Research Radar — 2026-08-05
Methods & AI
Computational
End-to-end multimodal pathology foundation model with clinical dialogue
Nature Medicine Published 2026-07-31 Research Article DOI: 10.1038/s41591-026-04521-4
computational pathology foundation model whole-slide imaging clinical AI
Summary: Trained on 2.3 million whole-slide images and 14 million clinical question–answer pairs, PRISM2 matched clinical-grade cancer-detection performance without task-specific training, demonstrating the value of clinical dialogue supervision for computational pathology.
Why it matters: The study tests whether clinical language supervision can add useful grounding to large-scale pathology image representations without task-specific training.
Why for Yiru: This connects tissue imaging with clinically grounded multimodal learning, directly relevant to your biomedical AI and spatial biology interests.
SpatialFormer: universal spatial representation learning from subcellular molecular to multicellular landscapes
Nature Computational Science Published 2026-07-30 Research Article DOI: 10.1038/s43588-026-01016-7
spatial transcriptomics representation learning cell-cell relationships cell communication
Summary: SpatialFormer is a transformer-based model for spatial transcriptomics that learns cell–cell relationships to predict cell co-localization, annotate cell types and niches, and reveal cell communication gene pairs in pulmonary fibrosis.
Why it matters: It makes spatial organization a learned representation and links cellular neighborhoods with communication signals in diseased tissue.
Why for Yiru: This is closely aligned with your spatial omics and tumor microenvironment interests, especially learning organized cell states and communication patterns.
CellTune: an integrative software for accurate cell classification in spatial proteomics
Nature Methods Published 2026-07-31 Research Article DOI: 10.1038/s41592-026-03162-2
spatial proteomics cell classification active learning human-in-the-loop
Summary: CellTune enables high precision analysis of spatial proteomics datasets through a human-in-the-loop active learning workflow.
Why it matters: The workflow uses iterative human feedback to address cell classification in spatial proteomics rather than treating annotation as a one-shot automated step.
Why for Yiru: It is a practical tool for your spatial omics work, where accurate cell labels are foundational for interpreting tissue organization.
Continual integration of single-cell multimodal data with MIRACLE
Nature Computational Science Published 2026-07-31 Research Article DOI: 10.1038/s43588-026-01030-9
single-cell multimodal integration continual learning batch correction
Summary: MIRACLE is an online continual learning framework that efficiently integrates single-cell multimodal data across batches, modalities, tissues and diseases.
Why it matters: Continual integration addresses the practical problem of keeping heterogeneous single-cell references usable as new data arrive.
Why for Yiru: This is relevant to your computational biology and spatial multi-omics interests because it targets integration across growing, heterogeneous atlases.
Biomedical discoveries
Biomedicine
Targeting peripheral 5-HT2AR enhances antitumor immunity in colorectal cancer
Cell Published 2026-08-04 Research Article DOI: 10.1016/j.cell.2026.07.028
colorectal cancer 5-HT2AR CD8+ T cells antitumor immunity
Summary: By selectively targeting peripheral 5-HT2AR without inducing psychedelic effects, a non-brain-penetrant agonist boosts antitumor CD8+ T cell immunity and improves immunotherapy responses in preclinical models of colorectal cancer.
Why it matters: The study reports a peripheral receptor-targeting strategy that improves antitumor immunity in the reported colorectal cancer models while avoiding brain penetration.
Why for Yiru: This fits your tumor microenvironment and computational immunology interests by connecting a targetable host pathway with CD8+ T cell response.
CRISPR screens identify targets to rescue age-related T cell dysfunction in cancer
Cell Published 2026-07-29 Research Article DOI: 10.1016/j.cell.2026.07.016
CRISPR screen T cell dysfunction aging tumor microenvironment
Summary: Dusp5 and Zfp219 serve as critical regulators of age-related T cell dysfunction, with loss of either driving T cells into an effector-like state that improves tumor control through distinct mechanisms within the aged tumor microenvironment.
Why it matters: It links genetic perturbations to age-dependent T cell state and tumor control, identifying experimentally testable targets in older hosts.
Why for Yiru: This directly fits your computational immunology, tumor microenvironment and CAR-T interests by connecting CRISPR perturbation with T cell state engineering.
Subclinical cholestasis is a hallmark of gut dysbiosis causing resistance to cancer immunotherapy
Cancer Cell Published 2026-07-29 Research Article DOI: 10.1016/j.ccell.2026.06.022
gut microbiome gut–liver axis immunotherapy resistance tumor immunosurveillance
Summary: Gut dysbiosis impairs tumor immunosurveillance through the gut–liver axis, causing subclinical cholestasis and immunotherapy resistance in tumors distant from gastrointestinal and hepatic tissues. The reported human and mouse analyses include changes in γ-glutamyl transferase, bile acids and sMAdCAM-1.
Why it matters: The findings connect gut microbial state with systemic metabolic and vascular changes associated with cancer immunotherapy response.
