Browse Articles
Discover research articles across all indexed journals
A synergistic enhancement of the Ivy algorithm for GAN-based imbalanced classification
Abstract The Ivy Algorithm (IVYA), a swarm intelligence algorithm inspired by plant growth, presents a novel framework for optimization. To unlock its full potential in complex, high-dimensional problems, it is crucial to address the fundamental challenge of balancing exploration and exploitation, which can impact overall search efficiency and solution quality. To this end, this paper proposes an Enhanced Ivy Algorithm (E-IVYA) that integrates three synergistic mechanisms. First, a dynamic perturbation framework combining symmetric and asymmetric exploration is introduced to maintain population diversity. Second, a dynamic escape mechanism based on elite differential mutation is employed to prevent search stagnation and effectively escape from local optima. Third, an adaptive movement strategy inspired by the Sine-Cosine Algorithm is integrated to achieve a more adaptive balance between global exploration and local exploitation. The performance of the proposed E-IVYA was rigorously evaluated through two distinct phases. Initially, its optimization capabilities were benchmarked against a wide range of classic and advanced algorithms on the challenging IEEE CEC 2014 and 2017 test suites. Subsequently, its practical utility was validated by applying it to the complex task of automating the hyperparameter optimization of Generative Adversarial Networks (GANs) for imbalanced data classification. The experimental results demonstrate E-IVYA’s superior performance. On the standard benchmarks, E-IVYA consistently ranked as a top-performing algorithm. In the practical application, the E-IVYA-optimized GAN model achieved a minority class F1-Score of 0.87 on the highly imbalanced Credit-Card Fraud dataset, significantly outperforming models augmented with standard techniques like SMOTE (0.71). These findings confirm that E-IVYA is a robust and efficient tool for tackling complex optimization problems, particularly in the domain of automated machine learning.
Advanced biomimetic robotic hand with EMG lifelong learning and recognition
Marine fishery resource dynamic prediction based on CNN-XGBoost fusion model
Evidential deep learning for interatomic potentials
Simple electronegativity-based model for predicting formation of stable compounds across the periodic table
Insights into the self-assembly and interaction of sars-cov-2 fusion peptides with biomimetic plasma membranes
Abstract First identified in late 2019, the COVID-19 pandemic, caused by the SARS-CoV-2 coronavirus, rapidly escalated into a global health crisis. SARS-CoV-2 is a single-stranded RNA virus encased in a lipid envelope that houses key structural proteins, including the Spike glycoprotein, which mediates viral entry into host cells. Within Spike, the S2 subunit, and particularly its fusion domain, plays a critical role in merging viral and host membranes. To explore how receptor-driven Spike clustering influences this process, we investigated the self-assembly of S2 fusion peptides (FPs) and their interactions with biomimetic plasma membrane (PM) models composed of phospholipids, sphingomyelin, and cholesterol. Atomic force microscopy, laser direct infrared spectroscopy, neutron reflectometry, and grazing-incidence X-ray diffraction reveal that FPs form supramolecular assemblies that exclude cholesterol-rich nanodomains, increase membrane fluidity, and disrupt raft-like order associated with ACE2 binding. The appearance of spiral FP fibers supports a loaded-spring mechanism for membrane remodeling, offering a model for cooperative peptide-driven fusion, highlighting opportunities for antiviral and nanobiotechnological applications.
Cryo-EM structure of TMEM164 reveals distinct phospholipid remodeling mechanisms with anti-ferroptotic potential
Lithium-ion battery recycling through an integrated electro-membrane crystallization technology
Generalizable and scalable protein stability prediction with rewired protein generative models
Trait-like visual cortical hyperactivity in trait anxiety
Abstract Sensory processing varies across individuals, with some traits—particularly sensory hypersensitivity to basic non-valenced stimuli—linked to emotional traits and psychiatric risk. Traditional accounts attribute this sensory–emotion linkage to limbic or prefrontal modulation, but empirical support is limited. Growing evidence suggests sensory cortex itself flexibly encodes value beyond labeled-line analysis. Across four high-density EEG experiments with multi-wave assessments, we identified reliable visual cortical hyperactivity in high trait anxiety. The effect emerged as early as 46 ms, localized to V1/V2, and was specific to the parvocellular pathway. It was reproducible across arousal states, stimulus valence, extended intervals, and paradigms, and evident for both simple (grating) and complex real-world images. Importantly, cortical excitation–inhibition balance (EEG aperiodic exponent/1/f slope) predicted parvocellular responses in low- but not high-anxiety individuals, implicating disrupted E/I modulation. Thus, trait anxiety alters early visual processing, aligning cortical computations with an individual’s intrinsic biological propensity from the outset.
