@article{19844, abstract = {Context. Rapidly rotating classical OBe stars have been proposed as the products of binary interactions, and the fraction of Be stars with compact companions implies that at least some are. However, to constrain the interaction physics spinning up the OBe stars, a large sample of homogeneously analyzed OBe stars with well-determined binary characteristics and orbital parameters are required. Aims. We investigated the multiplicity properties of a sample of 18 Oe, 62 Be, and two Of?p stars observed within the BLOeM survey in the Small Magellanic Cloud. We analyzed the first nine epochs of spectroscopic observations obtained over approximately three months in 2023. Methods. Radial velocities (RVs) of all stars were measured using cross-correlation based on different sets of absorption and emission lines. Applying commonly used binarity criteria, we classified objects as binaries, binary candidates, and apparently single (RV stable) objects. We further inspected the spectra for double-lined spectroscopic binaries and cross-matched with catalogs of X-ray sources and photometric binaries. Results. We classify 14 OBe stars as binaries, and an additional 11 as binary candidates. The two Of?p stars are apparently single. We find two more objects that are most likely currently interacting binaries. Without those, the observed binary fraction for the remaining OBe sample of 78 stars is fobs+candOBe = 0.18 ± 0.04 (fobs+candOBe = 0.32±0.05 including candidates). This binary fraction is less than half of that measured for OB stars in BLOeM. Combined with the lower fraction of SB2s, this suggests that OBe stars do indeed have fundamentally different present-day binary properties than OB stars. We find no evidence for OBe binaries with massive compact companions, in contrast to expectations from binary population synthesis. Conclusions. Our results support the binary scenario as an important formation channel for OBe stars, as post-interaction binaries may have been disrupted or the stripped companions of OBe stars are harder to detect. Further observations are required to characterize the detected binaries, their orbital parameters, and the nature of their companions.}, author = {Bodensteiner, J. and Shenar, T. and Sana, H. and Britavskiy, N. and Crowther, P. A. and Langer, N. and Lennon, D. J. and Mahy, L. and Patrick, L. R. and Villaseñor, J. I. and Abdul-Masih, M. and Bowman, D. M. and De Koter, A. and De Mink, S. E. and Deshmukh, K. and Fabry, M. and Gilkis, A. and Götberg, Ylva Louise Linsdotter and Holgado, G. and Izzard, R. G. and Janssens, S. and Kalari, V. M. and Keszthelyi, Z. and Kubát, J. and Mandel, I. and Maravelias, G. and Oskinova, L. M. and Pauli, D. and Ramachandran, V. and Rocha, D. F. and Renzo, M. and Sander, A. A.C. and Schneider, F. R.N. and Schootemeijer, A. and Sen, K. and Stoop, M. and Toonen, S. and Van Loon, J. T. and Valli, R. and Vigna-Gómez, A. and Vink, J. S. and Wang, C. and Xu, X. T.}, issn = {1432-0746}, journal = {Astronomy & Astrophysics}, publisher = {EDP Sciences}, title = {{Binarity at LOw Metallicity (BLOeM): Multiplicity properties of Oe and Be stars}}, doi = {10.1051/0004-6361/202452623}, volume = {698}, year = {2025}, } @article{19842, abstract = {Given the uncertain evolutionary status of blue supergiant stars, their multiplicity properties hold vital clues to better understand their origin and evolution. As part of The Binarity at LOw Metallicity (BLOeM) campaign in the Small Magellanic Cloud, we present a multi-epoch spectroscopic survey of 128 supergiant stars of spectral type B5–F5, which roughly correspond to initial masses in the 6–30 M⊙ range. The observed binary fraction for the B5–9 supergiants is 25 ± 6% (10 ± 4%) and 5 ± 2% (0%) for the A–F stars, which were found using a radial-velocity (RV) variability threshold of 5 km s−1 (10 km s−1) as a criterion for binarity. Accounting for observational biases, we find an intrinsic multiplicity fraction of less than 18% for the B5–9 stars and 8−7+9% for the AF stars, for the orbital periods up to 103.5 days and mass ratios (q) in the 0.1 < q < 1 range. The large stellar radii of these supergiant stars prevent short orbital periods, but we demonstrate that this effect alone cannot explain our results. We assessed the spectra and RV time series of the detected binary systems and find that only a small fraction display convincing solutions. We conclude that the multiplicity fractions are compromised by intrinsic stellar variability, such that the true multiplicity fraction may be significantly smaller. Our main conclusions from comparing the multiplicity properties of the B5–9- and AF-type supergiants to that of their less evolved counterparts is that such stars cannot be explained by a direct evolution from the main sequence. Furthermore, by comparing their multiplicity properties to red supergiant stars, we conclude that the AF supergiant stars are neither progenitors nor descendants of red supergiants.