Skip to content
中文EN

Methods & resources

New techniques, tools, datasets and atlases — and methodological finds that discount prior data.

184 picksRelatedSingle-cell & spatial omicsAI drug discoveryPrimary research

Latest picks

41–60 of 184
Oct 10Sat
  1. bioRxiv62

    Systematic mapping of enhancers, silencers, and topological elements with prime editing deletion screens

    Swap-seq, a method combining pooled twin prime editing with flow-sorting, deletes hundreds of endogenous regulatory sequences and measures their quantitative effects on nearby gene expression. Applied to PPIF regulation in THP-1 monocytes, it quantifies distal and intronic enhancers, characterizes CTCF insulators, and uncovers a silencer that represses neighboring enhancers through HDAC-mediated mechanisms, with hundreds of genomic elements sharing its signatures. The resulting data show sequence-to-function models capture local element activity but fail to link elements to distal target genes.

    Why it matters: It delivers a generalizable deletion-screening tool and a benchmark showing sequence-to-function models miss distal gene targeting.

Oct 9Fri
  1. bioRxiv66

    Random mtDNA mutagenesis maps genotype-phenotype associations across metabolic contexts and lineages

    mtDIVE, a random C→T mutagenesis platform, enables pooled mitochondrial forward genetics in human epithelial cells and zebrafish embryos, generating molecules from single substitutions to more than 100 mutations. Selection on mutant genomes and their transcriptional consequences depend on metabolic state and developmental lineage: obligate OXPHOS eliminates highly mutated cells by apoptosis, while glycolytic conditions promote their competitive displacement. Individual loci have modest, non-redundant effects that predict cell state beyond mutation burden, and cancer-recurrent variants are enriched among mutations linked to reduced mitochondrial transcription.

    Why it matters: It replaces one-allele-at-a-time mtDNA models with pooled forward genetics separating mutation identity from load.

  2. Cell71

    Mapping transcriptional responses to cellular perturbations with RNA fingerprinting

    The authors propose RNA fingerprinting, a statistical framework that infers candidate causal regulators by mapping the transcriptional responses of single-cell experiments onto reference perturbation dictionaries.

    Why it matters: Readers can apply this statistical framework to their own single-cell perturbation data to infer candidate causal regulators from transcriptional responses.

  3. bioRxiv60

    STEER-seq: a recombinase based platform for pooled multimodal single-cell perturbation screening

    STEER-seq is a virus-free platform for pooled ORF gain-of-function screening in human induced pluripotent stem cells, using Bxb1-mediated integration at a defined safe-harbour locus so each perturbation is expressed from a single copy in a common genomic context. Joint RNA and chromatin accessibility profiling linked transcription factor responses to chromatin remodelling. In a haematopoietic screen, SPI1 alone drove differentiation to HSC/MPP-like progenitors without exogenous cytokines, and scaling to around 2,000 human coding sequences uncovered factors inducing neural progenitor-, melanocyte- and vascular smooth muscle-like states.

    Why it matters: It removes random viral integration as a confound in pooled gain-of-function screens and works in hiPSCs at ~2,000-ORF scale.

  4. WeChat81

    DNA Typewriter Reconstructs Mouse Cell Lineage Tree from Zygote to E13.5

    A DNA Typewriter recorded cell lineage in a living mouse, and from a single embryo the team reconstructed a phylogenetic tree rooted at the zygote, time-calibrated to E13.5, containing 1.28 million cells annotated with transcriptomes. The work, by Chengxiang Qiu with Jay Shendure's group at the University of Washington, published in Science on October 9, 2026, closes a 50-year gap since Sulston manually traced a nematode's cell divisions. The tree reveals clonal dominance during gastrulation, hierarchical structure coupled to cell type, and ancestral state paths.

    Why it matters: First whole cell lineage tree of a mammal from zygote to organogenesis, closing a 50-year gap in developmental biology.

Oct 8Thu
  1. Cell69

    A fluctuation test for gastruloid heterogeneity

    Gastruloid variability is largely inherited, not induced, according to a fluctuation test adapted to gastruloids grown from single stem cells. The variability traces back to fluctuations in the state of the founding cell. This reframes a longstanding obstacle for organoids grown under identical conditions as a problem of founding-cell state rather than culture conditions.

    Why it matters: Identifies founding-cell state fluctuations as the source of organoid heterogeneity, a known barrier to organoid use.

