Research Radar — 2026-07-30
Methods & AI
Computational
RETROFIT: Reference-free deconvolution of cell-type mixtures in spatial transcriptomics
Nature Communications Published 2026-07-24 Research Article DOI: 10.1038/s41467-026-74928-7
spatial transcriptomics deconvolution Bayesian method reference-free
Summary: RETROFIT is a reference-free Bayesian method that infers cell-type composition and spatial expression profiles from spatial transcriptomics data without relying on external single-cell references. It leverages spatial autocorrelation and gene expression covariance to resolve mixed cell populations directly.
Why it matters: Solves a key bottleneck in spatial transcriptomics by enabling reference-free deconvolution, making spatial analysis applicable to tissues where matched single-cell data is unavailable.
Why for Yiru: Deconvolution is central to spatial transcriptomics analysis. This reference-free approach could be immediately useful in your spatial data analysis workflow, especially for tumor microenvironment studies.
Full-length single-cell spatial transcriptomics reveals spatial and cell-type-specific transcript isoforms in the primate brain
Nature Methods Published 2026-07-24 Research Article DOI: 10.1038/s41592-026-03174-y
spatial transcriptomics single-cell long-read sequencing isoform detection
Summary: Fullscope-seq combines Stereo-seq with long-read sequencing to achieve single-cell isoform-level resolution in spatial transcriptomics on macaque brain slices, enabling detection of spatial and cell-type-specific transcript isoforms across large tissue areas.
Why it matters: Adds a new dimension to spatial transcriptomics by revealing full-length isoform diversity in spatial context, bridging the gap between bulk isoform studies and single-cell spatial analysis.
Why for Yiru: Directly relevant to your spatial transcriptomics research. The isoform-resolution approach could inspire new analytical strategies for detecting functionally distinct transcript variants in the tumor microenvironment.
Spatialproteomics: an interoperable toolbox for analyzing highly multiplexed fluorescence image data
Nature Methods Published 2026-07-24 Research Article DOI: 10.1038/s41592-026-03155-1
spatial proteomics multiplexed imaging Python toolbox image analysis
Summary: Spatialproteomics is a Python-based toolbox supporting end-to-end analysis of highly multiplexed fluorescence imaging data, providing standardized workflows for cell segmentation, feature extraction, and spatial analysis across diverse platforms.
Why it matters: Provides a unified Python framework for analyzing highly multiplexed imaging data, standardizing previously fragmented analysis workflows and improving reproducibility.
Why for Yiru: Spatial proteomics complements spatial transcriptomics in your multi-omics toolkit. This toolbox could streamline your multiplexed imaging analysis pipeline for tumor microenvironment studies.
A benchmark study of vision and pathology foundation models for computational pathology
Nature Communications Published 2026-07-24 Research Article DOI: 10.1038/s41467-026-76004-6
foundation model computational pathology benchmarking AI
Summary: Benchmarks 32 pathology and vision AI foundation models on large-scale cancer datasets, showing that generalisation is heterogeneous and task-dependent. Ensemble approaches can combine the strengths of different models for improved performance.
Why it matters: Provides the most comprehensive comparison of pathology and vision foundation models to date, revealing critical insights about when domain-specific pretraining matters for computational pathology tasks.
Why for Yiru: Foundation models are a key interest. Understanding when pathology-specific models outperform general vision models is valuable for selecting AI tools in your cancer imaging and spatial analysis workflows.
Single-nucleus multimodal spatial transcriptomics reveals spatial colocalization of neoantigen-expressing tumor cells and cognate T cells
Nature Biotechnology Published 2026-07-22 Research Article DOI: 10.1038/s41587-026-03194-1
spatial transcriptomics neoantigen T cell tumor microenvironment
Summary: Combines single-nucleus multimodal spatial transcriptomics with TCR sequencing to directly visualize spatial colocalization of neoantigens and cognate T cells in the tumor microenvironment at single-cell resolution.
Why it matters: First direct spatial evidence of neoantigen-T cell colocalization in human tumors at single-nucleus resolution, establishing a framework for understanding how spatial immune organization governs anti-tumor immunity.
Why for Yiru: Directly aligns with your spatial transcriptomics and tumor immunology interests. The multimodal approach to studying T cell-tumor spatial interactions is highly relevant to your research.
Expanding the scope of protein language modeling to protein-protein interactions with MSA Pairformer
Cell Published 2026-07-20 Research Article DOI: 10.1016/j.cell.2026.06.029
deep learning protein language model protein-protein interactions AI
Summary: MSA Pairformer is a parameter-efficient protein language model that generalizes to protein-protein interactions despite being trained exclusively on monomer sequences. It enables accurate prediction of interface variant effects and subfamily-specific homo-oligomeric binding modes.
Why it matters: Bridges the gap between single-chain protein language models and interaction prediction, enabling AI-driven design of protein complexes without requiring training data on protein interfaces.
Why for Yiru: The extension to pairwise interactions demonstrates a principled approach that could inspire similar advances in modelling ligand-receptor interactions in spatial transcriptomics and cell-cell communication in the tumor microenvironment.
