Browse Articles
Discover research articles across all indexed journals
Daily briefing: Three decades of Dolly
Deep neural networks as discrete dynamical systems: Implications for physics-informed learning
We revisit the analogy between feed-forward deep neural networks (DNNs) and discrete dynamical systems derived from neural integral equations and their corresponding partial differential equation forms. A comparative analysis between the numerical/exact solutions of the Burgers’ and Eikonal equations and those obtained via physics-informed neural networks (PINNs) is presented. We show that PINN learning provides a different computational pathway compared to standard numerical discretization in approximating essentially the same underlying dynamics of the system. Within this framework, DNNs can be interpreted as discrete dynamical systems whose layerwise evolution approaches attractors, and multiple parameter configurations may yield comparable solutions, reflecting the degeneracy of the inverse mapping. In contrast to the structured operators associated with finite-difference procedures, PINNs learn dense parameter representations that are not directly associated with classical discretization stencils. This distributed representation generally involves a larger number of parameters, leading to reduced interpretability and increased computational cost. However, the additional flexibility of such representations may offer advantages in high-dimensional settings where classical grid-based methods become impractical.
Genomic and structural evidence of SARS-CoV-2 and MERS-CoV in migratory birds
Migratory birds are the natural reservoir of influenza A virus (IAV), but their role as a carrier of SARS-CoV-2 remains unclear. Here, we report the identification of three almost full-length viral genome sequences of SARS-CoV-2 variants of concern (VOCs) in Tundra swans. These sequences are named hCoV-19/Tundra swan/Jiangxi/IMCAS_M1/2021 (IMCAS_M1), hCoV-19/Tundra swan/Jiangxi /IMCAS_M2/2021 (IMCAS_M2), and hCoV-19/Tundra swan/Jiangxi/IMCAS_M3/2021 (IMCAS_M3). IMCAS_M1 and IMCAS_M3 have the same mutations as the Beta VOC (K417N, E484K, and N501Y) in the receptor-binding domain (RBD) of the viral spike (S) protein, whereas IMCAS_M2 shares the same mutations as the Gamma VOC (K417T, E484K, and N501Y) in the RBD with all three showing their distinct mutations in the genomes. Virus receptor angiotensin-converting enzyme 2 (ACE2) proteins from both Tundra swan (tsACE2) and Black swan (bsACE2) can bind to the RBDs of all three viruses and the Alpha VOC, but not to RBD of the prototype (PT) virus. The polar contacts and hydrophobic interactions revealed by cryo-electron microscopy (cryo-EM) structures of the RBD–ACE2 complex, play key roles in virus–receptor engagement. Furthermore, HeLa cells expressing bsACE2 and tsACE2 proteins could be transduced by pseudotyped SARS-CoV-2 variants (Alpha, Beta, and Gamma) but not PT SARS-CoV-2. In addition, we obtained one partial genome of MERS-CoV named Bar-headed goose/Tibet/IMCAS_M4/2022 (IMCAS_M4) with 20,180 bp (~70.0% coverage). Our findings highlight the importance of migratory birds as potential carrier of both SARS-CoV-2 and MERS-CoV, thereby posing potential threat to public health.
Local Ecological Knowledge offers opportunities to study and monitor elusive carnivores in coexistence landscapes
Vesicle size and membrane composition control monomer transfer pathways in multicomponent lipid vesicles
Lipid exchange between populations of non-fusing protocells, mediated by monomer diffusion through the aqueous phase, is a key process underlying protocell growth and compositional equilibration. While most theoretical descriptions consider single-component membranes with constant desorption rates, the influence of vesicle size and lipid composition on lipid-transfer dynamics remains poorly understood. Here, we develop a kinetic model for lipid exchange in multicomponent protocell populations that incorporates lipid species with different desorption rates and composition-dependent membrane packing effects. We show that lipid composition heterogeneity introduces multiple equilibration timescales and can generate non-monotonic dynamics, including transient overshoots of intermediate lipid species. Vesicle-size asymmetry controls the direction and rate of lipid transfer, determining the donor or acceptor character of individual vesicle populations. Membrane compression further modifies lipid-transfer kinetics and vesicle-area evolution. These results demonstrate how vesicle size, lipid composition, and membrane packing jointly govern lipid-transfer dynamics in heterogeneous protocell populations.
