deep learning

Predictors Recover Most of the Metagenomic Signal for Antibiotic Resistance Gene Occurrence: A Cross-City Test of Geographic Transferability in Urban Wastewater

This study uses environmental predictors to predict antibiotic resistance gene occurrence in urban wastewater, achieving 78% of the predictive performance of metagenomic sequencing.

OmiCoreTumorDetector: an open, molecularly validated model for mapping tumour regions in colorectal cancer H&E sections

OmiCoreTumorDetector maps tumour regions in colorectal cancer H&E sections with high accuracy, validated against transcriptomic data and showing strong performance on unseen slides.

Daily locomotor rhythms and ecdysial stage-dependent behaviors in the tardigrade Hypsibius exemplaris

This study identifies an endogenous circadian-like rhythm in the tardigrade Hypsibius exemplaris, independent of light, and links it to ecdysial stages.

A structure-informed deep learning framework for modeling TCR-peptide-HLA interactions

StriMap models TCR-peptide-HLA interactions using structural and sequence features, achieving state-of-the-art performance and identifying disease-associated peptides in ankylosing spondylitis.

A closed-loop language-model agent for target-specific multi-objective hit-to-lead optimization

This paper develops a language-model agent for hit-to-lead optimization that outperforms existing methods on four kinase targets in a computational benchmark.

The phage stress test: a proving ground for genome language models in biological prediction

This study evaluates a genome language model's ability to predict the effects of sequence changes in phages using experimental data, finding mixed success in distinguishing functional from nonfunctional mutations.

Vir2vec: A Genome-Wide Viral Embedding

Vir2vec is a genome-wide viral embedding model that outperforms existing models in viral genome understanding tasks and provides explainability for host adaptation mutations.

An open imaging and AI resource enabling unbiased quantification of extrachromosomal DNA at scale

The paper provides an open resource for automated ecDNA quantification, comparing methods and introducing ecCount to reduce count bias.

Efficient primer design for broad-range virus detection: A semi-automated workflow using sequence clustering and varVAMP

This paper presents a semi-automated workflow using sequence clustering and varVAMP to design broad-range RT-qPCR primers for diverse viruses, improving representation and reducing redundancy.

TopoFlow: Evolutionarily Conditioned Flow Matching for Protein Conformational Ensemble Generation

TopoFlow generates protein conformational ensembles using evolutionary and structural information, improving benchmark metrics and aligning with experimental data.

De novo design of small cyclic peptide inhibitors from non-canonical amino acids by cofolding-guided search

This paper introduces nCycle-Forge, a computational method for designing cyclic peptides that outperforms existing approaches in hit rate and potency against two targets.

ISP Platform powered by Geneformer: Framework for Cross-Species, State-feedback, and Multi-Gene In Silico Perturbation Screens with Application to iPS Cell State Transitions

MoTRUST: adaptive semantic protection for mosaic single-cell integration and molecular recovery

MoTRUST integrates mosaic single-cell multi-omics data and recovers missing molecular information, achieving top performance in integration and recovery tasks.

A Counterfactual Framework for Directional Cell-Cell Interaction Analysis in Spatial Transcriptomics

This paper introduces a counterfactual framework to analyze directional cell-cell interactions in spatial transcriptomics, identifying significant asymmetric influences between tumor, immune, and stromal compartments in cholangiocarcinoma tissue.

NeoToxPred: a fine-tuned protein language model with orthologous and length-stratified negative controls for robust toxicity classification

NeoToxPred improves toxicity classification by using a protein language model with carefully curated negative controls, outperforming existing methods on both internal and external benchmarks.

ASOFormer: a Transformer-based model for predicting antisense oligonucleotide efficacy to support therapeutic candidate prioritization

ASOFormer predicts ASO efficacy using sequence, structure, and modification features, outperforming existing methods in accuracy and candidate prioritization.

PepSpace: An Automated, Physics-Driven Directed Evolution Platform for the De Novo Design of Specific Peptide Binders and Condensate Modulators

PepSpace designs specific peptides using physics-driven directed evolution, achieving stronger predicted interactions and better binding specificity than machine learning models.

An RNA Language Model trained on sequence alone reveals the structural logic of Internal Ribosome Entry Sites

Continued pretraining of an RNA language model on picornaviral 5′ UTRs alone predicts IRES base pairs with median precision 0.80 versus 0.47 for Infernal, and reveals a Type II subclass.

Development of Rosaceae Crop-Specific Nanopore Models for Community Use Through the Genome Database for Rosaceae

The paper develops crop-specific Nanopore basecalling models for Rosaceae species to improve sequencing quality and enables community use of these models.

Self-supervised Internal Learning Enhances Isotropic Resolution for Three-dimensional Fluorescence Microscopy

This paper introduces DeepIso, a self-supervised framework that enhances isotropic resolution in 3D fluorescence microscopy by restoring axial frequency content without requiring explicit PSF models or system calibration.

ProFinder: Candidate constitutive promoter prediction across the bacterial tree for genetic tool development in non-model species

ProFinder predicts constitutive promoters across bacterial species to aid genetic tool development, achieving high precision and robustness across diverse genomes.

Protein Language Models: Learning From Evolution, Designing Beyond It

This paper argues for evaluating protein language models through prospective experimental success rather than retrospective predictive accuracy.

CycDelta: Multimodal Residue-Graph Learning of Permeability Changes in Cyclic Peptides

CycDelta is a multimodal residue-graph model for predicting permeability changes in cyclic peptides by comparing within-source permeability differences, outperforming baselines on out-of-distribution and zero-shot tasks.

DeepCDS: Ab initio coding sequence prediction in prokaryotic short nucleotide sequences

DeepCDS improves coding sequence prediction in short prokaryotic sequences using deep learning and protein language model embeddings, outperforming existing methods across various sequence lengths and error profiles.

Multi-vector retrieval enables residue-resolved prediction of protein partners by ColBERT-PPI

ColBERT-PPI predicts protein partners and their binding sites at the residue level using a multi-vector framework, outperforming existing methods on benchmark datasets.

AMADEUS: Annotation-free multi-animal direction estimation using self-supervised learning

AMADEUS tracks multiple animals without annotations or markings, achieving better performance in identity tracking and keypoint localization across various species.

From single-sequence structure prediction to protein fitness landscape through a composable, epistasis-aware mutation atlas

The paper introduces Cerebra-Epistasis, a framework that improves multi-mutant fitness prediction by modeling epistatic effects and integrating structure-aware sequence analysis.

CMAC: A deep learning framework for absolute cardiomyocyte transcriptional age estimation

CMAC estimates absolute transcriptional age of cardiomyocytes using deep learning, achieving high accuracy across multiple datasets and platforms.

Impacts of Morphology and Elasticity on Cancer Cell Deformation in Shear-flows

This paper examines how cancer cell morphology and elasticity affect deformation under shear-flow, showing morphology sets stiffness-modulated deformation patterns that influence extracellular flow dynamics and traction.

Deciphering Mechanistic Signatures in Drug-Drug Interactions with Dual Topology Graphs

This paper introduces DualTopoDDI, a deep learning model for predicting drug-drug interactions with mechanistic insights, achieving high accuracy and interpretability.