Cell Reports Methods最新文献

We develop the spatial dissimilarity method to uncover complex bivariate relationships in single-cell and spatial transcriptomics data, addressing challenges such as alternative splicing and allele-specific gene expression. Applying this method to detect alternative splicing in neurons demonstrates improved accuracy and sensitivity compared to existing tools, notably identifying neuron subtypes. In tumor cells, spatial dissimilarity analysis reveals somatic variants that emerge during tumor progression, validated through whole-exome sequencing. These findings highlight how allele-specific genetic variants contribute to the subclone architecture of cancer cells, offering insights into cellular heterogeneity. Applied on a human cell atlas, we uncover numerous cases of allele-specific expression of genes in normal cells. We provide a software package for spatial dissimilarity analysis to enable enhanced understanding of cellular complexity and gene expression dynamics under homeostatic conditions and during states of transitions.

Existing microscopy approaches are often unable to identify and contextualize rare but biologically meaningful events due to limitations associated with simultaneously achieving both high-resolution imaging and a cm-scale field of view. Here, we present multiscale cleared tissue axially swept light-sheet microscopy (MCT-ASLM), a platform combining cm-scale imaging with targeted high-resolution interrogation of intact tissues in human-guided or autonomous modes. Capable of capturing fields of view up to 21 mm at micron-scale resolution, MCT-ASLM can seamlessly transition to a targeted imaging mode with an isotropic resolution that approaches ∼300 nm. This versatility enables detailed studies of hierarchical organization and spatially complex processes, including mapping neuronal circuits in rat brains, visualizing glomerular innervation in mouse kidneys, and examining metastatic tumor microenvironments. By bridging subcellular- to tissue-level scales, MCT-ASLM offers a powerful method for unraveling how local events contribute to global biological phenomena.

We evaluated whether predicted continuous disease representations could enhance genetic discovery beyond case-control genome-wide association study (GWAS) phenotypes across eight complex diseases in up to 485,448 UK Biobank participants. Predicted phenotypes had high genetic correlations with case-control phenotypes (median r_g = 0.66) but identified more independent associations (median 306 versus 125). While some predicted phenotype associations were spurious, multi-trait analysis of GWAS-boosted case-control phenotypes identified a median of 46 additional variants per disease, of which a median of 73% replicated in FinnGen, 37% reached genome-wide significance in a UK Biobank/FinnGen meta-analysis, and 45% had supporting evidence. Predicted phenotypes also identified 14 genes targeted by phase I-IV drugs not identified by case-control phenotypes, and combined polygenic risk scores (PRSs) using both phenotypes improved prediction performance, with a median 37% increase in Nagelkerke's R². Predicted phenotypes represent composite biomarkers complementing case-control approaches in genetic discovery, drug target prioritization, and risk prediction, though efficacy varies across diseases.

A reliable, efficient, high-throughput pipeline to evaluate viral protease inhibitors would enhance antiviral drug discovery. Methods such as crystallography and phenotypic screening are often constrained by complex assay conditions, limited physiological relevance, or live virus handling safety concerns. Proof-of-concept studies previously demonstrated synthetic gene circuits that produce a quantitative reporter upon protease inhibition, enabling functional virus-independent evaluation of viral protease inhibitors in live cells. Using the SARS-CoV-2 3-chymotrypsin-like protease (3CLpro) as a model, we advanced this approach into a high-throughput first-pass qualitative assay ("hit/no-hit") to rapidly identify promising drug candidates. Our optimized circuit design was used to produce stable HEK293T and HeLa designer cells that generate two distinct fluorescence outputs, simultaneously reporting protease inhibition and cytotoxicity. The screening pipeline is designed to minimize labor, costs, and false-positive observations, thus enabling versatile, safe, and efficient functional drug screening suitable for any conventional biological laboratory.

We identified a gene panel comprising 71 glycosyltransferases (GTs) that alter glycan patterns on cancer cells as they become more virulent. When these cancer-pattern GTs (CPGTs) were run through an algorithm trained on The Cancer Genome Atlas, they differentiated tumors from healthy tissue with 97% accuracy and clustered 27 cancers with 94% accuracy in external validation, revealing each variety's "biometric glycan ID." Using machine learning, we built four models for cancer classification, including two for detecting the molecular subtypes of breast cancer and glioma using even smaller CPGT sets. Our results reveal the power of using glyco-genes for diagnostics: Our breast cancer classifier was almost twice as effective in independent testing as the widely used prediction analysis of microarray 50 (PAM50) subtyping kit at differentiating between luminal A, luminal B, HER2-enriched, and basal-like breast cancers based on a comparable number of genes. Only four GT genes were needed to build a prognostic model for glioma survival.