Mapping cis- and trans-regulatory target genes of human-specific deletions
Abstract Deletion of functional sequence is predicted to represent a fundamental mechanism of molecular evolution. Comparative genetic studies of primates have identified thousands of human-specific deletions (hDels), and the cis -regulatory potential of short (≤31 base pairs) hDels has been assessed using reporter assays. However, how structural variant-sized (≥50 base pairs) hDels influence molecular and cellular processes in their native genomic contexts remains unexplored. Here, we design genome-scale libraries of single-guide RNAs targeting 7.2 megabases of sequence in 6358 hDels and present a systematic CRISPR interference (CRISPRi) screening approach to identify hDels that modify cellular proliferation in chimpanzee pluripotent stem cells. By intersecting hDels with chromatin state features and performing single-cell CRISPRi (Perturb-seq) to identify their cis - and trans -regulatory target genes, we discovered 20 hDels controlling gene expression. We highlight two hDels, hDel_2247 and hDel_585, with tissue-specific activity in the brain. Our findings reveal a molecular and cellular role for sequences lost in the human lineage and establish a framework for functionally interrogating human-specific genetic variants.
Emergence of mammalian-adaptive PB2 mutations enhances polymerase activity and pathogenicity of cattle-derived H5N1 influenza A virus
Mapping risks of health conditions in people with atopic eczema in English primary care and hospital data
Abstract Atopic eczema may be associated with multiple health conditions. Here, we systematically explore risks across the full health spectrum based on the International Classification of Diseases, assessing associations between eczema and 2058 ICD-10 codes, 1593 phecodes, and 201 Global Burden of Disease codes. In English primary care electronic health records (1997 − 2023) we identify cohorts of people with eczema (up to 3 million) and matched (by age, sex, general practice) comparators without eczema (up to 14 million). In up to 25 years of follow-up, we capture outcomes recorded during hospital admissions. People with eczema show higher rates of several outcomes across multiple organ systems. Among those followed up from eczema diagnosis in childhood, atopic/allergic conditions and infections account for most excess diagnoses. Consistent across cohorts and analyses, large relative risk increases are seen for inflammatory bowel conditions (e.g., K50 Crohn disease, crude hazard ratio 1.70 [1.63-1.77]) and eye diseases (e.g., H16 Keratitis, crude hazard ratio 1.71 [1.57-1.86]). We provide a dashboard to explore and browse the full range of results.
Active learning-guided optimization of cell-free biosensors for lead testing in drinking water
Abstract Point-of-use diagnostics based on allosteric transcription factors (aTFs) are promising tools for environmental monitoring and human health. However, biosensors relying on natural aTFs rarely exhibit the sensitivity and selectivity needed for real-world applications, and traditional directed evolution struggles to optimize multiple biosensor properties at once. To overcome these challenges, we develop a multi-objective, machine learning (ML)-guided cell-free gene expression workflow for engineering aTF-based biosensors. Our approach rapidly generates high-quality sequence-to-function data, which we transform into an augmented paired dataset to train an ML model using directional labels that capture how aTF mutations alter performance. We apply our workflow to engineer the aTF PbrR as a point-of-use diagnostic for lead contamination in water. We tune the sensitivity of PbrR to sense at the U.S. Environmental Protection Agency (EPA) action level for lead and modify the selectivity away from zinc, a common metal found in water supplies. Finally, we show that the engineered PbrR functions in freeze-dried cell-free reactions, enabling a diagnostic capable of detecting lead in drinking water down to ~5.7 ppb. Our ML-driven, multi-objective framework powered by directional tokens can generalize to other biosensors and proteins, accelerating the development of synthetic biology tools for biotechnology applications.
