Browse Articles

Accurate POI recommendation for random groups with improved graph neural networks and a multi-negotiation model

Neuroglobin regulates autophagy through mTORC1/RAPTOR/ULK-1 pathway in human neuroblastoma cells

Functional mapping of the molluscan brain guided by synchrotron X-ray tomography

Molluscan brains are composed of morphologically consistent and functionally interrogable neurons, offering rich opportunities for understanding how neural circuits drive behavior. Nonetheless, detailed component-level CNS maps are often lacking, total neuron numbers are unknown, and organizational principles remain poorly defined, limiting a full and systematic characterization of circuit operation. Here, we establish an accessible, generalizable approach, harnessing synchrotron X-ray tomography, to rapidly determine the three-dimensional structure of the multimillimeter-scale CNS of Lymnaea . Focusing on the feeding ganglia, we generate a full neuron-level reconstruction, revealing key design principles and revising cell count estimates upward threefold. Our atlas uncovers the superficial but also nonsuperficial ganglionic architecture, reveals the cell organization in normally hidden regions—ganglionic “dark sides”—and details features of single-neuron morphology, together guiding targeted follow-up functional investigation based on intracellular recordings. Using this approach, we identify three pivotal neuron classes: a command-like food-signaling cell type, a feeding central pattern generator interneuron, and a unique behavior-specific motoneuron, together significantly advancing understanding of the function of this classical control circuit. Combining our morphological and electrophysiological data, we also establish a functional CNS atlas in Lymnaea as a shared and scalable resource for the research community. Our approach enables the rapid construction of cell atlases in large-scale nervous systems, with key relevance to functional circuit interrogation in a diverse range of model organisms.

Reuse of bottom sediment from reservoirs to cropland is a promising agroecological practice that must be rationalized

Abstract In semi-arid areas, intermittent streams are often equipped with small reservoirs to store water for irrigation and/or groundwater recharge, and to capture sediments lost through erosion. These reservoirs must be periodically desilted to maintain their storage capacity. While bottom sediments are generally considered waste, their reuse in agricultural fields is a centuries-old practice in India. Our study aimed to test the hypothesis that local farmers’ knowledge and current practices can help in understanding and rationalizing this practice. The study relied on both interviews of farmers and physico-chemical analysis of soil and sediment samples collected in a cultivated watershed in South India. First, our results disprove our hypothesis as we found a wide diversity of (i) application rates ranging from light soil amendment to creation of anthropogenic soils, and costs, which were not explained by the distance between reservoirs and fields neither by the field size, suggesting that there is no consensus among farmers on the optimal dose, and (ii) opinions on the impact of sediments on soil functions with the majority citing an improvement in the physical and/or chemical properties of the soil, suggesting that there is no consensus on the sediment impact on soil. Secondly, our results highlight that (i) only farmers with access to irrigation implemented this practice and they sourced sediment from the nearest reservoir, (ii) a slight majority of farmers used less irrigation water and less fertilizer after sediment application, and (iii) differences in sediments and soils composition suggest that sediment application is more likely to improve soil physical structure than nutrient status. The reuse of sediments on cropland could therefore be a promising agroecological practice, likely to increase the resource circularity and the sustainability of cropping systems. However, expressing its potential would require defining optimal application rates, assessing potential risks, sharing knowledge and promoting collective management of the resource.

Detection of obstetric anal sphincter injuries using machine learning-assisted impedance spectroscopy: a prospective, comparative, multicentre clinical study

Multiomics analysis unveils the cellular ecosystem with clinical relevance in aldosterone-producing adenomas with <i>KCNJ5</i> mutations

Aldosterone-producing adenomas (APA), a major endocrine tumor and leading subtype of primary aldosteronism, cause secondary hypertension with high cardiometabolic risks. Despite potentially producing multiple steroid hormones, detailed cellular mechanisms in APA remain insufficiently studied. Our multiomics analysis focusing on APA with KCNJ5 mutations, which represent the most common genetic form, revealed marked cellular heterogeneity. Tumor cell reprogramming initiated from stress-responsive cells to aldosterone-producing or cortisol-producing cells, with the latter progressing to proliferative stromal-like cells. These cell subtypes showed spatial segregation, and APA exhibited genomic intratumor heterogeneity. Among the nonparenchymal cells, lipid-associated macrophages, which were abundant in APA, might promote the progression of cortisol-producing and stromal-like cells, suggesting their role in the tumor microenvironment. Intratumor cortisol synthesis was correlated with increased blood cortisol levels, which were associated with the development of vertebral fractures, a hallmark of osteoporosis. This study unveils the complex cellular ecosystem with clinical relevance in APA with KCNJ5 mutations, providing insights into tumor biology that could inform future clinical approaches.