}, author = {Patrick, L. R. and Lennon, D. J. and Najarro, F. and Shenar, T. and Bodensteiner, J. and Sana, H. and Crowther, P. A. and Britavskiy, N. and Langer, N. and Schootemeijer, A. and Evans, C. J. and Mahy, L. and Götberg, Ylva Louise Linsdotter and De Mink, S. E. and Schneider, F. R.N. and O’Grady, A. J.G. and Villaseñor, J. I. and Bernini-Peron, M. and Bowman, D. M. and De Koter, A. and Deshmukh, K. and Gilkis, A. and González-Torà, G. and Kalari, V. M. and K̃Eszthelyi, Z. and Mandel, I. and Menon, A. and Moe, M. and Oskinova, L. M. and Pauli, D. and Renzo, M. and Sander, A. A.C. and Sen, K. and Stoop, M. and Van Loon, J. T. and Toonen, S. and Tramper, F. and Vink, J. S. and Wang, C.}, issn = {1432-0746}, journal = {Astronomy & Astrophysics}, publisher = {EDP Sciences}, title = {{Binarity at LOw Metallicity (BLOeM): The multiplicity properties and evolution of BAF-type supergiants}}, doi = {10.1051/0004-6361/202452949}, volume = {698}, year = {2025}, } @article{20352, abstract = {At high metallicity, a majority of massive stars have at least one close stellar companion. The evolution of such binaries is subject to strong interaction processes, which heavily impact the characteristics of their life-ending supernova and compact remnants. For the low-metallicity environments of high-redshift galaxies, constraints on the multiplicity properties of massive stars over the separation range leading to binary interaction are crucially missing. Here we show that the presence of massive stars in close binaries is ubiquitous, even at low metallicity. Using the Very Large Telescope, we obtained multi-epoch radial velocity measurements of a representative sample of 139 massive O-type stars across the Small Magellanic Cloud, which has a metal content of about one-fifth of the solar value. We find that 45% of them show radial velocity variations that demonstrate that they are members of close binary systems, and predominantly have orbital periods shorter than 1 year. Correcting for observational biases indicates that at least 70+11−6 % of the O stars in our sample are in close binaries, and that at least 68+7 −8% of all O stars interact with a companion star during their lifetime. We found no evidence supporting a statistically significant trend of the multiplicity properties with metallicity. Our results indicate that multiplicity and binary interactions govern the evolution of massive stars and determine their cosmic feedback and explosive fates.}, author = {Sana, H. and Shenar, T. and Bodensteiner, J. and Britavskiy, N. and Langer, N. and Lennon, D. J. and Mahy, L. and Mandel, I. and De Mink, S. E. and Patrick, L. R. and Villaseñor, J. I. and Dirickx, M. and Abdul-Masih, M. and Almeida, L. A. and Backs, F. and Berlanas, S. R. and Bernini-Peron, M. and Bowman, D. M. and Bronner, V. A. and Crowther, P. A. and Deshmukh, K. and Evans, C. J. and Fabry, M. and Gieles, M. and Gilkis, A. and González-Torà, G. and Gräfener, G. and Götberg, Ylva Louise Linsdotter and Hawcroft, C. and Hénault-Brunet, V. and Herrero, A. and Holgado, G. and Izzard, R. G. and De Koter, A. and Janssens, S. and Johnston, C. and Josiek, J. and Justham, S. and Kalari, V. M. and Klencki, J. and Kubát, J. and Kubátová, B. and Lefever, R. R. and Van Loon, J. Th and Ludwig, B. and Mackey, J. and Maíz Apellániz, J. and Maravelias, G. and Marchant, P. and Mazeh, T. and Menon, A. and Moe, M. and Najarro, F. and Oskinova, L. M. and Ovadia, R. and Pauli, D. and Pawlak, M. and Ramachandran, V. and Renzo, M. and Rocha, D. F. and Sander, A. A.C. and Schneider, F. R.N. and Schootemeijer, A. and Schösser, E. C. and Schürmann, C. and Sen, K. and Shahaf, S. and Simón-Díaz, S. and Van Son, L. A.C. and Stoop, M. and Toonen, S. and Tramper, F. and Valli, R. and Vigna-Gómez, A. and Vink, J. S. and Wang, C. and Willcox, R.