Microglia are crucial targets for therapeutic interventions in diseases like Alzheimer's and stroke, but efficient gene delivery to these immune cells is challenging. We developed an adeno-associated virus (AAV) vector that achieves specific and efficient gene delivery to microglia. This vector incorporates the mIba1 promoter, GFP, miRNA target sequences (miR.Ts), WPRE, and poly(A) signal. Positioning miR.Ts on both sides of WPRE significantly suppressed non-microglial expression, achieving over 90% specificity and more than 60% efficiency in microglia-specific gene expression 3 weeks post-administration. Additionally, this vector enabled GCaMP expression, facilitating real-time calcium dynamics monitoring in microglial processes. Using a blood-brain barrier-penetrant AAV-9P31 capsid variant, intravenous administration resulted in broad and selective microglial GFP expression across the brain. These results establish our AAV vector as a versatile tool for long-term, highly specific, and efficient gene expression in microglia, advancing microglial research and potential therapeutic applications.

We present a comprehensive gene fusion (GF) detection and analysis workflow that combines targeted panel-based and whole-transcriptome long-read sequencing. We first adapted libraries from the short-read CHOP Cancer Fusion Panel, which targets 119 oncogenes commonly implicated in cancer fusions, for use on Oxford Nanopore Technologies' long-read sequencing platform. Long-read sequencing successfully detected known GFs in panel-positive samples, confirming compatibility, and enabled reduced turnaround times. To expand GF discovery in clinically challenging cases, we analyzed 24 glioma samples with negative short-read fusion panel results using whole-transcriptome long-read sequencing. This identified 20 candidate GFs in panel-negative samples that were absent from current fusion databases, all of which were experimentally validated. In summary, we introduce a computational workflow that combines panel-based and whole-transcriptome long-read sequencing with tailored analysis pipelines to enable fast and comprehensive GF detection in cancer.

This paper reports a 3D-printed plug as a meso-scale interface solution that minimizes sample loss and enhances cell usage efficiency, seamlessly connecting microfluidic systems to conventional well plates. The plug concentrates cells near the region of interest for chemotaxis, reducing cell number requirements and featuring tapered structures for efficient manual or robotic liquid handling. Comprehensive testing showed that the plug increased cell usage efficiency in single-cell migration assays by 8-fold, maintaining accuracy and sensitivity. We also extended our approach to neuron axon guidance assays, where limited cell availability is a constraint, and observed substantial improvements in assay outcomes. This integration of 3D printing with microfluidics establishes low-loss interfaces for precious samples, advancing the capabilities of microfluidic technology.

The growing availability of public multi-domain medical image datasets enables training omnipotent image-to-image (I2I) translation models. However, integrating diverse protocols poses challenges in domain encoding and scalability. Therefore, we propose the "every domain all at once" I2I (EVA-I2I) translation model using DICOM-tag-informed contrastive language-image pre-training (DCLIP). DCLIP maps natural language scan descriptions into a common latent space, offering richer representations than traditional one-hot encoding. We develop the model using seven public datasets with 27,950 scans (3D volumes) for the brain, breast, abdomen, and pelvis. Experimental results show that our EVA-I2I can synthesize every seen domain at once with a single training session and achieve excellent image quality on different I2I translation tasks. Results for downstream applications (e.g., registration, classification, and segmentation) demonstrate that EVA-I2I can be directly applied to domain adaptation on external datasets without fine-tuning and that it also enables the potential for zero-shot domain adaptation for never-before-seen domains.

Novel insights into cardiac contractile dysfunction at the cellular level could deepen understanding of arrhythmia and heart injury, which are leading causes of morbidity and mortality worldwide. We present a comprehensive experimental and computational framework combining light-field microscopy and single-cell tracking to investigate real-time volumetric data in live zebrafish hearts, which share structural and electrical similarities to the human heart. Our system acquires 200 vol/s with lateral resolution of up to 5.02 ± 0.54 μm and axial resolution of 9.02 ± 1.11 μm across the whole depth using an expectation-maximization-smoothed deconvolution algorithm. We apply a deep-learning approach to quantify cell displacement and velocity in blood flow and myocardial motion and to perform real-time volumetric tracking from end-systole to end-diastole within a virtual reality environment. This capability delivers high-speed and high-resolution imaging of cardiac contractility at single-cell resolution over multiple cycles, supporting in-depth investigation of intercellular interactions in health and disease.