  2. WeChat67

    AI Agents Build scBaseCount, the Largest Standardized Single-Cell Database with Over 502 Million Cells

    scBaseCount, published in Cell by researchers from the Arc Institute, the University of California, Stanford University and others, is the largest public single-cell RNA sequencing database built to date. AI agents mined all public 10x Genomics datasets in the Sequence Read Archive, automatically annotated metadata and reprocessed counts under a uniform standard, covering 27 species, 75 tissues and more than 502 million cells. The resource supplies high-quality, large-scale data for training AI models of cell state and for cross-study integration.

    Why it matters: It turns the unstructured Sequence Read Archive into a standardized resource large enough for AI cell-state models.

Oct 7Wed
  1. Nature Biotechnology63

    Decoding systemic vascular health and hypertensive disorders in pregnancy through retinal imaging and Visionary AI.

    Visionary AI, a platform combining retinal imaging with graph-based vascular modeling, predicted hypertensive disorders of pregnancy before clinical onset. In a prospective development cohort of 1,267 pregnancies it reached AUC 0.91 and average precision 0.81 for preeclampsia, and a stability-optimized model transferred without retraining to external validation at AUC 0.81 and AP 0.68, outperforming clinical and deep learning benchmarks. The finding positions the maternal retina as a minimally invasive biosensor for first-trimester risk stratification.

    Why it matters: First-trimester preeclampsia prediction from retinal images outperformed clinical and deep learning benchmarks in external validation.

  2. Nature Biotechnology68

    Highly multiplexed mammalian metabolic engineering with a shotgun approach.

    Shotgun genetic engineering (SGE) delivers many barcoded small constructs into mammalian cells, so each cell tests a synthetic pathway varying gene content, stoichiometry and organellar localization, with functional pathways found by sequencing barcodes from phenotypic cells. Screening millions of pathway combinations enabled near-wild-type growth without valine and isoleucine-free growth in Chinese hamster ovary cells, plus valine-free growth in Jurkat cells. Functional pathways favored mitochondrial localization and involved tens of kilobases of synthetic DNA, and the datasets support machine-learning-guided engineering of complex biosynthetic traits.

    Why it matters: It makes unbiased combinatorial pathway screening in mammalian cells tractable at a scale conventional methods cannot reach.

  3. Cell81

    SPARCS enables scalable recovery of complex image-based phenotypes for genetic screening.

    Spatially resolved CRISPR screening (SPARCS) brings forward genetic screening to complex image-based phenotypes by isolating mutants in situ through automated laser microdissection, coupled with machine learning image analysis and proteomics. Applied to genome-wide CRISPR knockout screens of autophagosome formation and STING activation across 70 million cells, it recovered most macroautophagy genes and identified GPHR as a pH-dependent regulator of STING. Image-based genome-scale screening with proteome-level readout is now practical.

    Why it matters: It makes genome-wide screening feasible for complex image-based phenotypes, a scale and readout previously out of reach.

Oct 6Tue
  1. Cell70

    SPARCS enables scalable recovery of complex image-based phenotypes for genetic screening

    SPARCS automates laser microdissection to run pooled forward genetic screens that scale to large mutant libraries while measuring complex image-based phenotypes. It physically recovers individual mutants, so hits can be profiled beyond genotyping, including by mass spectrometry-based proteomics. This removes the trade-off between library scale and information-rich phenotype measurement in genetic screening.

    Why it matters: It removes a core trade-off in genetic screens by adding scalable, multimodal profiling of recovered mutants.

  2. Nature Biotechnology71

    Detectrons convert transient RNA sequences into stable DNA barcodes for high-throughput analysis of RNA-dependent processes.

    Detectrons couple programmable toehold switches with retron-mediated reverse transcription to convert transient RNA sequences into stable DNA barcodes in living cells. Built as a synthetic toehold retron library and refined with machine learning, the sensors enable multiplexed live-cell detection of specific phage infections and quantitative host susceptibility profiling in pooled bacterial populations. The work turns RNA sensing into a sequencing-readable, durable record.

    Why it matters: It converts RNA sensing into DNA barcodes readable by sequencing, enabling multiplexed phage screening in live cells.

  3. Nature74

    Highly multiplexed mammalian metabolic engineering with a shotgun approach

    Shotgun genetic engineering (SGE) delivers barcoded individual transcription units at high multiplicity of infection so each mammalian cell carries a unique pathway combination, letting a far larger design space be sampled in parallel than serial design-build-test cycles allow. After selection, sequencing of cells with the desired phenotype reveals which combinations of promoters, localization signals, enzymes, stoichiometry and copy number conferred the function. The method also generates large-scale synthetic datasets for training predictive machine learning models of pathway design.