}, issn = {2397-3366}, journal = {Nature Astronomy}, pages = {1337--1346}, publisher = {Springer Nature}, title = {{A high fraction of close massive binary stars at low metallicity}}, doi = {10.1038/s41550-025-02610-x}, volume = {9}, year = {2025}, } @unpublished{19546, abstract = {We study the sensitivity of the eigenvectors of random matrices, showing that even small perturbations make the eigenvectors almost orthogonal. More precisely, we consider two deformed Wigner matrices $W+D_1$, $W+D_2$ and show that their bulk eigenvectors become asymptotically orthogonal as soon as $\mathrm{Tr}(D_1-D_2)^2\gg 1$, or their respective energies are separated on a scale much bigger than the local eigenvalue spacing. Furthermore, we show that quadratic forms of eigenvectors of $W+D_1$, $W+D_2$ with any deterministic matrix $A\in\mathbf{C}^{N\times N}$ in a specific subspace of codimension one are of size $N^{-1/2}$. This proves a generalization of the Eigenstate Thermalization Hypothesis to eigenvectors belonging to two different spectral families.}, author = {Cipolloni, Giorgio and Erdös, László and Henheik, Sven Joscha and Kolupaiev, Oleksii}, booktitle = {arXiv}, title = {{Eigenvector decorrelation for random matrices}}, doi = {10.48550/arXiv.2410.10718}, year = {2025}, } @inproceedings{22827, abstract = {Quantized training of Large Language Models (LLMs) remains an open challenge, as maintaining accuracy while performing all matrix multiplications in low precision has proven difficult. This is particularly the case when fine-tuning pre-trained models, which can have large weight, activation, and error (output gradient) outlier values that make lower-precision optimization difficult. To address this, we present HALO, a new quantization-aware training approach for Transformers that enables accurate and efficient low-precision training by combining 1) strategic placement of Hadamard rotations in both forward and backward passes, which mitigate outliers, 2) high-performance kernel support, and 3) FSDP integration for low-precision communication. Our approach ensures that all large matrix multiplications during the forward and backward passes are executed in lower precision. Applied to LLaMa models, HALO achieves near-full-precision-equivalent results during fine-tuning on various tasks, while delivering up to 1.41x end-to-end speedup for full fine-tuning on RTX 4090 GPUs. HALO efficiently supports both standard and parameter-efficient fine-tuning (PEFT). Our results demonstrate the first practical approach to fully quantized LLM fine-tuning that maintains accuracy in INT8 and FP6 precision, while delivering performance benefits.}, author = {Ashkboos, Saleh and Nikdan, Mahdi and Tabesh, Soroush and Lopez Castro, Roberto and Hoefler, Torsten and Alistarh, Dan-Adrian}, booktitle = {39th Conference on Neural Information Processing Systems}, isbn = {9798331338275}, issn = {1049-5258}, location = {San Diego, CA, United States}, pages = {131755--131780}, publisher = {Neural Information Processing Systems Foundation}, title = {{HALO: Hadamard-assisted lower-precision optimization for LLMs}}, doi = {10.52202/085713-3966}, volume = {38}, year = {2025}, } @inproceedings{22832, abstract = {In many scientific experiments, the data annotating cost constraints the pace for testing novel hypotheses. Yet, modern machine learning pipelines offer a promising solution—provided their predictions yield correct conclusions. We focus on Prediction-Powered Causal Inferences (PPCI), i.e., estimating the treatment effect in an unlabeled target experiment, relying on training data with the same outcome annotated but potentially different treatment or effect modifiers. We first show that conditional calibration guarantees valid PPCI at population level. Then, we introduce a sufficient representation constraint transferring validity across experiments, which we propose to enforce in practice in Deconfounded Empirical Risk Minimization, our new model-agnostic training objective. We validate our method on synthetic and real-world scientific data, solving impossible problem instances for Empirical Risk Minimization even with standard invariance constraints. In particular, for the first time, we achieve valid causal inference on a scientific experiment with complex recording and no human annotations, fine-tuning a foundational model on our similar annotated experiment.