1504 papers
This study presents a novel hybrid machine learning framework combining LightGBM and PyTorch-based Attention mechanisms to accurately predict VBNC bacterial resuscitation, achieving superior performance over traditional methods and identifying nutrient concentration and cAMP signaling as key determinants for recovery.
The paper introduces Tabula, a privacy-preserving single-cell foundation model that leverages federated learning to explicitly capture the tabular structure of genomic data, enabling the discovery of regulatory logic and the identification of rejuvenation factors without sharing raw data across institutions.
This paper presents the Pacific eDNA Coastal Observatory, a cross-national network spanning 16 nodes from San Diego to Juneau that utilized 1,041 seawater samples over four years to generate a comprehensive, validated dataset of nearshore fish biodiversity across the Northeast Pacific, demonstrating the effectiveness of coordinated eDNA monitoring for large-scale comparative inference.
By applying a novel AI-mediated framework (AHA) to reproduce and audit thousands of Alzheimer's disease studies on a common dataset, the authors demonstrate that while the field's reported predictive performance can be replicated at scale, the available data is fundamentally information-insufficient to distinguish between competing hypotheses, revealing a ceiling on the evidence that explains the field's divergent research trajectories.
This systematic review of 65 murine studies concludes that music exposure significantly enhances behavioral, neurobiological, and physiological outcomes in rats, though methodological heterogeneity necessitates further standardized research to clarify mechanisms and confirm translational potential for humans.
This study presents three high-resolution cryo-EM structures of the human SET8 catalytic domain bound to nucleosomes, revealing how specific arginine-mediated anchoring to the acidic patch and dynamic conformational transitions facilitate the recognition and monomethylation of histone H4 lysine 20.
The paper introduces PRIDE-AP, a new FAIR-compliant, open-data repository within the PRIDE infrastructure that addresses the lack of a dedicated archive for high-throughput affinity proteomics by enabling standardized submission, validation, and cross-technology integration of datasets to accelerate biomarker discovery.
This paper reframes the spread of artificial intelligence literacy within organizations as a contagion process, demonstrating through network epidemiology and machine learning that learning retention is primarily driven by structural communication network properties and an individual's position within them rather than just personal attributes.
This paper, which originally proposed a link between circadian clock-gene structure and PD-L1 across six cancer types to define melanoma subtypes, has been withdrawn by the author due to concerns regarding the accuracy, validity, and reliability of its content.
The study reveals that the conserved PRPF39/PRTF1 complex is essential for splicing difficult, short, and GC-poor introns across metazoans, with its loss causing distinct splicing defects (exon skipping in vertebrates and intron retention in invertebrates) and developmental abnormalities like wing duplication in flies.