Evaluation of a proposed link between the SARS-CoV-2 furin cleavage site and mouse-adapted MERS-coronavirus MA30
The origin of the polybasic furin cleavage site (FCS) of SARS-CoV-2 remains a central question in debates on the emergence of COVID-19. One hypothesis proposes a genetic relationship between the SARS-CoV-2 S1/S2 motif RRAR and the RRVR sequence found in the mouse-adapted MERS-CoV strain MERS-MA30. Here, we combined large-scale bioinformatic analysis with experimental virology to evaluate this scenario. Analysis of over 17 million SARS-CoV-2 genomes revealed that the S:684V substitution corresponding to RRVR occurred repeatedly but only sporadically, never became phylogenetically basal, and showed limited geographic and temporal spread. Using reverse genetics, we generated SARS-CoV-2 variants encoding RRVR and demonstrated that S:684V consistently reduced viral entry efficiency and competitive fitness in multiple cell systems, including human respiratory epithelial cultures. RRVR variants did not evolve toward RRAR but instead accumulated alternative substitutions. These findings do not support an evolutionary relationship between MERS-MA30 and the SARS-CoV-2 FCS.
Tagraxofusp, Azacitidine, and Venetoclax in Blastic Plasmacytoid Dendritic Cell Neoplasm
Blastic plasmacytoid dendritic cell neoplasm (BPDCN) is an orphan, aggressive hematologic malignancy characterized by high CD123 expression. Tagraxofusp (TAG), a CD123-directed toxin, is the only approved therapy. Prior studies demonstrated BPDCN dependence on BCL2, sensitivity to venetoclax (VEN), and reversal of TAG resistance with azacitidine (AZA). We conducted a phase 2 study evaluating combination therapy with TAG, AZA, and VEN in patients with BPDCN (NCT03113643). Patients with previously untreated (1L) or relapsed/refractory (R/R) BPDCN received 28-day cycles of AZA (75 mg/m² days 1-7), VEN (400 mg days 1-21), and TAG (12 µg/kg days 4-6). Eligibility followed TAG guidelines to mitigate risk of capillary leak syndrome (CLS). Twenty-seven patients were enrolled (16 1L, 11 R/R), with median age of 70 years (range 21-81). Composite complete remission (CR/CRi/CRc) rates were 88% in 1L and 64% in R/R cohorts. Median duration of response was not reached in 1L and was 7.2 months in R/R patients. CLS occurred in 15% of patients; most were grade 2. In the 1L cohort, median overall and progression-free survival were not reached; the 2-year overall survival was 65% and 2-year progression-free survival was 53%. Median overall survival in R/R patients was 8.4 months. A high proportion of patients proceeded to allogeneic stem cell transplantation in remission (63% 1L, including 10 of 11 patients age 75 or younger; and 55% R/R). TAG-AZA-VEN is highly active in both untreated and relapsed BPDCN with a predictable and manageable safety profile, supporting its use as a new therapeutic option.
Cryptanalysis and improvement of a distributed zero trust scheme for airborne wireless sensor networks
<i>In situ</i> polymer nanoparticle densitometry via real-time 3D single-particle tracking
Single-particle techniques have the potential to measure the heterogeneous dynamics at the nanoscale within reaction mixtures. However, new tools are needed to gain insight into the molecular structure of evolving chemical systems at the single-particle level. Here, a single-particle method for measuring the density and mass of freely diffusing polymer nanoparticles is introduced and applied to the ring-opening metathesis polymerization of norbornene. Using 3D single-molecule active real-time tracking, growing polymer nanoparticles are tracked in situ, enabling simultaneous real-time measurement of both particle size and particle density. Particle density is extracted from the descent velocity of growing nanoparticles, giving insight into particle composition at the single-particle level. Using this technique, it is found that polynorbornene nanoparticles have a density of 850 ± 30 kg/m3, validated through control measurements on polystyrene beads. Density and size analysis indicate each nanoparticle contains ∼3.5 × 109 densely packed monomers, rather than loosely packed polymer chains. Simulations show that the accuracy of this method depends on particle size, the distance traveled before contact with the coverslip, solvent properties, and the number of trajectories. Overall, this paper presents a technique that enables in situ characterization of the density and mass of individual growing polymer nanoparticles.