Fine-tuning diffusion model to generate new kite designs for the revitalization and innovation of intangible cultural heritage

Cryo-EM heterogeneity analysis using regularized covariance estimation and kernel regression

Proteins and the complexes they form are central to nearly all cellular processes. Their flexibility, expressed through a continuum of states, provides a window into their biological functions. Cryogenic electron microscopy (cryo-EM) is an ideal tool to study these dynamic states as it captures specimens in noncrystalline conditions and enables high-resolution reconstructions. However, analyzing the heterogeneous distributions of conformations from cryo-EM data is challenging. We present RECOVAR, a method for analyzing these distributions based on principal component analysis (PCA) computed using a REgularized COVARiance estimator. RECOVAR is fast, robust, interpretable, expressive, and competitive with state-of-the-art neural network methods on heterogeneous cryo-EM datasets. The regularized covariance method efficiently computes a large number of high-resolution principal components that can encode rich heterogeneous distributions of conformations and does so robustly thanks to an automatic regularization scheme. The reconstruction method based on adaptive kernel regression resolves conformational states to a higher resolution than all other tested methods on extensive independent benchmarks while remaining highly interpretable. Additionally, we exploit favorable properties of the PCA embedding to estimate the conformational density accurately. This density allows for better interpretability of the latent space by identifying stable states and low free-energy motions. Finally, we present a scheme to navigate the high-dimensional latent space by automatically identifying these low free-energy trajectories. We make the code freely available at https://github.com/ma-gilles/recovar .

MicroNAS for memory and latency constrained hardware aware neural architecture search in time series classification on microcontrollers

Abstract The use and research of neural networks on very small processor systems are currently still limited. One of the main reasons is that the design of microcontroller-architecture-aware ML models that take into account user-defined constraints on memory consumption and run-time are very difficult to implement. Therefore, we adapt the concept of differentiable neural architecture search (DNAS) to solve the time series classification problem on resource-constrained microcontrollers (MCUs). This paper explores and demonstrates for the first time that this problem can be solved using Neural Architecture Search (NAS). The key of our specific hardware-aware approach, MicroNAS, is an integration of a DNAS approach, Latency Lookup Tables, Dynamic Convolutions and a novel search space specifically designed for time series classification on MCUs. The resulting system is hardware-aware and can generate neural network architectures that satisfy user-defined limits on execution latency and peak memory consumption. To support our findings, we evaluate MicroNAS under different latency and peak memory constraints. The experiments highlight the ability of MicroNAS to find trade-offs between latency and classification performance across all dataset and microcontroller combinations. As an example, on the UCI-HAR dataset, MicroNAS achieves an accuracy of 94.62% when allowed 25 ms and 98.86% when allowed 50 ms when running on the Nucleo-L552ZE-Q. The much more powerful Arduino Portenta, on the other hand, achieves an accuracy of 95.88% with an allowance of 3 ms and 99.37% when allowed 25 ms displaying the ability of MicroNAS to adapt to different microcontrollers. MicroNAS is also able to find architectures which perform similarly to state-of-the-art systems designed to run on desktop computers (99.62% vs. 99.65% accuracy on the UCI-HAR dataset and 97.83% vs. 97.46% accuracy on the SkodaR dataset).

Directed evolution of a sequence-specific covalent protein tag for RNA labeling

Efficient methods for conjugating proteins to RNA are needed for RNA delivery, imaging, editing, interactome mapping, and barcoding applications. Noncovalent coupling strategies using viral RNA binding proteins such as MS2/MCP have been widely applied but are limited by tag size, sensitivity, and dissociation over time. We took inspiration from a sequence-specific, covalent protein–DNA conjugation method based on the Rep nickase of a porcine circovirus called “HUH tag”. Though wild-type HUH protein has no detectable activity toward an RNA probe, we engineered an RNA-reactive variant, called “rHUH”, through 7 generations of yeast display–based directed evolution. Our 13.4 kD rHUH has 12 mutations relative to HUH and forms a covalent tyrosine-phosphate ester linkage with a 10-nucleotide RNA recognition sequence (“rRS”) within minutes. We engineered the sensitivity down to 1 nM of target RNA, shifted the metal ion requirement from Mn 2+ toward Mg 2+ , and demonstrated efficient labeling in mammalian cell lysate. This work paves the way toward a potentially powerful methodology for sequence-specific covalent protein–RNA conjugation in biological systems.

Author Correction: Adaptive evolution and early diversification of photonic nanomaterials in marine diatoms

GenAI synthetic data create ethical challenges for scientists. Here’s how to address them.

Investigating the expression of anti/pro-inflammatory cytokines in the pathogenesis and treatment of ulcerative colitis and its association with serum level of vitamin D

Radiation-induced cellular plasticity primes glioblastoma for forskolin-mediated differentiation

Glioblastoma (GBM) is the deadliest brain cancer in adults, and all patients succumb to the tumor. While surgery followed by chemoradiotherapy delays disease progression, these treatments do not lead to tumor control, and targeted therapies or biologics have failed to further improve survival. Utilizing a transient radiation-induced state of multipotency, we used the adenylcyclase activator forskolin to alter the fate of irradiated glioma cells. The effects of the combined treatment on neuronal marker expression, cell cycle distribution, and proliferation were studied. Gene expression profiling was conducted using bulk RNA-seq. Changes in cell populations were investigated using single-cell RNA-seq. Effects on glioma stem cells (GSCs) were studied in extreme limiting dilution assays, and the effects on median survival were studied in both syngeneic and PDOX mouse models of GBM. The combined treatment induced the expression of neuronal markers in glioma cells, reduced proliferation, and led to a distinct gene expression profile. scRNA-seq revealed that the combined treatment forced glioma cells into a microglia- and neuron-like phenotype. In vivo, this treatment led to a loss of GSCs and prolonged median survival. Collectively, our data suggest that revisiting a differentiation therapy with forskolin in combination with radiation could lead to clinical benefit.

