Spatially Resolved in vivo CRISPR Screen Sequencing via Perturb-DBiT
Alev BaysoyiD, Xiaolong TianiD, Feifei ZhangiD, Paul A. RenaueriD, Zhiliang BaiiD, Hao ShiiD, Haikuo LiiD, Bo Tao, Mingyu YangiD, Archibald EnninfuliD, Fu Qiang Gao, Guangchuan WangiD, Wanqiu ZhangiD, Thao Tran, Nathan Heath PattersoniD, Shuozhen BaoiD, Chuanpeng DongiD, Xin ShaniD, Mei ZhongiD, Sherri L. Rankin, Cliff S. GuyiD, Yan WangiD, Jon P. ConnellyiD, Shondra M. Pruett‐MilleriD, Hongbo ChiiD, , Rong FaniD
Abstract
Perturb-seq enabled the profiling of transcriptional effects of genetic perturbations in single cells but lacks the ability to examine the impact on tissue environments. We present Perturb-DBiT for simultaneous co- sequencing of spatial transcriptome and guide RNAs (gRNAs) on the same tissue section for in vivo CRISPR screen with genome-scale gRNA libraries, offering a comprehensive understanding of how genetic modifications affect cellular behavior and tissue architecture. This platform supports a variety of delivery vectors, gRNA library sizes, and tissue preparations, along with two distinct gRNA capture methods, making it adaptable to a wide range of experimental setups. In applying Perturb-DBiT, we conducted un-biased knockouts of tens of genes or at genome-wide scale across three cancer models. We mapped all gRNAs in individual colonies and corresponding transcriptomes in a human cancer metastatic colonization model, revealing clonal dynamics and cooperation. We also examined the effect of genetic perturbation on the tumor immune microenvironment in an immune-competent syngeneic model, uncovering differential and synergistic perturbations in promoting immune infiltration or suppression in tumors. Perturb-DBiT allows for simultaneously evaluating the impact of each knockout on tumor initiation, development, metastasis, histopathology, and immune landscape. Ultimately, it not only broadens the scope of genetic inquiry, but also lays the groundwork for developing targeted therapeutic strategies.
02 /Specification & Classification
Click any tag to browse matching papers- Citations
- 45
- Organism
- Assay
- Research Focus
- Approach
In-depth summary
Claude Sonnet 4.6 · max effort · 8 Nov 2026 · from the full text of version 1 (19 Nov 2024) · CC BY 4.0
Perturb-DBiT co-sequences spatial transcriptomes and gRNAs on one tissue section, supporting genome-wide in vivo CRISPR screens in FFPE and fresh-frozen tumors across three cancer models.
Pooled single-cell CRISPR screens lose tissue context, and earlier spatial perturbation methods did not jointly map perturbations and transcriptome on the same section, or were restricted to small predefined libraries. Perturb-DBiT offers unbiased gRNA and mRNA co-profiling that is compatible with genome-scale libraries.
Key results
ShownsgRNA capture was efficient in small and medium libraries, and the liver GEMM data recovered every sgRNA that the earlier pooled molecular inversion probe (MIP) analysis had found enriched.
B16-OVA: 78 of 86 sgRNAs, detected in 90% of tissue-bearing pixels. Liver GEMM (DC): 263 of 288 sgRNAs; 86 sgRNAs overlapped MIP data, 100% of MIP-enriched sgRNAs recapitulated, roughly 3-fold more sgRNAs than MIP in a 1.25x1.25 mm ROI. · Perturb-DBiT ensures precise sgRNA fidelity and…
SuggestedOn an adjacent section of the same treated liver block, PAC co-profiling detected more sgRNA diversity than pooled MIP sequencing, with moderate agreement on which sgRNAs were truly present.
PAC captured 43% of total sgRNAs vs 30% for MIP, and 95% of the 57 mTSG target genes vs 80% for MIP. ROC AUC was 0.724 (MIP detection >25% across lobes as predictor). · Perturb-DBiT enhanced sgRNA capture and…; Comparative analysis of pooled and perturb-DBiT…
ShownIn a genome-wide Brunello screen in HT29 lung colonies, most top sgRNAs lay within tumor colonies and some formed colocalized pairs. Intronic reads, likely from polyadenylation, raised per-pixel gene and UMI yield.
ROI had 235 unique sgRNAs and 5,740 sgRNA UMIs. Pairs S100A4/SLC22A16, SEMA6B/FGL2, MT1E/KCTD16. Exonic reads gave 3,371 genes and 6,429 UMIs per pixel; with intronic reads, 7,903 genes and 21,821 UMIs per pixel. · Perturb-DBiT reveals clonality within HT29 lung…
ShownIn vitro clonogenic assays supported tumor suppressor roles for four top HT29 hits. Pixel-level differential expression for those sgRNAs showed down-regulated growth-inhibiting genes.
HT29-Cas9 cells (5,000 cells seeded, colonies scored after 14 days, vs NTC and AAVS1 controls) supported SEMA6B, MT1E, S100A4 and ASAH2. Pixel DEG thresholds were Padj<0.05 with Log2FoldChange<-1 or >1. · Perturb-DBiT reveals clonality within HT29 lung…; Clonogenic assay of HT29 cells in vitro
SuggestedIn immune-competent E0771 lung tumors, spatial sgRNA enrichment separated two neighboring tumor clusters. Zc3h12a sgRNA was enriched in C1 near CD8+ T cells, while C2 had more myeloid cells and higher PD-L1.
