4244 papers
This study integrates multiple transcriptomic datasets and machine learning algorithms to identify robust gene expression signatures that effectively distinguish cardiovascular disease samples from healthy controls, offering potential biomarkers for early diagnosis and improved understanding of disease mechanisms.
A comprehensive GPS tracking study of 2,346 red kites across Europe reveals that human activities, particularly poisoning and collisions, account for nearly 70% of adult mortality, underscoring the urgent need for coordinated international conservation efforts to mitigate these anthropogenic threats.
The study introduces InsectExpress, a multimodal deep learning framework that successfully predicts cross-species tissue-specific gene expression across diverse insects, revealing a layered regulatory architecture where conserved tissue-defining programs are superimposed on lineage-specific promoter syntax.
This study demonstrates that microtubule–kinesin active fluids exhibit fundamentally different responses to localized activity suppression depending on their flow state, where incoherent flows show only local effects while coherent flows display a striking system-wide, long-range coupling that alters flow speed, order, and response timescales.
This study demonstrates that in the seagrass *Cymodocea nodosa*, heat-induced DNA methylation changes establish a stable, vegetatively transmitted epigenetic memory that primes offspring for enhanced resilience against ocean warming.
The paper introduces SAMECAT, a structure-aware deep learning framework that robustly integrates metagenomic and clinical data to predict bone mineral density with superior generalizability and interpretability by revealing coherent functional themes and hub taxa through a novel stability-oriented attribution workflow.
This study reveals that in postmenopausal women, the relationship between endogenous estradiol and hippocampal functional connectivity during memory tasks differs by APOE4 status, where higher estradiol levels correlate with maladaptive brain patterns and increased Alzheimer's risk in APOE4 carriers but with beneficial effects in non-carriers.
This paper demonstrates that a biologically inspired spiking neural network achieves robust continual multi-goal navigation by employing a target-context evidence-gated plasticity mechanism that accumulates and selectively consolidates synaptic changes based on reward evidence, thereby effectively mitigating interference between competing goals.
This study demonstrates that elevated SQSTM1 expression induced by high glucose promotes retinal pigment epithelium cell apoptosis in diabetic retinopathy by activating the p53/PIDD1/caspase-8 signaling pathway, suggesting SQSTM1 as a potential therapeutic target.
The paper introduces MacaSurfer, a fully automated and containerized framework that enables robust, standardized surface-volume mapping of macaque MRI across the entire lifespan and diverse acquisition sites, validated on over 1,300 imaging sessions to establish normative morphometric trajectories for translational and comparative neuroscience.