A network-based deep learning model integrating subclonal architecture for therapy response prediction in cancer Kim et al. develop SubNetDL, a deep learning framework that decodes cancer therapy responses by integrating tumor subclonal architecture with molecular networks. By capturing the complex interplay between genetic diversity and functional interactions, this approach provides a robust, interpretable path toward personalized oncology across diverse cancer types.

A network-based deep learning model integrating subclonal architecture for therapy response prediction in cancer

Light-inducible FLPase reconstitution enables temporal control of gene expression in Drosophila melanogaster Shridharan et al. introduce a versatile building block to the Drosophila toolbox, a split-FLP that removes an FRT-stop cassette upon light stimulation. They show that this can be used to activate long-lasting gene expression, with minimal unwanted biological side effects and no transgene leakage in the absence of light.

Light-inducible FLPase reconstitution enables temporal control of gene expression in Drosophila melanogaster

HAK-actin, a U-ExM-compatible probe to image the actin cytoskeleton Expansion microscopy imaging of actin has largely relied on antibodies, limiting studies primarily to well-established model systems. Here, Mercey et al. introduce HAK-actin, an actin-targeting probe compatible with ultrastructure expansion microscopy that enables robust and efficient labeling across a broad range of species.

HAK-actin, a U-ExM-compatible probe to image the actin cytoskeleton

A tri-modal contrastive learning framework for protein representation learning Zhang et al. develop ProteinAligner, a foundation model that teaches AI to understand proteins by simultaneously looking at their genetic sequences, 3D shapes, and scientific descriptions. By merging these diverse perspectives, the model accurately predicts biological functions and properties, offering a powerful tool for drug discovery and biotechnology.

A tri-modal contrastive learning framework for protein representation learning

Differential expression analysis in single-cell and spatial RNA-seq without model assumptions Differential expression analysis is key for identifying genetic checkpoints in biomedical research and diagnostics. Margolin et al. introduce simple tools for more unbiased expression analysis in single-cell and spatial transcriptomics. This analysis relies only on the data and eliminates biases and artifacts caused by unnecessary implicit and model assumptions.

Differential expression analysis in single-cell and spatial RNA-seq without model assumptions

ProtFI, an efficient frailty-trained proteomics-based biomarker of aging, robustly predicts age-related decline Garst et al. systematically compare modeling strategies for aging biomarkers and introduce two efficient protein-based biomarkers trained on frailty and mortality. The findings show that simpler models and minimal protein sets can robustly predict diverse aging outcomes, and that external validation is crucial for ensuring reliable biomarkers.

ProtFI, an efficient frailty-trained proteomics-based biomarker of aging, robustly predicts age-related decline

Spatial mapping of microbial communities by an integrated automation platform of sequential FISH Cao et al. report on SEER-Map, an automated sequential FISH platform with optimized protocols that improve hybridization efficiency and are cost efficient. This high-resolution spatial mapping platform has broad applicability for deciphering micron-scale biogeography and species interactions in complex microbiomes.

Spatial mapping of microbial communities by an integrated automation platform of sequential FISH

R-PLA enables enhanced detection of R-loops in mammalian cells using proximity ligation Sanchez et al. report the development of R-PLA, a proximity-ligation-assay-based approach for the detection of cellular R-loops by fluorescence microscopy. They demonstrate that this technique provides improved resolution and specificity over existing approaches for R-loop detection and can be applied in a variety of cellular settings.

R-PLA enables enhanced detection of R-loops in mammalian cells using proximity ligation

Cellular-state control using ribozyme-scaffolded miRNA-sensing and CRISPR-mediated actuation Kang and Bleris utilize an miRNA-directed CRISPR platform to convert endogenous miRNA signatures into state-specific response. Dual-sgRNA constructs with ribozyme scaffolds restrict effector activation to defined states. In an EMT model, the system filters heterogeneous populations to selectively eliminate mesenchymal cells.

Cellular-state control using ribozyme-scaffolded miRNA-sensing and CRISPR-mediated actuation

A lightweight TcrLM model predicts T cell receptor and epitope binding specificity Yu et al. develop a lightweight language model to predict how immune T cell receptors recognize disease-related targets. Trained on millions of immune sequences, this model demonstrates robust prediction accuracy despite its streamlined parameter size.