}, author = {Cadei, Riccardo and Demirel, Ilker and De Bartolomeis, Piersilvio and Lindorfer, Lukas and Cremer, Sylvia and Schmid, Cordelia and Locatello, Francesco}, booktitle = {39th Conference on Neural Information Processing Systems}, isbn = {9798331338275}, issn = {1049-5258}, location = {San Diego, CA, United States}, pages = {82200--82229}, publisher = {Neural Information Processing Systems Foundation}, title = {{Prediction-Powered Causal Inferences}}, doi = {10.52202/085713-2479}, volume = {38}, year = {2025}, } @inproceedings{22831, abstract = {Neural models learn representations of high-dimensional data on low-dimensional manifolds. Multiple factors, including stochasticities in the training process, model architectures, and additional inductive biases, may induce different representations, even when learning the same task on the same data. However, it has recently been shown that when a latent structure is shared between distinct latent spaces, relative distances between representations can be preserved, up to distortions. Building on this idea, we demonstrate that exploiting the differential-geometric structure of latent spaces of neural models, it is possible to capture precisely the transformations between representational spaces trained on similar data distributions. Specifically, we assume that distinct neural models parametrize approximately the same underlying manifold, and introduce a representation based on the pullback metric that captures the intrinsic structure of the latent space, while scaling efficiently to large models. We validate experimentally our method on model stitching and retrieval tasks, covering autoencoders and vision foundation discriminative models, across diverse architectures, datasets, pretraining schemes and modalities. Code is available at https://github.com/marc0git/RelativeGeodesics.}, author = {Yu, Hanlin and Inal, Berfin and Arvanitidis, Georgios and Hauberg, Søren and Locatello, Francesco and Fumero, Marco}, booktitle = {39th Conference on Neural Information Processing Systems}, isbn = {9798331338275}, issn = {1049-5258}, location = {San Diego, CA, United States}, pages = {125316--125360}, publisher = {Neural Information Processing Systems Foundation}, title = {{Connecting neural models latent geometries with relative geodesic representations}}, doi = {10.52202/085713-3769}, volume = {38}, year = {2025}, } @article{22929, abstract = {The circle method has been successfully used over the last century to study rational points on hypersurfaces. More recently, a version of the method over function fields, combined with spreading out techniques, has led to a range of results about moduli spaces of rational curves on hypersurfaces. In this paper a version of the circle method is implemented in the setting of the Grothendieck ring of varieties. This allows us to approximate the classes of these moduli spaces directly, without relying on point counting, and leads to a deeper understanding of their geometry.}, author = {Bilu, Margaret and Browning, Timothy D}, issn = {1873-2151}, journal = {Annales Scientifiques de l’École Normale Supérieure}, keywords = {Circle method, moduli spaces of curves, hypersurfaces, Grothendieck ring of varieties, motivic integration}, number = {5}, pages = {1179--1242}, publisher = {Société Mathématique de France}, title = {{A motivic circle method}}, doi = {10.24033/asens.2628}, volume = {58}, year = {2025}, } @article{18074, abstract = {The Aharonov–Casher theorem is a result on the number of the so-called zero modes of a system described by the magnetic Pauli operator in R2. In this paper we address the same question for the Dirac operator on a flat two-dimensional manifold with boundary and Atiyah–Patodi–Singer boundary condition. More concretely we are interested in the plane and a disc with a finite number of circular holes cut out. We consider a smooth compactly supported magnetic field on the manifold and an arbitrary magnetic field inside the holes.}, author = {Fialova, Marie}, issn = {1424-0637}, journal = {Annales Henri Poincare}, pages = {2859--2900}, publisher = {Springer Nature}, title = {{Aharonov–Casher theorems for Dirac operators on manifolds with boundary and APS boundary condition}}, doi = {10.1007/s00023-024-01482-7}, volume = {26}, year = {2025}, } @article{17293, abstract = {Voltage-gated CaV2.1 (P/Q-type) Ca2+ channels play a crucial role in regulating neurotransmitter release, thus contributing to synaptic plasticity and to processes such as learning and memory. Despite their recognized importance in neural function, there is limited information on their potential involvement in neurodegenerative conditions such as Alzheimer's disease (AD). Here, we aimed to explore the impact of AD pathology on the density and nanoscale compartmentalization of CaV2.1 channels in the hippocampus in association with GABAB receptors. Histoblotting experiments showed that the density of CaV2.1 channel was significantly reduced in the hippocampus of APP/PS1 mice in a laminar-dependent manner. CaV2.1 channel was enriched in the active zone of the axon terminals and was present at a very