光学混合CRISPR筛选:结合多重gRNA检测与单细胞空间多组学

Optical pooled screening (OPS) enables high-throughput measurement of cellular responses to genetic perturbations directly in situ. However, most sequencing-by-synthesis (SBS)-based OPS approaches remain limited to single-gene perturbation readouts or narrow phenotypic panels, restricting their ability to resolve the full complexity of perturbation-driven cellular states. In contrast, multiplexed in situ hybridization-based platforms offer an orthogonal detection strategy capable of simultaneously measuring multiple gRNAs and rich multiomic phenotypes within the same cell. We establish a framework for large-scale pooled CRISPR screening using the CosMxⓇ Spatial Molecular Imager (SMI), enabling spatially resolved, single-cell whole-transcriptome profiling integrated with multiplexed gRNA detection. We evaluated diverse vector architectures expressing barcoded gRNAs to minimize guide-barcode decoupling, a major obstacle in pooled optical screens. Using a reporter-based functional assay, we identified an optimized vector design that preserves Cas9 activity while enabling sensitive, accurate, high-fidelity detection of barcoded gRNAs in situ. We further demonstrate seamless coupling of gRNA detection with high-dimensional RNA and protein profiling at subcellular resolution within the standard CosMx workflow. This provides a practical and scalable path to pooled optical combinatorial CRISPR screening with omics-level phenotypes. By leveraging whole-transcriptome spatial readouts, this approach captures global gene expression changes, emergent cell states, microenvironment-dependent phenotypes, and spatially coordinated transcriptional responses not detected using targeted phenotyping or dissociated single-cell methods. Importantly, spatially resolved whole-transcriptome OPS enables knockouts to be interpreted in their native architectural context, revealing how genetic perturbations influence cell-cell communication, signaling cascades, neighborhood formation, and ligand-receptor interactions. These capabilities position CosMx-enabled pooled CRISPR OPS as a transformative platform for cancer research. It allows investigators to map causal genetic mechanisms within complex in vitro and ex vivo models, uncover cryptic phenotypes that manifest only in spatially organized environments, and identify therapeutic targets invisible to conventional pooled screens. Together, these results establish a strong foundation for next-generation multimodal spatial CRISPR screening. The ability to perform pooled perturbations with unbiased, spatially resolved whole-transcriptome readouts will accelerate discovery of cancer-relevant pathways, support mechanism-of-action and resistance studies, and enable development of spatially aware functional genomics atlases and AI-driven predictive models of tumor biology.