A lightweight TcrLM model predicts T cell receptor and epitope binding specificity

A high-speed attention network for MHC-bound peptide identification and 3D modeling Baakman et al. introduce SwiftMHC, an ultra-fast AI method that predicts peptide-immune molecule interactions and their 3D structures. The framework enables rapid screening of potential cancer therapy targets, helping bridge the gap between computational speed and structural insight.

A high-speed attention network for MHC-bound peptide identification and 3D modeling

Microfluidics-enabled proteomic profiling reveal iron-driven immune evasion by an antimicrobial-resistant pathogen Nguyen et al. introduce a microfluidic-proteomics workflow that separates immune cells from bacteria during infection. This approach reveals how Klebsiella pneumoniae uses iron-associated proteins to evade macrophage clearance, providing a powerful way to study host-pathogen interactions and uncover mechanisms of immune escape.

Microfluidics-enabled proteomic profiling reveal iron-driven immune evasion by an antimicrobial-resistant pathogen

Multiscale domain identification for spatial transcriptomics via persistent homology Beamer et al. introduce PHD-MS, a tool for spatial transcriptomics that finds spatial patterns at multiple scales. By scanning clustering results from broad to fine substructures, it links large tissue domains to their subregions and reveals overlooked within-region heterogeneity in gene expression.

Multiscale domain identification for spatial transcriptomics via persistent homology

3D imaging with enhanced transparency, signal-to-background ratios, and antigen detection using HyPer-3D Choi et al. develop HyPer-3D, a technique that enhances 3D imaging of cleared organs by significantly increasing signal-to-background ratios, optical clearing capacity, and visualization of poorly recognized antigens and fluorescent reporters. HyPer-3D enables enhanced, high-contrast multiplex interrogation of complex tissue architectures across experimental models, including human specimens, with improved signal discernability.

3D imaging with enhanced transparency, signal-to-background ratios, and antigen detection using HyPer-3D

A homecage operant paradigm reveals behavioral dynamics of social motivation in mice Chen et al. describe an automated homecage task that measures social motivation in mice over days. By combining operant responses with proximity sensing, the method distinguishes social seeking from environmental exploration and captures reciprocal social interactions, revealing how social motivation varies across the day and is shaped by dopamine signaling.

A homecage operant paradigm reveals behavioral dynamics of social motivation in mice

Topography-aware optimal transport for alignment of spatial omics data Ceccarelli et al. introduce TOAST, a topography-aware optimal transport framework for spatial omics alignment. By incorporating spatial coherence and neighborhood consistency, TOAST improves slice alignment, multimodal integration, and spatial reconstruction across diverse transcriptomic and proteomic datasets, enabling biologically interpretable mappings of tissue architecture.

Topography-aware optimal transport for alignment of spatial omics data

Wnt activation and dual SMAD inhibition for induction and maintenance of hindbrain-like neural stem cell from hiPSCs Al-Akashi et al. develop a robust and simplified protocol to generate hindbrain-specific neural stem cells from human iPSCs using a defined small-molecule combination. These cells maintain their regional identity and functional potential for over a year, providing a scalable tool for studying brain development, disease modeling, and potential cell therapies.

Wnt activation and dual SMAD inhibition for induction and maintenance of hindbrain-like neural stem cell from hiPSCs

Biobank of genetically defined murine prostate cancer tumoroids uncovers oncogenic pathways and drug vulnerabilities driven by PTEN-loss Kalla et al. establish and characterize a biobank of murine prostate organoids and prostate cancer tumoroids that model key patient-relevant mutations, providing a versatile platform for drug discovery. A medium-throughput drug screen identifies targeted compounds that effectively inhibit prostate cancer growth and enhance antiandrogen therapy for future treatment development.

Biobank of genetically defined murine prostate cancer tumoroids uncovers oncogenic pathways and drug vulnerabilities driven by PTEN-loss

Assessing leaky expression of Cre-dependent DNA constructs in the mouse genome using sensitive bioluminescent reporters Using the highly sensitive AkaBLI system, Nakashiba et al. reveal that Cre-dependent DNA constructs with a transcriptional STOP cassette exhibit significant leakage when integrated into the mouse genome. The combination of transcriptional and translational controls successfully suppresses this leakage, enabling comprehensive whole-body mapping of Cre recombination sites.

