994 papers
This paper introduces eccDNAMamba, a bidirectional state space model built on the Mamba-2 framework that overcomes the limitations of existing genomic foundation models by efficiently scaling to ultra-long sequences and preserving the intrinsic circular topology of extrachromosomal circular DNA (eccDNA) to achieve superior performance in cancer-related prediction tasks.
The paper introduces Astro-BEATS, an astronomy-inspired algorithm that outperforms existing methods in detecting and segmenting subtle miniature synaptic calcium transients in fluorescence microscopy, thereby facilitating the creation of training datasets for deep learning applications.
OmicClaw is an executable natural-language framework built on the unified OmicVerse ecosystem and J.A.R.V.I.S. runtime that transforms fragmented multi-omics analysis into reproducible, traceable workflows by converting user requests into validated actions over live data objects.
CROCHET is a versatile, open-source, end-to-end pipeline designed to democratize spatial omics by enabling the automated construction of large-scale, interactive spatially resolved cell atlases from complex single-cell data across diverse sample cohorts.
The paper introduces st-Xprop, a cross-propagative graph network that effectively integrates gene expression and histological image data through dual-graph embedding and alternating cross-modal propagation to identify more accurate and biologically meaningful spatial tissue domains in multi-modal spatial transcriptomics.
VarDCL is a novel multimodal framework that integrates protein language model embeddings with self-distilled contrastive learning to effectively capture sequence and structural differences caused by mutations, achieving state-of-the-art accuracy in distinguishing pathogenic missense variants from benign ones.
RIBEX is a multimodal framework that integrates protein language model embeddings with graph-derived positional encodings from the human interactome to significantly improve the prediction and interpretation of RNA-binding proteins, particularly those lacking canonical domains or enriched in intrinsically disordered regions.
This study develops a phenotype-structured mathematical model to demonstrate how intra-tumor proliferation heterogeneity and life-history trade-offs drive eco-evolutionary dynamics, revealing that while all treatments slow tumor growth, they induce distinct evolutionary trajectories by selectively enriching either fast- or slow-proliferating clones depending on the specific therapeutic targeting strategy.
The authors present Glydentify, an explainable deep learning platform that accurately predicts the donor substrate specificity of glycosyltransferases by integrating protein sequence and chemical features, thereby enabling the functional characterization of uncharacterized enzymes through both computational modeling and experimental validation.
This paper introduces 3D-Manhattan, a browser-based interactive tool that visualizes multiple GWAS results in a unified 3D coordinate system to facilitate the comparative analysis of genetic associations across time, traits, or experimental conditions.
BioOS is a novel computational runtime that simulates emergent plant development by executing gene regulatory networks within Formal Cell abstractions, achieving high accuracy in predicting *Arabidopsis thaliana* phenotypes across multiple biological benchmarks without relying on hardcoded behavioral rules.
This study presents an integrated artificial intelligence and quantum chemistry framework to rationally design and computationally validate "Solres," a novel antibacterial agent targeting key virulence proteins in *Ralstonia solanacearum* to combat antimicrobial resistance in agriculture.
LigandForge is a high-throughput discrete diffusion model that generates high-affinity peptide binders for diverse receptor targets in a single forward pass without iterative structure prediction, achieving a >10,000-fold speedup over existing methods while producing structurally diverse candidates with superior predicted binding quality.
RNAElectra is a novel RNA foundation model that leverages ELECTRA-style replaced-token detection pretraining on diverse non-coding RNAs to achieve superior cross-task generalization and interpretability in RNA regulatory inference compared to traditional masked language modeling approaches.
NYX is a format-aware, learned compression system built on the OpenZL framework that significantly improves both compression speed and ratios for various omics file types by exploiting their inherent data structures, outperforming existing format-specific compressors while maintaining lossless efficiency.
This paper introduces the first formal definition of contigs for variable-order de Bruijn graphs, termed "(, )-tigs," and presents an efficient algorithm to enumerate them, demonstrating that this approach significantly improves assembly contiguity on PacBio HiFi data while remaining computationally lighter than full genome assemblers.
Nerpa 2 is a freely available probabilistic framework that utilizes hidden Markov models to accurately and scalably link nonribosomal peptide biosynthetic gene clusters to their chemical structures, outperforming existing methods in both accuracy and pathway reconstruction.
The paper introduces ATOMICA, a geometric deep learning model trained on over two million complexes to generate universal, multiscale atomic representations of intermolecular interfaces across five molecular modalities, demonstrating superior performance in structure-function benchmarks and successfully predicting functional ligands for previously uncharacterized "dark" protein pockets.
This study presents AFESM, a comprehensive dataset of 820 million predicted protein structures derived from AlphaFold2 and ESMfold, which reveals millions of structural clusters, uncovers novel domain folds and combinations, and highlights the critical role of metagenomic data and prediction quality in expanding our understanding of the protein structure universe.
This study reveals that alternative intronic polyadenylation serves as a widespread mechanism for escaping nonsense-mediated mRNA decay (NMD) by removing downstream exon-exon junctions, thereby enabling tissue-specific regulation of gene expression in human genes such as VRK3, NFX1, TM2D3, and DDX31.