Junctophilin-2-orchestrated calcium signalosome regulates brown adipocyte thermogenesis and energy metabolism
Dysfunctional adipocyte calcium handling is implicated in obesity and thermogenesis. Junctophilins (JPs) stabilize calcium microdomain junctions between the plasma membrane and endoplasmic reticulum, but whether JPs are required for adipocyte function is not known. We show that JP2 is enriched in thermogenic brown adipose tissue (BAT) relative to other fat depots and is downregulated under conditions of nutrient overload. Conditional knockdown of JP2 in adipocytes, and more selectively in BAT, exacerbates cold intolerance and susceptibility to diet induced obesity. Mechanistically, JP2-depleted brown adipocytes exhibit calcium handling dysfunction with elevated cytosolic calcium levels at baseline but diminished norepinephrine-induced calcium transients, reduced store-operated calcium entry. Basal cytosolic calcium overload accounts for an increase in calpain activation and ensuing downregulation of STIM1 and hormone-sensitive lipase in JP2-depleted cells. Furthermore, JP2 silencing in brown adipocytes reduced oxygen consumption rates and compromised mitochondrial structure and quality. Together, these findings demonstrate that JP2 is essential for normal calcium homeostasis in brown adipocytes and reveal a critical role for JP2 in thermogenesis and resistance to diet-induced metabolic dysregulation.
Single Cell Analysis of the Tumor Microenvironment Landscape Across the Disease Spectrum of Multiple Myeloma
Multiple myeloma progresses from precursor states to active disease, and studying tumor microenvironment (TME) evolution across these stages is key to understanding immune dysregulation and therapeutic resistance. Here, we integrated paired single-cell RNA, T-, and B-cell receptor sequencing data from bone marrow samples of 235 patients spanning the disease spectrum. This dataset revealed dynamic changes in the abundance and functional states of diverse immune populations, including T-, natural killer, B-, and myeloid cells. Using non-negative factorization of cell-subset composition, we identified five reproducible TME subtypes, or "ecotypes," defined by coordinated cellular architectures. These ecotypes captured structured variation beyond disease stage and were associated with distinct cell-cell communication networks, cytokine signaling landscapes, transcription factor programs, and shared gene modules, reflecting coordinated immune adaptation to microenvironmental constraints. By linking tumor features with immune changes and ecotype distributions, we identified context-dependent associations influenced by both disease biology and treatment effects on the TME. We found that an ecotype enriched for bone marrow-resident populations and limited immune infiltration was associated with tumor expansion and inferior clinical outcomes, whereas ecotypes reflecting T-cell functional states showed distinct associations with immunotherapy response and survival. Our study provides a comprehensive single-cell immune atlas of multiple myeloma and its precursors, offering insights into TME organization and progression, and may guide development of stage-specific or ecotype-targeted immunotherapies.
Effect of Brewster window placement on polarization, beam quality, and output power in a diode-side-pumped ring Nd:YAG/KTP green laser
Influence of guest molecules on shock-induced intrinsic reactivity and reaction pathways of CL-20 based on neuroevolution potential
Incorporating guest molecules into energetic crystal frameworks has proven to be an effective strategy for modulating microstructure and reaction mechanism. In this study, the regulatory mechanisms of guest molecules on the shock-induced intrinsic reactivity and reaction pathways of CL-20 were systematically investigated based on the neuroevolution potential. The results indicate that NCCH3 exhibits the strongest reaction-promoting effect, driving the system into a violent reaction stage rapidly. However, due to its low oxygen balance, both the detonation pressure and velocity are significantly reduced. In contrast, N2O and CO2 behave as weakly reactive, endothermic buffering guests at early stages, which delay the decomposition of CL-20 and enhance its detonation pressure. The temperature rise and detonation performance of CL-20/H2O2 are close to those of α-CL-20, while showing a certain inhibitory effect on the initial temperature increase. Reaction pathway analysis reveals that high pressure can activate additional pathways, including H abstraction, O migration, and NO2 abstraction, thereby promoting parallel reactions and radical capture. H2O and H2O2 promote radical reaction chains through endothermic processes and the release of OH/H radicals, with H2O2 exhibiting faster onset and higher intensity. CO2 and N2O primarily suppress decomposition by diluting active species and through reversible reorganization reactions, with N2O showing a more pronounced retardation effect. NCCH3 strongly promotes early stage decomposition by engaging in multiple reaction pathways, but the overall reaction proceeds incompletely at later stages. The polarity and size of guest molecules can modify the initial crystal structure, while their radical-generating capability and chemical reactivity determine how they participate in reactions, thereby jointly influencing the intrinsic reactivity and decomposition pathways of host-guest crystals.