low density over the surface of dendritic tree of the CA1 pyramidal cells, as shown by quantitative SDS-digested freeze-fracture replica labelling (SDS-FRL). In APP/PS1 mice, the density of CaV2.1 channel in the active zone was significantly reduced in the strata radiatum and lacunosum-moleculare, while it remained unaltered in the stratum oriens. The decline in Cav2.1 channel density was found to be associated with a corresponding impairment in the GABAergic synaptic function, as evidenced by electrophysiological experiments carried out in the hippocampus of APP/PS1 mice. Remarkably, double SDS-FRL showed a co-clustering of CaV2.1 channel and GABAB1 receptor in nanodomains (~40–50 nm) in wild type mice, while in APP/PS1 mice this nanoarchitecture was absent. Together, these findings suggest that the AD pathology-induced reduction in CaV2.1 channel density and CaV2.1-GABAB1 de-clustering may play a role in the synaptic transmission alterations shown in the AD hippocampus. Therefore, uncovering these layer-dependent changes in P/Q calcium currents associated with AD pathology can benefit the development of future strategies for AD management.}, author = {Martín‐Belmonte, Alejandro and Aguado, Carolina and Alfaro‐Ruiz, Rocío and Kulik, Akos and de la Ossa, Luis and Moreno‐Martínez, Ana Esther and Alberquilla, Samuel and García‐Carracedo, Lucía and Fernández, Miriam and Fajardo‐Serrano, Ana and Aso, Ester and Shigemoto, Ryuichi and Martín, Eduardo D. and Fukazawa, Yugo and Ciruela, Francisco and Luján, Rafael}, issn = {1750-3639}, journal = {Brain Pathology}, number = {2}, publisher = {Wiley}, title = {{Nanoarchitecture of CaV>2.1 channels and GABAB receptors in the mouse hippocampus: Impact of APP/PS1 pathology}}, doi = {10.1111/bpa.13279}, volume = {35}, year = {2025}, } @inproceedings{20292, abstract = {In automated decision-making, it is desirable that outputs of decision-makers be robust to slight perturbations in their inputs, a property that may be called input-output robustness. Input-output robustness appears in various different forms in the literature, such as robustness of AI models to adversarial or semantic perturbations and individual fairness of AI models that make decisions about humans. We propose runtime monitoring of input-output robustness of deployed, black-box AI models, where the goal is to design monitors that would observe one long execution sequence of the model, and would raise an alarm whenever it is detected that two similar inputs from the past led to dissimilar outputs. This way, monitoring will complement existing offline ''robustification'' approaches to increase the trustworthiness of AI decision-makers. We show that the monitoring problem can be cast as the fixed-radius nearest neighbor (FRNN) search problem, which, despite being well-studied, lacks suitable online solutions. We present our tool Clemont, which offers a number of lightweight monitors, some of which use upgraded online variants of existing FRNN algorithms, and one uses a novel algorithm based on binary decision diagrams--a data-structure commonly used in software and hardware verification. We have also developed an efficient parallelization technique that can substantially cut down the computation time of monitors for which the distance between input-output pairs is measured using the L∞norm. Using standard benchmarks from the literature of adversarial and semantic robustness and individual fairness, we perform a comparative study of different monitors in Clemont, and demonstrate their effectiveness in correctly detecting robustness violations at runtime.}, author = {Gupta, Ashutosh and Henzinger, Thomas A and Kueffner, Konstantin and Mallik, Kaushik and Pape, David}, booktitle = {Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining}, isbn = {9798400714542}, issn = {2154-817X}, location = {Toronto, Canada}, pages = {790--801}, publisher = {Association for Computing Machinery}, title = {{Monitoring robustness and individual fairness}}, doi = {10.1145/3711896.3737054}, volume = {2}, year = {2025}, } @inproceedings{21090, abstract = {Fairness in AI is traditionally studied as a static property evaluated once, over a fixed dataset. However, real-world AI systems operate sequentially, with outcomes and environments evolving over time. This paper proposes a framework for analysing fairness as a runtime property. Using a minimal yet expressive model based on sequences of coin tosses with possibly evolving biases, we study the problems of monitoring and enforcing fairness expressed in either toss outcomes or coin biases. Since there is no one-size-fits-all solution for either problem, we provide a summary of monitoring and enforcement strategies, parametrised by environment dynamics, prediction horizon, and confidence thresholds. For both problems, we present general results under simple or minimal assumptions. We survey existing solutions for the monitoring problem for Markovian and additive dynamics, and existing solutions for the enforcement problem in static settings with known dynamics.