Single-cell and spatial transcriptomics

Integrative Single-Cell Transcriptomics and Explainable AI for Disease Mechanism and Biomarker Discovery in Triple-Negative Breast Cancer

The paper uses single-cell RNA-seq and explainable AI to identify transcriptional features distinguishing invasive from non-invasive TNBC cells in a cell line, finding EMT-related and inflammatory genes as key predictors.

CATFISHR: integration of transcriptome and inferred CNV profiles for comprehensive malignant cell identification

CATFISHR integrates transcriptomic and copy-number data to improve malignant cell identification in scRNA-seq, achieving high accuracy across multiple datasets and identifying previously missed malignant populations.

Mitigating Bias in Spatial Transcriptomic Pipelines via Human Feedback

The paper introduces CSDE, a method that improves differential expression analysis in spatial transcriptomics by incorporating expert feedback to mitigate preprocessing biases.

Quantifying epigenetic reprogramming and cellular plasticity in preneoplastic lesions reveals dedifferentiation as a key precancer hallmark

PLASTIC quantifies cellular plasticity and epigenetic reprogramming in preneoplastic lesions, revealing dedifferentiation as a key factor in cancer risk.

Single-cell spatiotemporal transcriptomics reveals the origin, cellular heterogeneity and regulatory network of seed fiber cells in poplar

This study uses single-nucleus RNA sequencing and spatial transcriptomics to map seed fiber development in poplar, identifying their origin and regulatory network.

Building virtual cell dynamics that evolve with biological measurements

CellCompass, an agentic AI system, models cell dynamics from single-cell measurements and improves predictions by integrating new biological constraints and experimental data.

How Much Can AI Virtual Cells Gain from Unified Perturbational Modeling?

This paper evaluates how unified perturbational modeling improves AI virtual cell predictors by integrating heterogeneous Perturb-seq datasets and achieving better performance across benchmarks.

Evaluation design shapes estimates of cross-donor generalization in single-cell classification of cell-type

The paper evaluates how different evaluation designs affect cross-donor generalization in single-cell cell-type classification and finds that random cell-level splitting overestimates performance compared to donor-held-out evaluation.

BLOOM: Biologically Informed Flow Matching for Transcriptomic Virtual Embryo Modeling

BLOOM generates developmental cell-expression distributions across time using conditional flow matching, outperforming baselines in gene and cell-expression correlations.

Inferring somatic copy number aberrations from spatial transcriptomics data to study early cancer development

This study infers somatic copy number aberrations from spatial transcriptomics data to identify spatially coherent SCNA-defined populations in early cancer development.

Dimensionality Reduction for Cyclic Compositional Data

clrcycle is a dimensionality reduction method for cyclic data that captures phase and magnitude in biological systems, showing effectiveness in circadian and cell-cycle analysis.

An evidence-governed agent harness for spatial transcriptomics prioritizes ICAM-1 as a target in lung fibrosis

DeepMAST, an AI system for spatial transcriptomics, identifies ICAM-1 as a therapeutic target in lung fibrosis through evidence-based analysis and validation.

Prediction transfer does not imply reliability transfer in single-cell foundation models

The paper investigates reliability transfer in single-cell foundation models across biological contexts and finds it varies independently from prediction transfer.

RFind-sc maps single cells onto biological axes defined by external gene signatures

RFind-sc maps single cells onto biological axes defined by external gene signatures, revealing cell states and candidate genes for stress responses.

ALHENA, an H&E encoder distilled from transcriptomic foundation models for multi-scale slide analysis

ALHENA improves transcriptomic signature prediction and cell-type separability using H&E images alone, without requiring spatial transcriptomics data.

Probabilistic uncertainty in virtual spatial transcriptomics from histology is largely redundant with the predicted mean

This study benchmarks uncertainty estimation in virtual spatial transcriptomics, finding that predicted means largely explain prediction error, and uncertainty provides limited additional information.

Language models annotate mass cytometry clusters within the range where human experts disagree

This paper uses large language models to annotate mass cytometry clusters, achieving high concordance with expert annotations and outperforming baselines.

A knowledge graph-guided framework for multitask single-cell analysis

scKNIFE integrates gene expression with functional annotations to identify tumor programs and treatment-related functional variation in breast cancer.

Multi-scale microenvironment covariance for unsupervised spatial domain identification

The paper introduces CoSMIM, a method for spatial domain identification that uses multi-scale gene-gene covariance matrices, achieving top performance on multiple benchmark datasets.

corrdeconv corrects deconvolution's misreading of disease cell states as abundance, overturning benchmarks and outperforming every published deconvolution method

corrdeconv corrects deconvolution errors by addressing misread disease cell states as abundance, outperforming existing methods and reducing refuted effects.

Dissecting cellular and molecular networks of aging and senescence in the human hippocampus

This study identifies spatially resolved aging and senescence signatures in the human hippocampus using multi-omics approaches, revealing cell-type and sex-specific gene expression changes.

Denoising single-cell ATAC-seq data with Shannon entropy

This paper introduces a Shannon entropy-based method to filter high-quality nuclei in scATAC-seq data, improving signal-to-noise ratio and downstream analysis reliability.

Robust and accessible in situ sequencing with an open-source automated sequencer

The paper introduces Seq-o-Matic, an open-source automated in situ sequencer that reduces hands-on time and improves sequencing consistency across tissue sections.

General Purpose AI for Public Data Analysis and Experimental Hypothesis Generation

The paper demonstrates how general-purpose AI can analyze public molecular datasets to generate hypotheses and identify biological relationships, though not all findings were validated experimentally.

LiverHomo: An Integrated and Interactive Human Liver Single-Cell Atlas with Protein Interaction and Cell-Cell Communication Networks

LiverHomo integrates human liver scRNA-seq data with PPI and cell-cell communication networks to provide an interactive resource for exploring liver health and disease.

Cell-level integration of gene expression and B cell receptor data captures complementary B cell information

The paper integrates gene expression and B cell receptor data to capture complementary B cell information, revealing substructure supported by both transcriptional and receptor features.

GeoSinkhorn Flow: Geometry-Aware Flow Matching for Perturbation-Conditioned Dynamics in Single cell Data Phenoscapes

GeoSinkhorn Flow converts static phenoscapes into dynamic ones by learning perturbation-conditioned transport trajectories that respect the phenoscape's geometry and generalize to unseen conditions.

SPATNIC enables H&E-grounded cancer-cell identification across spatial and single-cell transcriptomic modalities in colorectal cancer

SPATNIC is a variational autoencoder that identifies cancer cells from spatial and single-cell transcriptomic data in colorectal cancer, outperforming existing tools and identifying a rare cancer stem cell population.

LoCoFlow: Local-Context Flow Matching for Single-Cell Dynamics and Fate Inference

LoCoFlow improves trajectory inference in single-cell data by incorporating local geometry and fate uncertainty, achieving better performance than existing methods.

A Community-Driven Single-Cell PBMC Reference Integrating Landmark Datasets Spanning Health and Disease

HARP integrates PBMC data from multiple studies to create a reference atlas with consensus annotations and identifies age, sex, and disease-related cellular modules.