Biomedical discoveries
Biomedicine
Disease-associated microglia adopt stage-specific phenotypes that regulate T cell fate and immunity in glioma
Immunity Published 2026-07-23 Research Article DOI: 10.1016/j.immuni.2026.06.024
microglia glioma T cell tumor microenvironment
Summary: Disease-associated microglia (DAMs) in glioma are not static suppressors but evolve with tumor stage, transitioning through antigen-presenting, checkpoint-regulatory, and T cell-clearing programs that directly regulate T cell fate and immunity.
Why it matters: Reveals that brain-resident microglia dynamically reprogram their phenotypes during glioma progression, directly shaping local T cell responses and providing new targets for brain cancer immunotherapy.
Why for Yiru: Your interest in macrophage biology and the tumor microenvironment makes this highly relevant. Dynamic microglia phenotypes could inform myeloid-targeting strategies for glioblastoma immunotherapy.
Localized PD-1 CAR T therapy reprograms neuroinflammation
Cell Published 2026-07-20 Research Article DOI: 10.1016/j.cell.2026.06.036
CAR-T immunotherapy T cell neuroinflammation
Summary: Programmable PD-1-targeting CAR T cells selectively eliminate pathogenic CD4 T cells while delivering IL-10 at sites of inflammation, effectively suppressing neuroinflammation in preclinical models of multiple sclerosis.
Why it matters: Demonstrates that CAR T cell therapy can be adapted for non-oncological indications by combining targeted cell elimination with localized delivery of therapeutic payloads.
Why for Yiru: Highly relevant to your CAR-T and immunotherapy interests. The concept of localized CAR T delivery with payload release could be adapted for solid tumors to overcome trafficking barriers and improve the therapeutic index.
Tertiary lymphoid structures harbour stem-like tumour-specific T cells
Nature Published 2026-07-22 Research Article DOI: 10.1038/s41586-026-10808-w
tertiary lymphoid structures T cell renal cell carcinoma tumor microenvironment
Summary: Renal cell carcinoma tumours containing tertiary lymphoid structures (TLSs) are enriched for exhausted CD8+ T cells with stem-like progenitor features and reduced terminal exhaustion, revealing that TLSs serve as niches for maintaining stem-like tumour-specific T cells.
Why it matters: Provides mechanistic insight into why TLSs correlate with favourable immunotherapy outcomes, establishing them as functional immune hubs that maintain progenitor T cell populations.
Why for Yiru: Directly relevant to your tumor microenvironment and immunotherapy interests. The stem-like T cell population in TLSs represents a potential target for combination strategies preserving progenitor T cell function.
Blocking the m6Am methyltransferase PCIF1 releases STAT1-mediated Th1 immunity to potentiate cancer immunotherapy
Nature Communications Published 2026-07-22 Research Article DOI: 10.1038/s41467-026-75269-1
immunotherapy epitranscriptomics T cell PCIF1
Summary: Deleting PCIF1 in T cells releases STAT1-mediated Th1 immunity by altering m6Am methylation on STAT1 mRNA, identifying a new epitranscriptomic mechanism that restricts anti-tumor immunity and can be targeted to potentiate cancer immunotherapy.
Why it matters: Discovers PCIF1 as a T cell-intrinsic epitranscriptomic checkpoint, revealing that RNA methylation regulates STAT1 translation and T cell activation with immediate therapeutic implications.
Why for Yiru: Connects epitranscriptomics with immunotherapy — directly relevant to your interests in immune mechanisms and T cell biology. Targeting RNA modifications could enhance CAR-T and checkpoint immunotherapy efficacy.
A self-amplifying nerve-fibroblast circuit drives colorectal cancer progression
Cancer Cell Published 2026-07-23 Research Article DOI: 10.1016/j.ccell.2026.06.018
colorectal cancer cancer-associated fibroblast tumor microenvironment nerve-tumor crosstalk
Summary: Demonstrates a self-amplifying circuit where cholinergic signaling induces NTN1 secretion from colorectal cancer-associated fibroblasts, enhancing intratumoral cholinergic innervation and accelerating cancer growth through CHRM3 and UNC5B receptors.
Why it matters: Reveals a previously unknown self-amplifying nerve-CAF-cancer circuit that drives colorectal cancer progression, identifying CHRM3 and NTN1 as potential therapeutic targets.
Why for Yiru: Tumor microenvironment interactions are central to your research. This nerve-CAF-cancer axis reveals a new dimension of TME crosstalk beyond immune-stromal interactions.
Quiescent tumor cells shape the immunosuppressive microenvironment in pancreatic cancer
Nature Communications Published 2026-07-21 Research Article DOI: 10.1038/s41467-026-75883-z
pancreatic cancer CAR-T macrophage immunosuppression
Summary: A rare population of quiescent pancreatic cancer cells increases after CAR-T therapy, expressing high levels of Epiregulin (EREG) which induces an immunosuppressive tumor microenvironment via ErbB4-expressing macrophages. Targeting EREG enhances CAR-T sensitivity.
Why it matters: Identifies a therapy-induced resistance mechanism in which quiescent tumor cells actively remodel the microenvironment through EREG-ErbB4 macrophage signaling, providing a targetable axis to improve CAR-T efficacy.
