker than beak keratin), which slots over the internal bony casque like a tight‐fitting sheath. The bony casque's central cavity contains a network of trabeculae—heavily aligned and predominantly rod‐like, among the thickest described in vertebrates—forming a massive rostrocaudal strut spanning the casque's length, bridging rostral (impact), and caudal (braincase) surfaces. Quantitative network characterizations indicate no differences between male and female trabecular architectures. This suggests that females may also joust or that casques play other roles. Our results argue that the casque's impact loading demands and shapes a high‐safety‐factor construction that involves extreme trabecular morphologies among vertebrates, architectures that also have the potential for informing the design of collision‐resistant materials.

Trower M., Djurdjevac Conrad N., Klus S.

Time-evolving graphs arise frequently when modeling complex dynamical systems such as social networks, traffic flow, and biological processes. Developing techniques to identify and analyze communities in these time-varying graph structures is an important challenge. In this work, we generalize existing spectral clustering algorithms from static to dynamic graphs using canonical correlation analysis to capture the temporal evolution of clusters. Based on this extended canonical correlation framework, we define the spatiotemporal graph Laplacian and investigate its spectral properties. We connect these concepts to dynamical systems theory via transfer operators and illustrate the advantages of our method on benchmark graphs by comparison with existing methods. We show that the spatiotemporal graph Laplacian allows for a clear interpretation of cluster structure evolution over time for directed and undirected graphs.

Hege H., Chen M.

Vu-Han T., Weiß C., Köhli P., Schönnagel L., Perka C., Pumberger M.

Abstract Purpose 5q-spinal muscular atrophy (SMA) is a treatable neuromuscular disorder associated with scoliosis in up to 90% of patients. New SMA therapies could mark a paradigm shift in scoliosis management, but their effects on scoliosis development remain unclear. This study aims to observe scoliosis progression in the current treatment landscape to inform management strategies. Methods We conducted a cross-sectional retrospective analysis of 94 SMA patients treated at our center. Scoliosis development was evaluated in 75 patients using spine radiographs and electronic health records. Statistical analysis was performed using Python and GraphPad Prism. One-way ANOVA and Pearson correlation were used for group comparisons and correlation analysis, respectively. Results Scoliosis parameters in 5q-SMA patients who had received either nusinersen, onasemnogene abeparvovec, risdiplam, or their combinations showed mean ages at scoliosis detection were 23.94, 55.52, and 168.11 months for SMA types 1, 2, and 3, respectively. Cobb angles at detection showed no significant intergroup differences. The mean ages at scoliosis surgery were 60, 88.43, and 124.8 months. Pelvic obliquity (PO) was highest in type 1 and lowest in type 3. A strong correlation (r = 0.9) was found between PO measurement techniques. HFMSE scores correlated moderately with scoliosis severity (r = -0.38), while CHOP-INTEND showed no correlation. Conclusion The observations made in this study suggest that the effects of SMA therapies do not prevent scoliosis development. The improved prognosis may lead to a growing cohort of SMA type 1 and 2 patients with early onset scoliosis who require early growth-friendly surgical interventions.

Mayer J., Baum D., Ambellan F., von Tycowicz C.

AbstractShape analysis provides methods for understanding anatomical structures extracted from medical images. However, the underlying notions of shape spaces that are frequently employed come with strict assumptions prohibiting the analysis of incomplete and/or topologically varying shapes. This work aims to alleviate these limitations by adapting the concept of functional maps. Further, we present a graph-based learning approach for morphometric classification of disease states that uses novel shape descriptors based on this concept. We demonstrate the performance of the derived classifier on the open-access ADNI database differentiating normal controls and subjects with Alzheimer’s disease. Notably, the experiments show that our approach can improve over state-of-the-art from geometric deep learning.