Nucleotide diversity is a poor predictor of short-term adaptive potential
The capacity to adapt is essential for a population to avoid extinction in a changing world and is recognized as a global conservation priority. Adaptation requires additive (heritable) genetic variation for traits that influence survival and fecundity, but measuring this variation is difficult, particularly in species of conservation concern. Instead, molecular genetic diversity is often used to infer adaptive potential. However, previous research has cast doubt on the suitability of traditional molecular markers (allozymes and microsatellites) for this purpose given their weak relationship with heritability—a common measure of additive genetic variance. Advances in sequencing technology have since shifted focus toward nucleotide diversity and variation in functional regions, but their practicality for predicting adaptive potential remains debated. Furthermore, heritability itself is a poor proxy for adaptive potential because it depends on environmental variance. We collated 2,113 published estimates of evolvability—a measure of additive genetic variance that avoids environmental confounding—across 193 eukaryotic species, and evaluated how well evolvability is predicted by molecular diversity. We find that microsatellite and nucleotide diversity are not significantly correlated to each other, and neither predict evolvability. Nucleotide diversity explains 1.1% of interspecific differences in evolvability and doubling nucleotide diversity only corresponds to a 11.7% increase in evolvability. With theoretical work, we show that such weak associations are expected. Together, our results suggest that simple molecular measures of genetic variation are insufficient for predicting adaptive potential and reliance on these metrics risks misinforming conservation management.
Concurrent administration of BCMA and GPRC5D chimeric antigen receptor (CAR) T cells in advanced multiple myeloma
BCMA and GPRC5D-directed therapies have shown high efficacy in the treatment of relapsed or refractory myeloma with multiple new immune therapies now approved. Preclinical studies have shown that expression of BCMA and GPRC5D are heterogenous in malignant plasma cells. We conducted a phase I dose escalation trial of the BCMA CAR T cell therapy MCARH125 alone or in combination with the GPRC5D CAR T cell therapy MCARH109 in patients with heavily pre-treated relapsed or refractory multiple myeloma. Patients were treated across three dose levels and the primary objective of establishing safety of the concurrent infusion. The trial is registered in ClinTrials.gov, NCT05431608 and has now completed accrual. We treated 15 patients across 3 dose levels; this included 6 patients who received MCARH125 alone and 9 patients treated with concurrent infusion of MCARH125 and MCARH109. 1 patient each with MCARH 125 alone and the co-infusion of MCARH125 and MCARH109 developed immune effector cell associated hemophagocytic syndrome (IEC-HS). There were no instances of grade 3 or higher cytokine release syndrome (CRS) or immune effector cell associated neurologic syndrome (ICANS). The overall response for co-infusion was 78% with a median PFS of 18.2 months. Correlative analysis suggests that antigen loss maybe an important mechanism of resistance even with dual-antigen targeting and we show for the first time that expansion of one CAR population can limit expansion of the second population, as dual-CAR treated patients had significantly less BCMA-CAR expansion than BCMA-CAR alone treated patients, despite equivalent BCMA CAR T cell doses at infusion.
Predicting properties of antifungal drug molecules using neighborhood degree topological indices
Abstract Topological indices are an important part of chemical graph theory as a representation of the molecular structure and a predictor of the physicochemical properties. In this paper, nine neighborhood degree sum based topological indices are use as indicators of predictive potential in a QSPR analysis of twenty antifungal drug molecules. Linear, quadratic, and cubic regression models were used to test the relationships between the structural properties and important physicochemical properties, such as boiling point, density, enthalpy of vaporization, flash point, refractive index, molar refractivity, polarizability, surface tension, and molar volume. The findings also revealed that a cubic regression model provided the most optimal overall performance with the highest level of predictability in the form of molar refractivity and polarizability being the neighborhood inverse product index $$(ND_{4} )$$ ( $$R^{2}$$ = 0.98). These results assert that neighborhood based topological descriptors are effective, especially in modeling the refractivity related and electronic properties of drug molecules.