YOLO-BS: a traffic sign detection algorithm based on YOLOv8

Leaky ribosomal scanning enables tunable translation of bicistronic ORFs in green algae

Advances in sequencing technology have unveiled examples of nucleus-encoded polycistrons, once considered rare. Exclusively polycistronic transcripts are prevalent in green algae, although the mechanism by which multiple polypeptides are translated from a single transcript is unknown. Here, we used bioinformatic and in vivo mutational analyses to evaluate competing mechanistic models for translation of bicistronic mRNAs in green algae. High-confidence manually curated datasets of bicistronic loci from two divergent green algae, Chlamydomonas reinhardtii and Auxenochlorella protothecoides , revealed a preference for weak Kozak-like sequences for ORF 1 and an underrepresentation of potential initiation codons before the ORF 2 start codon, which are suitable conditions for leaky ribosome scanning to allow ORF 2 translation. We used mutational analysis in A. protothecoides to test the mechanism. In vivo manipulation of the ORF 1 Kozak-like sequence and start codon altered reporter expression at ORF 2, with a weaker Kozak-like sequence enhancing expression and a stronger one diminishing it. A synthetic bicistronic dual reporter demonstrated inversely adjustable activity of green fluorescent protein expressed from ORF 1 and luciferase from ORF 2, depending on the strength of the ORF 1 Kozak-like sequence. Our findings demonstrate that translation of multiple ORFs in green algal bicistronic transcripts is consistent with episodic leaky scanning of ORF 1 to allow translation at ORF 2. This work has implications for the potential functionality of upstream open reading frames (uORFs) found across eukaryotic genomes and for transgene expression in synthetic biology applications.

Evaluation method of interface bonding strength of cemented carbide with additives

A spectral machine learning approach to derive central aortic pressure waveforms from a brachial cuff

Analyzing cardiac pulse waveforms offers valuable insights into heart health and cardiovascular disease risk, although obtaining the more informative measurements from the central aorta remains challenging due to their invasive nature and limited noninvasive options. To address this, we employed a laboratory-developed cuff device for high-resolution pulse waveform acquisition and constructed a spectral machine learning model to nonlinearly map the brachial wave components to the aortic site. Simultaneous invasive aortic catheter and brachial cuff waveforms were acquired in 115 subjects to evaluate the clinical performance of the developed wave-based approach. Magnitude, shape, and pulse waveform analysis on the measured and reconstructed aortic waveforms were correlated on a beat-to-beat basis. The proposed cuff-based method reconstructed aortic waveform contours with high fidelity (mean normalized-RMS error = 11.3%). Furthermore, continuous signal reconstruction captured dynamic aortic systolic blood pressure (BP) oscillations (r = 0.76, P &lt; 0.05). Method-derived central pressures showed strong correlation with the independent invasive measurement for systolic BP (R 2 = 0.83; B [LOA] = −0.3 [−17.0, 16.4] mmHg) and diastolic BP (R 2 = 0.58; B [LOA] = −0.7 [−13.1, 11.6] mmHg). Shape-based features are effectively captured by the spectral machine learning method, showing strong correlations and no systemic bias for systolic pressure–time integral (r = 0.91, P &lt; 0.05), diastolic pressure–time integral (r = 0.95, P &lt; 0.05), and subendocardial viability ratio (r = 0.86, P &lt; 0.05). These results suggest that the nonlinear transformation of wave components from the distal to the central site predicts the morphological waveform changes resulting from complex wave propagation and reflection within the cardiovascular network. The proposed wave-based approach holds promise for future applications of noninvasive devices in clinical cardiology.

Porcine β-defensin 5 (pBD-5) modulates the inflammatory and metabolic host intestinal response to infection

Empowering genome-wide association studies via a visualizable test based on the regional association score

The genome-wide association studies identified genes associated with many diseases, but the identification and verification of disease variants are still challenging due to small effects and large number of individual variants. In this paper, we propose a powerful method that first quantifies the strength of regional associations at each single nucleotide polymorphism and converts these measures into time series data before using a change point detection algorithm to identify significant regions. In our extensive simulation study, the proposed method consistently demonstrates greater power than existing alternatives, achieving a relative increase of over 20% in challenging scenarios where true causal variants are sparse and multiple association regions exist at the same time, while maintaining a lower false positive rate.