Gaps in tropical science from unrepresentative distribution of sampling and citation across natural terrestrial environments
Abstract Effective environmental policies for the tropics depend on accurate, representative scientific data. However, there is strong evidence from particular disciplines and regions that existing research is patchily distributed. Here, we show that poor representation of sampling and citation in some biomes and across key environmental gradients from all disciplines for the entire tropics may lead to flawed scientific paradigms and inappropriate policy prescriptions. We map sampling locations and citations from 2 738 published studies in natural terrestrial tropical environments across all disciplines to identify gaps in field sampling effort and research attention. Five ecoregions – all in moist broadleaf forests – generate 22% of the total citations but cover only 3% of the tropical land area. By contrast, drier biomes with low tree cover account collectively for 57% of the tropical area but generate only 20% of total citations. Locations that are drier, colder, with greater plant species richness, lower tree cover and facing greater climate change extremes are under-sampled and under-cited. Our results will help to correct these imbalances to improve the scientific basis for environmental policies across the tropics.
Determinants of sensitivity to HER2-targeted antibody drug conjugates in urothelial cancer
Abstract HER2, encoded by the ERBB2 gene, is a receptor tyrosine kinase frequently activated in human cancers via gene amplification, mutation, and/or protein overexpression. In an analysis of 42,415 prospectively analyzed solid tumors, we show that 14.5% of urothelial cancers (n = 295/2,035) have oncogenic or likely oncogenic ERBB2 alterations (6.7% ERBB2 mutation, 6.3% amplification of wildtype ERBB2 , and 1.5% concurrent mutation and amplification). Discordance of ERBB2 mutational status between primary and metastatic disease sites is common in patients with urothelial cancer as is discordance of ERBB2 mutational status between patient-derived organoid/xenograft models and the tumors from which they were derived. In patient-derived urothelial cancer models, the HER2-targeted antibody-drug conjugate (ADC) trastuzumab deruxtecan is significantly more effective than the HER kinase inhibitor neratinib. In a real-world cohort of patients with urothelial cancer treated with trastuzumab deruxtecan, co-mutation and amplification of ERBB2 is associated with exceptional clinical response. Our data support expanded clinical trials of HER2-targeted ADCs for urothelial cancers with low HER2 expression, the clinical testing of HER2 ADCs with alternative cytotoxic payloads, and the development of functional precision oncology platforms capable of assessing payload sensitivity pre-treatment as a guide to individualized therapy selection.
The Ku80-p53-SIRT1 axis in DNA damage response contributes to sporadic and familial ALS and FTD
Genome-scale spatial mapping of the Hodgkin lymphoma microenvironment identifies tumor cell survival factors
Kinetochore mutations and histone phosphorylation pattern changes accompany holo- and macro-monocentromere evolution
Abstract Centromeres are essential for kinetochore assembly and spindle attachment. While chromosomes of most species are monocentric with a single centromere, a minority exhibit holocentricity, with a centromere along the chromatid length. Sporadic emergence of holocentricity suggests multiple independent transitions. To explore this, we compare the centromere and (epi)genome organization of two sister genera with contrasting centromere types: Chamaelirium luteum with large macro-monocentromeres and Chionographis japonica with holocentromeres. Both exhibit chromosome-wide histone phosphorylation patterns distinct from typical monocentric species. Kinetochore analysis reveals similar chimeric Borealin in both species, with additional KNL2 loss and NSL1 chimerism in Cha. luteum . The broad-scale synteny between both genomes supports de novo holocentromere formation in Chi. japonica . Despite sharing features with both centromere types, macro-monocentromeres do not represent a direct link between mono- and holocentromeres. We propose a model for the divergent evolution involving kinetochore gene mutations, altered histone phosphorylation patterns, and centromeric satellite DNA amplification.