Research Radar — 2026-08-13
Methods & AI
Computational
Reliable single-cell perturbations explain and improve model performance
bioRxiv Published 2026-08-12 Preprint DOI: 10.64898/2026.08.11.744177v1
single-cell perturbation model evaluation computational biology
Summary: The study examines how reliable single-cell perturbation measurements explain and improve computational model performance.
Why it matters: It connects experimental perturbation reliability with the validity of downstream predictive models.
Why for Yiru: This directly matches your interest in robust, agent-ready single-cell analysis and perturbation modeling.
Spatial multi omics enables single cell transcriptome metabolome inference
bioRxiv Published 2026-08-12 Preprint DOI: 10.64898/2026.08.06.743252v1
spatial omics single-cell transcriptomics metabolomics inference
Summary: The article presents a spatial multi-omics approach for inferring single-cell transcriptome and metabolome states.
Why it matters: Joint molecular inference could help connect spatial cell identity with local metabolic programs.
Why for Yiru: It is closely aligned with your spatial-omics and multimodal computational biology interests.
AnchorR: A QuPath and R interface for collaborative exploration of spatial transcriptomics and histology
bioRxiv Published 2026-08-11 Preprint DOI: 10.64898/2026.08.05.742985v1
spatial transcriptomics histology QuPath R
Summary: AnchorR provides a QuPath and R interface for collaborative exploration of spatial transcriptomics data alongside histology.
Why it matters: A shared interface can reduce friction between image-based pathology review and transcriptomic analysis.
Why for Yiru: This is a practical fit for your interest in usable, collaborative spatial-analysis infrastructure.
pysigscore: gene signatures scoring across bulk and single-cell transcriptomics
bioRxiv Published 2026-08-09 Preprint DOI: 10.64898/2026.08.04.742537v1
gene signatures bulk transcriptomics single-cell transcriptomics Python
Summary: pysigscore is presented as a tool for scoring gene signatures across bulk and single-cell transcriptomic datasets.
Why it matters: Consistent signature scoring across data modalities supports reproducible biological interpretation.
Why for Yiru: It could be useful infrastructure for your transcriptomic workflows and automated analysis agents.
REFCON: Reference-free and robust copy number inference in single-cell tumor transcriptomes
bioRxiv Published 2026-08-07 Preprint DOI: 10.64898/2026.08.03.742406v1
single-cell transcriptomics copy-number inference tumor genomics
Summary: REFCON proposes reference-free copy-number inference for single-cell tumor transcriptomes.
Why it matters: Reducing dependence on external references may make copy-number analysis more robust across tumor datasets.
Why for Yiru: This matches your interests in scalable single-cell tumor genomics and reliable preprocessing.
PIANO: Probabilistic Inference Autoencoder Networks for multi-Omics enables robust generative modeling of gene expression and scales single-cell integration to 100 million cells
bioRxiv Published 2026-08-12 Preprint DOI: 10.64898/2026.08.06.743394v1
multi-omics generative modeling single-cell integration scalability
Summary: PIANO uses probabilistic autoencoder networks for multi-omics gene-expression modeling and large-scale single-cell integration.
Why it matters: The reported scale targets a major systems challenge in integrating very large single-cell datasets.
Why for Yiru: This is directly relevant to your interest in scalable, model-driven single-cell infrastructure.
Biomedical discoveries
Biomedicine
Acquired resistance to the RAS(ON) multi-selective inhibitor daraxonrasib guides rational combination therapy strategies in pancreatic cancer
www.nature.com Published 2026-08-11 Research article DOI: 10.1038/s41591-026-04537-w
RAS inhibition drug resistance pancreatic cancer combination therapy
Summary: The study investigates acquired resistance to the RAS(ON) inhibitor daraxonrasib in pancreatic cancer and uses the resistance mechanisms to guide combination strategies.
Why it matters: It links treatment-emergent resistance to rational design of follow-up therapies.
Why for Yiru: This fits your drug-related omics and pharmacobiology interests, especially resistance-aware therapeutic modeling.
MALT1 protease inhibition restrains glioblastoma progression by reversing tumor-associated macrophage-dependent immunosuppression in mice
www.nature.com Published 2026-08-10 Research article DOI: 10.1038/s41467-026-76572-7
glioblastoma MALT1 tumor-associated macrophages immunosuppression
Summary: In mouse models, MALT1 protease inhibition restrains glioblastoma progression while reversing tumor-associated macrophage-dependent immunosuppression.
Why it matters: The work identifies a macrophage-linked immunosuppressive mechanism that may be therapeutically reversible.
Why for Yiru: It directly connects computational immunology themes with tumor-microenvironment intervention.
AI-Driven Computational Design of Peptide-Based WWP1 Inhibitors as Promising Therapeutic Agents Against Breast Cancer, Including Triple-Negative Subtype
bioRxiv Published 2026-08-10 Preprint DOI: 10.64898/2026.08.08.742959v1
AI drug design peptides WWP1 breast cancer
Summary: The preprint describes AI-driven computational design of peptide-based WWP1 inhibitors for breast cancer, including triple-negative disease.
Why it matters: It illustrates how computational design is being applied to nominate peptide therapeutics against a cancer target.
Why for Yiru: This is a direct match for your biomedical-AI and drug-discovery interests.