    Why it matters: It addresses the slow iteration that has kept pathway-scale metabolic engineering rare in mammalian cells.

  4. Nature65

    Decoding systemic vascular health and hypertensive disorders in pregnancy through retinal imaging and Visionary AI

    Hypertensive disorders of pregnancy affect 2–15% of pregnancies and are usually diagnosed after 20 weeks of gestation, once hypertension, proteinuria or organ dysfunction appear, even though low-dose aspirin started before 16 weeks can reduce risk of preterm delivery with preeclampsia by up to 62%. The paper develops retinal imaging with AI to detect the systemic and placental vascular changes that precede symptoms, building on evidence that women with preeclampsia show retinal arteriolar narrowing, increased tortuosity and reduced capillary density before diagnosis.

    Why it matters: It tests a noninvasive retinal imaging route to early preeclampsia risk identification, which current late diagnostics fail to provide.

  5. U.S. FDA64

    M15 Resource Center

    The FDA has posted a resource center for ICH M15, a guidance on general principles for model-informed drug development (MIDD). The guidance describes MIDD principles, including planning, model evaluation and evidence documentation, sets a harmonized framework for assessing MIDD-derived evidence, and gives recommendations on MIDD-related regulatory interactions, reporting and submission. FDA recommends sponsors use it when preparing meeting requests, background packages and regulatory submissions.

    Why it matters: FDA now expects ICH M15 principles to shape MIDD evidence in submissions and meetings.

  6. YouTube69

    Nucleosome Crystal Structure Breakthrough Hinged on Sample Homogeneity and 37°C Heating

    In an iBiology lecture, Karolin Luger recounted how her team solved the high-resolution X-ray crystal structure of the nucleosome. The keys were homogenizing the sample and heating it at 37°C for 30 minutes to convert heterogeneous particles with off-center DNA to centered positioning.

    Why it matters: Packages refolding, thermodynamic equilibrium and synchrotron radiation into a sample preparation route for high-resolution nucleosome crystallography, transferable to other heterogeneous samples.

  7. YouTube69

    ESM Team Describes Training ESMC and Opening Protein Language Models with Sparse Autoencoders

    A Biohub webinar by the ESM team explains the motivation for protein language models and the training of ESMC, which masks single amino acids and predicts them from context with stacked transformer layers. The talk highlights sparse autoencoders from NLP as a way to extract monosemantic features from dense embeddings, with 16000 features chosen to balance specificity against cross-family generalization, enabling explicit clustering of protein space distinct from standard sequence or structure clustering.

    Why it matters: Explains how interpretability tools from NLP are being applied to protein foundation models.

  8. YouTube70

    NN-UNet Author Fabian Isensee Demonstrates nnInteractive Promptable 3D Segmentation at Janelia

    NN-UNet author Fabian Isensee presented nnInteractive at Janelia, a vision-transformer-based promptable model that turns 2D prompts (points, boxes, lasso) into instant 3D segmentation masks. He compared it with general SAM foundation models, which still beat organ-specific checkpoints on some medical tasks, and demonstrated the tool on brain tumor MRI, electron microscopy neuron stacks, and insect brain micro-CT. Automated masks serve as initialization, with user interaction correcting model errors.

    Why it matters: NN-UNet's author shows how interactive prompting fixes 3D biomedical segmentation, directly relevant to image-analysis workflows.

Oct 5Mon
  1. 顶刊论文（社媒热度）68

    Once yearly cell-based therapy for sustained and dose-tunable delivery of monoclonal antibodies.

    An encapsulated cell-based "biologics factory" using an immunomodulatory alginate formulation sustained stable serum titers of the HIV-neutralizing mAb 3BNC117 for 1 year in immunocompetent mice, supported by an anti-inflammatory, proresolving immune niche that attenuates fibrosis. The platform produced 13 diverse mAbs from an allogeneic cell chassis, and in a nonhuman primate subcutaneous implant maintained stable ipilimumab titers for more than 6 months with no detectable toxicity, antidrug antibodies, or adverse events across a three-dose escalation. A retrievable macrodevice allows on-demand termination and dose-proportional tuning, pointing toward replacing frequent injections with single-administration therapy.

    Why it matters: It shows a single implant can deliver monoclonal antibodies for months to a year, replacing repeated injections.

CBACBA Frontier Insights (CFI)Chinese Biopharmaceutical Association (CBA)www.cba-usa.org