Why for Yiru: This is relevant to your tumor microenvironment and computational immunology interests because it connects host state, tissue crosstalk and therapy response.
p38β-mediated BiP phosphorylation drives stemness and chemoresistance by suppressing UPR activation in hepatocellular carcinoma
Nature Communications Published 2026-08-05 Research Article DOI: 10.1038/s41467-026-76073-7
hepatocellular carcinoma p38β/MAPK11 BiP chemoresistance
Summary: Tumor-initiating cells drive chemoresistance in hepatocellular carcinoma. The study reports that p38β/MAPK11 phosphorylation of BiP promotes stemness and chemoresistance in hepatocellular carcinoma.
Why it matters: It identifies a p38β–BiP mechanism connecting stress-response regulation with tumor-initiating-cell behavior and treatment resistance.
Why for Yiru: This is relevant to your tumor biology interests as a mechanistic example of how signaling and proteostasis can shape resistant cancer cell states.
Dietary arginine drives codon-dependent MHC class I translation and improves immunity in colon tumorigenesis and respiratory viral infection
Cell Published 2026-07-30 Research Article DOI: 10.1016/j.cell.2026.07.020
arginine metabolism MHC class I antigen presentation tumor immunity
Summary: Arginine availability regulates arginyl tRNA levels and codon-dependent translation of MHC class I, tuning antigen presentation and shaping anti-viral and anti-tumor immunity.
Why it matters: The study provides a reported link between nutrient availability, codon-dependent translation and MHC class I antigen presentation.
Why for Yiru: This connects metabolic context to immune recognition, a useful lens for your tumor microenvironment and computational immunology interests.
Cross-disciplinary watchlist
Other Fields
Tissue-resident immune cells drive genetic risk in autoimmune and lung diseases
Nature Immunology Published 2026-08-03 Research Article DOI: 10.1038/s41590-026-02596-2
tissue-resident immune cells lung immunity autoimmune disease genetic risk
Summary: The study presents a human lung immune cell atlas linking disease-associated genetic variants to gene expression in tissue-resident immune cells.
Why it matters: It places disease-associated genetic variation in the context of immune cells residing in human tissue rather than only circulating populations.
Why for Yiru: This is a strong reference for your spatial and computational immunology interests, especially linking cell states with disease genetics.
Milk osteopontin alters the infant microbiome to drive DC hematopoiesis and disease tolerance
Cell Published 2026-08-03 Research Article DOI: 10.1016/j.cell.2026.07.022
milk osteopontin infant microbiome dendritic cells disease tolerance
Summary: Milk-derived osteopontin confers protection from viral or bacterial severe lower respiratory illness by altering the neonatal gut microbiome.
Why it matters: The study connects a milk-derived factor, infant microbial ecology and dendritic-cell hematopoiesis with disease tolerance.
Why for Yiru: This is useful for your host–microbe and immune-state interests as an example of developmental systems biology spanning nutrition and immunity.
Cumulative transcription factor binding and p300-mediated histone acetylation drive enhancer activation frequency
Nature Genetics Published 2026-07-31 Research Article DOI: 10.1038/s41588-026-02703-x
enhancers transcription factors histone acetylation single-molecule footprinting
Summary: Using single-molecule footprinting in mouse embryonic stem cells, the study quantifies chromatin accessibility at enhancers and promoters and dissects contributions from transcription factor binding and chromatin context.
Why it matters: It examines how cumulative transcription-factor binding and chromatin context relate to enhancer activation frequency at molecular resolution.
Why for Yiru: This complements your single-cell and regulatory-omics interests by offering a mechanistic layer for interpreting enhancer activity.
Metabolic atlas of early human cortex reveals glycolytic remodeling and pentose phosphate pathway control of cell fate transitions
Cell Published 2026-08-04 Resource DOI: 10.1016/j.cell.2026.07.023
human cortex metabolic atlas glycolysis cell fate
Summary: The study maps the metabolic landscape of the developing human cortex and identifies glycolysis and pentose phosphate pathway activity as regulators of radial glia fate, controlling the generation of outer radial glia and inhibitory neurons.
Why it matters: It connects developmental cell-fate transitions in the human cortex with distinct metabolic pathways.
Why for Yiru: The atlas is relevant to your interest in integrating molecular state, metabolism and tissue development in computational biology.
BCAT2 links branched-chain amino acids metabolism to interferon signaling to sustain macrophage inflammation
Nature Immunology Published 2026-07-31 Research Article DOI: 10.1038/s41590-026-02604-5
macrophages branched-chain amino acid metabolism interferon signaling inflammation
Summary: BCAT2-dependent branched-chain amino acid catabolism sustains IFN-driven macrophage activation and chronic inflammation during autoimmune arthritis.
Why it matters: It identifies a metabolic dependency coupling nutrient processing to persistent interferon signaling and inflammatory macrophage states.
Why for Yiru: The metabolism–macrophage connection is useful for your tumor microenvironment and computational immunology interests, where myeloid state and nutrient context co-vary.