91 unique sgRNAs and 17,548 sgRNA UMIs; 1,995 genes and 4,458 UMIs per pixel. A 27-plex CODEX panel on an adjacent section showed PD-L1 nearly undetectable in C1 but significantly elevated in C2. · Perturb-DBiT identifies genes modulating the…
Methods, data and reproducibility
Two microfluidic chips apply orthogonal spatial barcodes to tissue, producing paired mRNA and sgRNA libraries via polyadenylated capture (PAC) or direct capture (DC). sgRNA calls are integrated with Seurat/Mixscape, Monocle3, iStar, histology and CODEX.
- B16-OVA melanoma tumors with an 86-sgRNA dual-guide retroviral library (38 epigenetic genes plus 10 NTCs), PAC profiling · mouse
- Spleen sections from CD4+ SMARTA T cell transfer with a 2-NTC sgRNA library, DC profiling, 6 days after LCMV infection · mouse
- AAV-CRISPR liver GEMM with the 288-sgRNA mTSG library, PD-1 treated; DC and PAC on adjacent sections · mouse · n = one treated liver block (adjacent sections)
- HT29 lung metastatic colonization with Brunello library (77,441 sgRNAs targeting 19,114 genes), 35 days; 50-micron and 20-micron devices · human cells in NSG mice
- E0771 syngeneic lung colonization with Brie library (78,637 sgRNAs, 19,674 genes), 21 days · mouse
- CODEX 27-plex immunophenotyping on a section adjacent to the E0771 Perturb-DBiT section · mouse
- Methods & tools
- Perturb-DBiT PAC and DC, Seurat v5.0.1 (SCTransform, Mixscape), STAR V2.7.7a, ST_Pipeline V1.7.6, Cutadapt V3.4 and Bowtie v1.3.0, Monocle 3 v1.3.1, iStar, Connectome V1.0.0, pROC v1.18.0, fgsea v1.22.0, StarDist in QuPath, Weave platform (Aspect Analytics)
- Reagents & environment
- Versions stated for Cutadapt V3.4, STAR V2.7.7a, ST_Pipeline V1.7.6, Seurat v5.0.1, Monocle 3 v1.3.1, Connectome V1.0.0, pROC v1.18.0, fgsea v1.22.0, Presto v1.0.0, ComplexHeatmap v2.12.1. No container is described.
- Reproducibility gaps
- GEO accession pending. Code is said to be on the authors' GitHub, but no URL is printed. Other reanalysis information is available from the lead contact on request.
Limitations (4)
- The method does not reach true single-cell resolution. Near single-cell pixels may obscure spatial heterogeneity in sgRNA distribution and the exact cell populations where sgRNA activity occurs. · stated by the authors
- The specific genes detected can vary between samples, even adjacent sections. The authors attribute this to clonal heterogeneity, microenvironmental influences or stochastic CRISPR editing. · stated by the authors
- Genome-wide analyses rest on a few sections with sparse sgRNA coverage (235 sgRNAs in HT29, 91 in E0771). Perturbation effects are characterized mainly for top hits, and E0771 has no replicate sections reported. · noted in this summary
- Immune and stemness interpretations of E0771 sgRNA patterns (Zc3h12a, Cotl1, Neu4, Aldh1a1) rest on prior literature and one CODEX section. No functional in vivo validation of these hits is reported. · noted in this summary
Who should read:Read this if you want to jointly map pooled CRISPR perturbations and transcriptomes in tissue sections, including genome-wide libraries in FFPE or frozen tumors.
Reproducibility ledger
Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.
Sign in to file a reproducibility report.
Related papers
- Spatial Perturb-Seq: Single-Cell Functional Genomics within Intact Tissue Architecture
Kimberle Shen, Wan Yi Seow, … Wei Leong Chew·Nature CommunicationsPeer-reviewed·Feb 21, 2026
- In vivo perturb-seq of cancer and microenvironment cells dissects oncologic drivers and radiotherapy responses in glioblastoma
Siyuan John Liu, Christopher Zou, … Luke A. Gilbert·Genome biologyPeer-reviewed·Oct 7, 2024
- Direct capture of CRISPR guides enables scalable, multiplexed, and multi-omic Perturb-seq
Joseph M. Replogle, Thomas M. Norman, … Britt Adamson·Nature BiotechnologyPeer-reviewed·Mar 30, 2020
- Spatially resolved, multimodal in vivo Perturb-seq using antibody-based cell hashing
Alexander A. Nevue, George C. Hartoularos, … David Lara‐Astiaso·bioRxiv·May 26, 2026
- Perturb-map enables CRISPR genomics with spatial resolution and identifies regulators of tumor immune composition
M. Dhainaut, S. A. Rose, … B. D. Brown·bioRxiv·Jul 13, 2021
- SPACE: multimodal spatial CRISPR screening with whole-transcriptome readout at subcellular resolution in 3D models
Mengwei Hu, Yi Cui, … Vanessa M. Peterson·bioRxiv·Sep 14, 2025