Therapeutic signature mapping of paired direct and indirect LPS injury in an ex vivo human lung perfusion platform reveals injury-specific druggable programs
bioRxiv Published 2026-08-07 Preprint DOI: 10.64898/2026.08.03.739838v1
lung injury LPS ex vivo perfusion drug targets
Summary: An ex vivo human lung perfusion platform maps therapeutic signatures of paired direct and indirect LPS injury and identifies injury-specific druggable programs.
Why it matters: The platform separates injury contexts that may otherwise be combined under a single inflammatory signature.
Why for Yiru: This suits your interest in translational systems and context-aware drug-related omics.
RNA terminal uridylyl transferases are druggable vulnerabilities in AML but are dispensable for normal hematopoiesis
www.science.org Published 2026-08-12 Research article DOI: 10.1126/sciadv.aec3399
AML RNA regulation drug targets hematopoiesis
Summary: The article identifies RNA terminal uridylyl transferases as druggable vulnerabilities in AML while reporting that they are dispensable for normal hematopoiesis.
Why it matters: A therapeutic dependency with limited apparent normal-hematopoiesis requirement could offer a useful selectivity window.
Why for Yiru: It aligns with your interests in RNA biology, cancer dependencies and pharmacobiology.
Modeling maternal immune activation in 3D ex vivo human fetal brain cerebroids reveals IL-17A-driven disruption of cortical development
www.nature.com Published 2026-08-11 Research article DOI: 10.1038/s41593-026-02400-2
maternal immune activation fetal brain IL-17A cortical development
Summary: Using 3D ex vivo human fetal brain cerebroids, the study models maternal immune activation and reports IL-17A-driven disruption of cortical development.
Why it matters: The model provides a human ex vivo context for studying immune-mediated effects on developing cortical tissue.
Why for Yiru: This complements your computational immunology interests with a tractable human developmental model.
Cross-disciplinary watchlist
Other Fields
megaMine: a scalable, rule-based framework for mining gene-cancer-drug evidence from biomedical literature
bioRxiv Published 2026-08-12 Preprint DOI: 10.64898/2026.08.06.743392v1
literature mining gene-drug evidence biomedical NLP knowledge bases
Summary: megaMine is a scalable, rule-based framework for mining gene-cancer-drug evidence from biomedical literature.
Why it matters: Structured evidence mining can make large biomedical literatures more searchable and auditable.
Why for Yiru: This is especially relevant to your interest in agent-native biomedical research infrastructure.
Virtual-cell verification enables self-auditing AI discovery for immune rejuvenation
bioRxiv Published 2026-08-11 Preprint DOI: 10.64898/2026.08.04.742916v1
virtual cells AI discovery immune rejuvenation verification
Summary: The preprint describes virtual-cell verification as a way to support self-auditing AI discovery for immune rejuvenation.
Why it matters: Verification inside an AI discovery loop is important for distinguishing plausible computational hypotheses from unsupported predictions.
Why for Yiru: It is directly aligned with your digital-twin AI and self-auditing biomedical-agent interests.
Moirai: single-cell trajectory inference grounded in gene-level expression dynamics
bioRxiv Published 2026-08-11 Preprint DOI: 10.64898/2026.08.05.742709v1
single-cell trajectories gene dynamics cell-state transitions
Summary: Moirai presents single-cell trajectory inference grounded in gene-level expression dynamics.
Why it matters: Gene-level dynamics may provide a more mechanistic basis for ordering and interpreting cell-state transitions.
Why for Yiru: This matches your interest in trajectory-aware single-cell analysis infrastructure.
Integrated coding-noncoding genome annotation expands single-cell transcriptomic discovery and identifies clinically relevant noncoding RNAs in multiple myeloma
bioRxiv Published 2026-08-11 Preprint DOI: 10.64898/2026.08.09.743753v1
single-cell transcriptomics noncoding RNA genome annotation multiple myeloma
Summary: The study integrates coding and noncoding genome annotation to expand single-cell transcriptomic discovery in multiple myeloma.
Why it matters: Including noncoding annotations can broaden the biological signals recovered from single-cell datasets.
Why for Yiru: This fits your interests in transcriptomic analysis, regulatory biology and clinically relevant cancer atlases.
Spatial and Multi-Omics Analysis of Human Breast Cancer Reveals the Spatiotemporal Dynamics of Basal Layer Disruption
bioRxiv Published 2026-08-11 Preprint DOI: 10.64898/2026.08.10.744069v1
breast cancer spatial omics multi-omics tissue architecture
Summary: The preprint uses spatial and multi-omics analysis to examine spatiotemporal dynamics of basal-layer disruption in human breast cancer.
Why it matters: Basal-layer disruption is examined as a spatially organized process rather than only a bulk molecular signal.
Why for Yiru: This is a strong match for your spatial-omics and tumor-architecture interests.
Molecular basis of HACD-TECR complex mediated very-long-chain fatty acid elongation reveal a potential target in colorectal cancer
www.science.org Published 2026-08-12 Research article DOI: 10.1126/sciadv.aeh5593
fatty-acid elongation HACD-TECR colorectal cancer metabolism
Summary: The article examines the molecular basis of HACD-TECR complex-mediated very-long-chain fatty-acid elongation and identifies a potential colorectal-cancer target.
Why it matters: It links lipid metabolic machinery to a possible cancer vulnerability.
Why for Yiru: This connects your drug-related omics interests with cancer metabolism and target discovery.