Spin–orbit-resolved strong-field ionization from real-time relativistic dynamics
Strong-field ionization exhibits rich ultrafast dynamics arising from intricate electron-field interactions. However, the role of spin–orbit coupling in the strong-field ionization of many-electron systems remains largely unexplored. Here, we develop a fully relativistic framework that combines a complex absorbing potential (CAP) with the exact two-component (X2C) real-time time-dependent density functional theory. We apply this approach to investigate the real-time strong-field dynamics of atomic krypton with a variational treatment of spin–orbit coupling. The CAP-X2C method captures the spin–orbit-resolved ionization-rate trends of the Kr 4p shell. Moreover, we uncover a previously inaccessible coherent population transfer channel between the 4p12 and 4p32 states, opened by spin–orbit coupling, which lifts the nonrelativistic symmetry restriction that forbids their mixing. This work provides a practical framework for exploring strong-field dynamics in systems where spin–orbit coupling plays a significant role.
Correction for Otsuki et al., Extracellular sulfatases support cartilage homeostasis by regulating BMP and FGF signaling pathways
Method for AFM investigation of lateral forces required for the surface detachment of neurites
Abstract Resprouting of neurites to establish direct connections between stimulation electrodes and spiral ganglion neurons is of considerable interest in current cochlear implant research. The adhesion strength of neurites to implant surfaces is a critical factor determining whether such connections develop, remain stable or are disrupted shortly after formation. In this study, an in vitro protocol was established to quantify the lateral forces required to detach neurites from the surface of precoated substrates using atomic force microscopy. Spiral ganglion neurons were isolated from mixed primary cell cultures using cytosine β-D-arabinofuranoside hydrochloride as a mitotic inhibitor. After several days of cultivation, the neurites were mechanically hooked with the cantilever tip and laterally dragged across the substrate surface to measure the forces required for detachment. Four different parameters and their influence on the detachment process were evaluated: exposure time to measurement conditions, the distance between growth cone and cantilever, the distance between cantilever and soma, and the total length of the neurite. The mean lateral detachment force was 2.47 ± 0.19 nN, and no significant correlation was observed between the lateral forces and any of the four investigated parameters. Overall, this study provides quantitative information on neurite-surface adhesion behaviour and strength in vitro.
Molecular structure, binding, and disorder in TDBC–Ag plexcitonic assemblies
Plexcitonic assemblies are hybrid materials composed of a plasmonic nanoparticle and molecular or semiconducting emitters whose electronic transitions are strongly coupled to the plasmonic mode. This coupling hybridizes the system modes into upper and lower polariton branches. The interaction strength depends on the number of emitters and on their orientation and spatial arrangement relative to the metallic surface. These structural factors have profound consequences for the ensuing photoexcited dynamics. Despite the extensive spectroscopic work on plexcitonic systems, direct understanding of the molecular geometry at the metal interface remains limited. We present a comprehensive structural characterization of a model plexciton formed by the cyanine dye 5,5′,6,6′-tetrachloro-1,1′-diethyl-3,3′-di(4-sulfobutyl)-benzimidazolocarbocyanine (TDBC) and silver nanodisks using NMR, THz-Raman spectroscopy, and density functional theory calculations. By comparing the signals from the monomeric and aggregated forms of TDBC with those of the plexciton, we identify shared spectral fingerprints that reveal how molecular packing is modified when the aggregate adsorbs on the silver surface. We observe Raman modes specific to plexciton systems and identify NOESY cross-peaks in the aliphatic region that, along with several Raman modes, are sensitive indicators of aggregation geometry and adsorption. We find that TDBC monomers adopt an asymmetric conformation in which both sulfobutyl chains lie on the same side of the chromophore, while J-aggregates adopt a symmetric up–down alternation of the chains from molecule to molecule, which becomes distorted and loses long range periodicity when adsorbed on Ag nanodisks. This work constrains the molecular geometry and interfacial arrangement of a prototypical TDBC–silver plexciton, providing a structural benchmark for understanding geometry-dependent photophysics in exciton–plasmon systems.