}, author = {Cano Cordoba, Filip and Henzinger, Thomas A and Kueffner, Konstantin}, booktitle = {25th International Conference on Runtime Verification}, issn = {1611-3349}, location = {Graz, Austria}, pages = {1--21}, publisher = {Springer Nature}, title = {{Algorithmic fairness: A runtime perspective}}, doi = {10.1007/978-3-032-05435-7_1}, volume = {16087}, year = {2025}, } @article{19627, abstract = {Differentially private gradient descent (DP-GD) is a popular algorithm to train deep learning models with provable guarantees on the privacy of the training data. In the last decade, the problem of understanding its performance cost with respect to standard GD has received remarkable attention from the research community, which formally derived upper bounds on the excess population risk RP in different learning settings. However, existing bounds typically degrade with over-parameterization, i.e., as the number of parameters p gets larger than the number of training samples n -- a regime which is ubiquitous in current deep-learning practice. As a result, the lack of theoretical insights leaves practitioners without clear guidance, leading some to reduce the effective number of trainable parameters to improve performance, while others use larger models to achieve better results through scale. In this work, we show that in the popular random features model with quadratic loss, for any sufficiently large p , privacy can be obtained for free, i.e., |RP|=o(1) , not only when the privacy parameter ε has constant order, but also in the strongly private setting ε=o(1) . This challenges the common wisdom that over-parameterization inherently hinders performance in private learning.}, author = {Bombari, Simone and Mondelli, Marco}, issn = {1091-6490}, journal = {Proceedings of the National Academy of Sciences}, number = {15}, publisher = {National Academy of Sciences}, title = {{Privacy for free in the overparameterized regime}}, doi = {10.1073/pnas.2423072122}, volume = {122}, year = {2025}, } @inproceedings{21324, abstract = {Learning models have been shown to rely on spurious correlations between non-predictive features and the associated labels in the training data, with negative implications on robustness, bias and fairness. In this work, we provide a statistical characterization of this phenomenon for high-dimensional regression, when the data contains a predictive core feature x and a spurious feature y. Specifically, we quantify the amount of spurious correlations C learned via linear regression, in terms of the data covariance and the strength λ of the ridge regularization. As a consequence, we first capture the simplicity of y through the spectrum of its covariance, and its correlation with x through the Schur complement of the full data covariance. Next, we prove a trade-off between C and the in-distribution test loss L, by showing that the value of λ that minimizes L lies in an interval where C is increasing. Finally, we investigate the effects of over-parameterization via the random features model, by showing its equivalence to regularized linear regression. Our theoretical results are supported by numerical experiments on Gaussian, Color-MNIST, and CIFAR-10 datasets.}, author = {Bombari, Simone and Mondelli, Marco}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, issn = {2640-3498}, location = {Vancouver, Canada}, pages = {4839--4873}, publisher = {ML Research Press}, title = {{Spurious correlations in high dimensional regression: The roles of regularization, simplicity bias and over-parameterization}}, volume = {267}, year = {2025}, } @article{20370, abstract = {The Huntingtin protein (HTT), named for its role in Huntington’s disease, has been best understood as a scaffolding protein that promotes vesicle transport by molecular motors along microtubules. Here, we show that HTT also interacts with the actin cytoskeleton, and its loss of function disturbs the morphology and function of the axonal growth cone. We demonstrate that HTT organizes F-actin into bundles. Cryo–electron tomography (cryo-ET) and subtomogram averaging (STA) structural analyses reveal that HTT’s N-terminal HEAT and Bridge domains wrap around F-actin, while the C-terminal HEAT domain is displaced; furthermore, HTT dimerizes via the N-HEAT domain to bridge parallel actin filaments separated by ~20 nanometers. Our study provides the structural basis for understanding how HTT interacts with and organizes the actin cytoskeleton.