Why for Yiru: Highly relevant to your CAR-T, macrophage, and tumor microenvironment interests. The finding that quiescent tumor cells communicate with macrophages to drive immunosuppression identifies a targetable vulnerability for solid tumor CAR-T therapy.
Cross-disciplinary watchlist
Other Fields
AI-redesigned starting points and outcomes enhance protein evolution
Nature Published 2026-07-22 Research Article DOI: 10.1038/s41586-026-10820-0
protein engineering AI directed evolution deep learning
Summary: A workflow using AI-redesigned starting points to evolve enzymes yields improved properties compared with those evolved from natural proteins, demonstrating that AI-generated scaffolds can serve as superior templates for directed evolution.
Why it matters: Establishes that AI-designed protein scaffolds outperform natural proteins as starting points for directed evolution, significantly accelerating the development of enzymes with industrial and therapeutic applications.
Why for Yiru: Relevant to your AI interests — the concept of AI-optimized starting points could extend to designing improved immune receptors, CAR constructs, or other therapeutic proteins with enhanced functionality.
Precise DNA base editing using AlphaFold3-based contact modelling
Nature Published 2026-07-22 Research Article DOI: 10.1038/s41586-026-10794-z
gene editing AlphaFold deep learning CRISPR
Summary: ContactSeek is an AlphaFold3-driven model that uses contact maps to predict off-target effects of base editors with high accuracy, demonstrating how protein structure prediction can directly enhance the safety and specificity of therapeutic genome editing.
Why it matters: Addresses a key clinical barrier for therapeutic genome editing by using AI-driven protein structure prediction to improve the precision of base editors.
Why for Yiru: The ContactSeek framework represents a template for how AI-driven protein structure prediction can improve molecular tool design — a paradigm applicable to engineering improved CAR constructs and other therapeutic proteins.
CRISPR–Cas regulates expression of embedded anti-phage defence systems
Nature Published 2026-07-22 Research Article DOI: 10.1038/s41586-026-10833-9
CRISPR gene regulation phage defence microbiology
Summary: CRISPR-Cas systems transcriptionally tune diverse innate defences using CRISPR RNA-like guides to balance antiviral protection with fitness. When compromised, these systems hyperactivate embedded defences, establishing a layered bacterial 'immunity guard' network.
Why it matters: Reveals a previously unrecognized regulatory function of CRISPR-Cas systems beyond adaptive immunity, showing they act as master regulators of a layered bacterial antiviral defence network.
Why for Yiru: While primarily in microbiology, the finding that CRISPR systems have regulatory functions beyond DNA targeting expands the conceptual framework for how these systems could be engineered for new genome engineering applications.
Genetic background sets the trajectory of experimental cancer evolution
Nature Published 2026-07-22 Research Article DOI: 10.1038/s41586-026-10821-z
cancer evolution genetic background tumor initiation cancer
Summary: Experimentally replaying tumour evolution in divergent mouse strains reveals that interactions between genetic ancestry and acquired cancer-driving mutations shape the earliest stages of cancer development, with different genetic backgrounds predisposing to distinct evolutionary trajectories.
Why it matters: Provides direct experimental evidence that genetic background fundamentally shapes cancer evolutionary trajectories, interacting with somatic mutations to determine tumour initiation and progression.
Why for Yiru: Understanding how genetic background influences cancer evolution provides important context for interpreting tumour heterogeneity and immune microenvironment differences in your cancer studies.
Spatial multi-omics identifies early synaptic pruning and context-specific dopaminergic vulnerability in synucleinopathies
Nature Communications Published 2026-07-21 Research Article DOI: 10.1038/s41467-026-74961-6
multi-omics spatial transcriptomics neurodegeneration synucleinopathy
Summary: Spatial multi-omics profiling identifies early complement-associated pruning of inhibitory synapses that precedes overt alpha-synuclein aggregation in prodromal synucleinopathy, with context-specific dopaminergic vulnerability across brain regions.
Why it matters: Reveals that synaptic dysfunction occurs before detectable protein aggregation in synucleinopathies, identifying early complement-mediated pruning as a potential therapeutic window.
Why for Yiru: The spatial multi-omics approach is directly relevant to your multi-omics interests. The methodology for integrating spatial transcriptomics with proteomics could inspire similar analyses in the tumor microenvironment.
Neural sampling from cognitive maps enables goal-directed imagination and planning
Nature Machine Intelligence Published 2026-07-21 Research Article DOI: 10.1038/s42256-026-01254-4
AI cognitive science neural networks deep learning
Summary: A brain-inspired generative model uses cognitive maps, stochastic computing, and compositional coding to enable goal-directed imagination and planning, demonstrating how neural sampling from internal representations can solve complex planning problems.
Why it matters: Bridges cognitive neuroscience and AI by demonstrating how neural sampling from cognitive maps can enable flexible planning and problem-solving without explicit reward models.
Why for Yiru: The AI and deep learning principles behind this neural sampling approach could inspire novel computational methods for modelling cellular decision-making and state transitions in the tumor microenvironment.