Assessing leaky expression of Cre-dependent DNA constructs in the mouse genome using sensitive bioluminescent reporters

Projection-specific intersectional optogenetics for precise excitation and inhibition in the marmoset brain Shaw et al. develop an intersectional optogenetic strategy that enables precise excitation and inhibition of defined long-range cortical projections in the marmoset, providing a scalable framework for causal analysis of distributed neural circuits in the primate brain.

Projection-specific intersectional optogenetics for precise excitation and inhibition in the marmoset brain

Benchmarking RNA velocity methods across 17 independent studies In a benchmark study utilizing both real and simulated single-cell RNA sequencing datasets, Luo et al. systematically evaluate 15 RNA velocity methods across three dimensions and propose scenario-based recommendations of best-practice.

Benchmarking RNA velocity methods across 17 independent studies

Electrophysiological development and functional plasticity in dissociated human cerebral organoids across multiple cell lines Microelectrode array (MEA) recordings of human neural organoids allow for functional studies of brain network development, but these studies are currently limited in throughput. Pavlinek et al. use organoid dissociation to longitudinally study the development of neural networks and the effects of drug treatment, cell line, and cell-type composition.

Electrophysiological development and functional plasticity in dissociated human cerebral organoids across multiple cell lines

Development of motif-specific monoclonal antibodies for global protein citrullination detection with minimal cross-reactivity to homocitrullination Laposchan et al. report a motif-based strategy to generate monoclonal antibodies for global detection of protein citrullination. The antibodies show strong sensitivity and selectivity over homocitrullination, enabling quantitative detection of global citrullination in cells and lysates. This approach addresses key limitations of existing anti-pan citrullination reagents.

Development of motif-specific monoclonal antibodies for global protein citrullination detection with minimal cross-reactivity to homocitrullination

Persistent homology-based segmentation tool for membrane images

Intensity modulation of trichromatic split fluorescent proteins for live cell mapping Ishii et al. present Caterpie, a rationally designed cell-labeling platform using engineered split fluorescent proteins. By modulating color and intensity combinations, Caterpie achieves precise identification of 20 distinct live-cell populations with 97% accuracy, enabling high-fidelity mapping and analysis of complex multicellular interactions.

Intensity modulation of trichromatic split fluorescent proteins for live cell mapping

Immune response to adeno-associated viral vectors in a human iPSC-derived microglia chimeric mouse model Detrez et al. generate a human microglia chimeric mouse model in which human microglia, generated from cryopreserved precursors, engraft in the mouse CNS, adopt primary human microglial features, and maintain functional responsiveness. This chimeric model reveals that AAV serotypes elicit divergent immune responses in mouse microglia and human microglia.

Immune response to adeno-associated viral vectors in a human iPSC-derived microglia chimeric mouse model

Integrating single-cell and single-nucleus datasets improves bulk RNA-seq deconvolution Ivich and Greene evaluate how reference modality choice influences bulk RNA sequencing deconvolution accuracy. By comparing single-cell and single-nucleus references across tissues, they show that modality-aware strategies enable effective integration and provide practical guidance for reference selection in deconvolution analyses.

Integrating single-cell and single-nucleus datasets improves bulk RNA-seq deconvolution

Single-cell assessment of iron content in primary human T cells using laser ablation inductively coupled plasma mass spectrometry Iron is an essential transition metal necessary for metabolic and signaling functions in immune cells. Carp et al. present a method for the preparation of human T cells for visual inspection and analysis of intracellular iron content at a single-cell level using laser ablation inductively coupled mass spectrometry (LA-ICP-MS).

Single-cell assessment of iron content in primary human T cells using laser ablation inductively coupled plasma mass spectrometry

MASTR-seq enables multiplexed analysis of short tandem repeats with sequencing Su et al. present MASTR-seq, a cost-effective, high-throughput method for precisely measuring short tandem repeat tract length and DNA methylation at a pre-defined locus at single-molecule resolution using nanopore sequencing.

MASTR-seq enables multiplexed analysis of short tandem repeats with sequencing

Metapipeline-DNA: A comprehensive germline and somatic genomics Nextflow pipeline Patel et al. develop an automated, extensible, and cloud-compatible DNA sequencing analysis pipeline for DNA sequencing data to transform raw sequencing reads into genetic characteristics and evolutionary features. They demonstrate and validate the pipeline using whole-genome and targeted sequencing data from normal and tumor samples.

Metapipeline-DNA: A comprehensive germline and somatic genomics Nextflow pipeline