}, author = {Carpentier, Rémi and Kim, Jaesung and Capizzi, Mariacristina and Kim, Hyeongju and Fäßler, Florian and Hansen, Jesse and Kim, Min Jeong and Denarier, Eric and Blot, Béatrice and Degennaro, Marine and Labou, Sophia and Arnal, Isabelle and Marcaida, Maria J. and Peraro, Matteo Dal and Kim, Doory and Schur, Florian KM and Song, Ji-Joon and Humbert, Sandrine}, issn = {2375-2548}, journal = {Science Advances}, number = {38}, publisher = {AAAS}, title = {{Structure of the Huntingtin F-actin complex reveals its role in cytoskeleton organization}}, doi = {10.1126/sciadv.adw4124}, volume = {11}, year = {2025}, } @article{20324, abstract = {We report relaxation oscillations in a one-dimensional array of Josephson junctions, wherein the array dynamically switches between low-current and high-current states. The oscillations are current-voltage dual to those ordinarily observed in single junctions. The current-voltage dual circuit quantitatively accounts for temporal dynamics of the array, including the dependence on biasing conditions. Injection locking of the oscillations results in well-developed current plateaux. A thermal model explains the self-consistent reduction of the superconducting gap due to overheating of the array in the high-current state. Our work suggests that overheating determines the switching from the high-current state to the low-current state.}, author = {Mukhopadhyay, Soham and Lancheros Naranjo, Diego A and Senior, Jorden L and Higginbotham, Andrew P}, issn = {2331-7019}, journal = {Physical Review Applied}, publisher = {American Physical Society}, title = {{Dual relaxation oscillations in a Josephson-junction array}}, doi = {10.1103/qvls-7s3q}, volume = {24}, year = {2025}, } @article{19370, abstract = {Sex-linked and autosomal loci experience different selective pressures and evolutionary dynamics. X (or Z) chromosomes are often hemizygous in males (or females), as Y (or W) chromosomes often degenerate. Such hemizygous regions can be under greater efficacy of selection, as recessive mutations are immediately exposed to selection in the heterogametic sex leading to faster adaptation and faster divergence on the X chromosome (the so-called Faster-X or Faster-Z effect). However, in young nonrecombining regions, Y/W chromosomes often have many functional genes, and many X/Z-linked loci are therefore diploid. The sheltering of recessive mutations on the X/Z by the Y/W homolog is expected to drive slower adaptation for diploid X/Z loci, i.e. a reduction in the efficacy of selection. While the Faster-X effect has been studied extensively, much less is known empirically about the evolutionary dynamics of diploid X or Z chromosomes. Here, we took advantage of published population genomic data in the female-heterogametic human parasite Schistosoma japonicum to characterize the gene content and diversity levels of the diploid and hemizygous regions of the Z chromosome. We used different metrics of selective pressures acting on genes to test for differences in the efficacy of selection in hemizygous and diploid Z regions, relative to autosomes. We found consistent patterns suggesting reduced Ne, and reduced efficacy of purifying selection, on both hemizygous and diploid Z regions. Moreover, relaxed selection was particularly pronounced for female-biased genes on the diploid Z, as predicted by recent theoretical work.}, author = {Mrnjavac, Andrea and Vicoso, Beatriz}, issn = {1759-6653}, journal = {Genome Biology and Evolution}, number = {2}, publisher = {Oxford University Press}, title = {{Reduced efficacy of selection on a young Z chromosome region of schistosoma japonicum}}, doi = {10.1093/gbe/evaf021}, volume = {17}, year = {2025}, } @article{18866, abstract = {Using JWST Near Infrared Camera (NIRCam) images of the globular cluster 47,Tucanæ, (or NGC,104), taken at two epochs just 7 months apart, we derived proper-motion membership down to F322W2 ∼ 27. We identified an intriguing feature at the very low-mass end of the main sequence, around ∼ 0.08,M_⋅, at magnitudes F322W2 ∼ 24 and m_ F150W2 ∼ 25. This feature, dubbed 'kink', is characterized by a prominent discontinuity in the slope of the main sequence. A similar discontinuity is seen in theoretical isochrones with oxygen-poor chemistries, related to the rapid onset of absorption. We therefore hypothesize that the cluster hosts disproportionately more oxygen-poor stars near the bottom of the main sequence compared to the upper main sequence and the red giant branch. Our results show no strong or conclusive evidence of a rise in the brown dwarf luminosity function at faint magnitudes, in contrast to previous findings likely affected by faint red background galaxies. In our analysis, we accounted for this contamination by using proper motion membership.}, author = {Scalco, M. and Gerasimov, R. and Bedin, L. R. and Vesperini, E. and Correnti, M. and Nardiello, D. and Burgasser, A. and Richer, H. and Caiazzo, Ilaria and Heyl, J. and Libralato, M. and Anderson, J. and Griggio, M.}, issn = {1432-0746}, journal = {Astronomy & Astrophysics}, publisher = {EDP Sciences}, title = {{JWST photometry and astrometry of 47 Tucanæ. Discontinuity in the stellar sequence at the star--brown dwarf transition}}, doi = {10.1051/0004-6361/202452907}, volume = {694}, year = {2025}, } @article{18936, abstract = {A major obstacle to predictive understanding of evolution stems from the complexity of biological systems, which prevents detailed characterization of key evolutionary properties. Here, we highlight some of the major sources of complexity that arise when relating molecular mechanisms to their evolutionary consequences and ask whether accounting for every mechanistic detail is important to accurately predict evolutionary outcomes. To do this, we developed a mechanistic model of a bacterial promoter regulated by 2 proteins, allowing us to connect any promoter genotype to 6 phenotypes that capture the dynamics of gene expression following an environmental switch. Accounting for the mechanisms that govern how this system works enabled us to provide an in-depth picture of how regulated bacterial promoters might evolve. More importantly, we used the model to explore which factors that contribute to the complexity of this system are essential for understanding its evolution, and which can be simplified without information loss. We found that several key evolutionary properties—the distribution of phenotypic and fitness effects of mutations, the evolutionary trajectories during selection for regulation—can be accurately captured without accounting for all, or even most, parameters of the system. Our findings point to the need for a mechanistic approach to studying evolution, as it enables tackling biological complexity and in doing so improves the ability to predict evolutionary outcomes.}, author = {Grah, Rok and Guet, Calin C and Tkačik, Gašper and Lagator, Mato}, issn = {1943-2631}, journal = {Genetics}, number = {2}, publisher = {Oxford University Press}, title = {{Linking molecular mechanisms to their evolutionary consequences: a primer}}, doi = {10.1093/genetics/iyae191}, volume = {229}, year = {2025}, } @article{18605, abstract = {The response of clouds and moist-convective processes to heat loss to space by long-wave radiative cooling is an important feedback in the Earth's atmosphere. It is known that moist convection increases roughly in equilibrium with radiative cooling, an assumption often made in simplified models of the tropical atmosphere. In this study, we use an idealised two-dimensional model of the atmosphere introduced by Vallis et. al. and incorporate a bulk-cooling term, which is an idealisation of radiative cooling in the atmosphere. We comment briefly on the static stability of the system to dry and moist convection and characteris its moist convective response to changes in the bulk cooling. We find that, while the clear-sky regions of the model respond directly to the change in the cooling term, the regions dominated by moist convective plumes are insensitive to changes in cooling. Similar to previous findings from cloud-resolving models, we too find in our idealised setting that the majority of the increase in convection occurs via an increase in the areal coverage of convection, rather than its intensity. We argue that these small-scale convective processes are an upper bound on how quickly convective intensity can change to stay in equilibrium with radiative cooling.}, author = {Agasthya, Lokahith N and Muller, Caroline J and Cheve, Mathis}, issn = {1477-870X}, journal = {Quarterly Journal of the Royal Meteorological Society}, number = {766}, publisher = {Wiley}, title = {{Moist convective scaling: Insights from an idealised model}}, doi = {10.1002/qj.4902}, volume = {151}, year = {2025}, }