Tracing the evolution of single-cell 3D genomes in Kras-driven cancers

Abstract

Although three-dimensional (3D) genome structures are altered in cancer, it remains unclear how these changes evolve and diversify during cancer progression. Leveraging genome-wide chromatin tracing to visualize 3D genome folding directly in tissues, we generated 3D genome cancer atlases of oncogenic Kras-driven mouse lung adenocarcinoma (LUAD) and pancreatic ductal adenocarcinoma. Here we define nonmonotonic, stage-specific alterations in 3D genome compaction, heterogeneity and compartmentalization as cancers progress from normal to preinvasive and ultimately to invasive tumors, discovering a potential structural bottleneck in early tumor progression. Remarkably, 3D genome architectures distinguish morphologic cancer states in single cells, despite considerable cell-to-cell heterogeneity. Analyses of genome compartmentalization changes not only showed that compartment-associated genes are more homogeneously regulated but also elucidated prognostic and dependency genes in LUAD, as well as an unexpected role for Rnf2 in 3D genome regulation. Our results highlight the power of single-cell 3D genome mapping to identify diagnostic, prognostic and therapeutic biomarkers in cancer.

Fig. 1. A genome-scale chromatin tracing strategy to visualize cancer 3D genomes.

a, Schematic illustration of the experimental procedure. b, Raw FISH foci (left) in bit 28 in a WT AT2 cell. Zoom-in images (right) of raw FISH foci showing the decoding procedure and reconstructed genomic loci for five loci (shown as dots with five pseudocolors each appearing twice). Representative results from five biological replicates (n = 4 mice) are shown in b and c. c, Reconstructed chromatin traces (left) superimposed with DAPI staining. The traces are 2D projections of x and y coordinates. The DAPI image is a maximum-intensity z projection from a 10-μm z stack. The 3D positions (right) of all decoded genomic loci in a single-cell nucleus. Different pseudocolors represent different autosomes. d, Matrix of mean interloci distances between all genomic loci in AT2 cells. n = 4,806 AT2 cells. e, Immunofluorescence staining (top) of cell-type markers, DAPI staining and fluorescent protein imaging of lung tissue from a WT mouse, a K-MADM-Trp53 mouse with adenomas and a K-MADM-Trp53 mouse with LUAD. Representative results from five biological replicates in WT (n = 4 mice) and adenoma (n = 4 mice) and six biological replicates in LUAD (n = 5 mice) states are shown. Representative cells (bottom) of each state. The images are maximum-intensity z projections from 10-μm z stacks. Chr, chromosome. Source data

Fig. 2. Systematic changes of 3D genome compaction and heterogeneity during lung cancer progression.

a, Mean interloci distances (top) and COV of interloci distances (bottom) of mouse chr13 in WT AT2 (n = 4,806), Trp53^+/+ adenoma (AdenomaR, n = 791), Trp53^+/− adenoma (AdenomaY, n = 1,603), Trp53^−/− adenoma (AdenomaG, n = 1,941) and Trp53^−/− LUAD (n = 17,711) cells. Cell numbers are identical in a–j. b, Distributions of decompaction (mean interloci distance) scores of each autosome (n = 19 autosomes) in each cell state. P values from two-sided Wilcoxon signed-rank tests versus AT2 with matched genomic loci are displayed. c, Mean log₂(FC) of decompaction scores of each chromosome, comparing each cancer state to AT2. An asterisk indicates FDR < 0.05, two-sided Wilcoxon signed-rank test. d, Clustered heatmap of z scores (scaled by column) of cis-chromosomal mean interloci distances. Percentages of locus pairs with different decompaction change patterns are shown below. e, Chr19 territory size measured by whole chromosome paint. P values from one-sided Wilcoxon rank-sum tests are shown. f, Distributions of heterogeneity (COV of interloci distance) scores of each autosome (n = 19) in each cell state. P values from two-sided Wilcoxon signed-rank tests versus AT2 with matched genomic loci are displayed. g, Mean log₂(FC) of heterogeneity scores of each chromosome, comparing each cancer state to AT2. An asterisk indicates FDR < 0.05, two-sided Wilcoxon signed-rank test. h, Clustered heatmap of z scores (scaled by column) of COV of cis-chromosomal interloci distances. Percentages of locus pairs with different chromatin folding heterogeneity change patterns are shown below. i,j, Distributions of 3D genome heterogeneity scores of autosomes (n = 19 autosomes) for cell pairs from the same mice (AT2, n = 4 mice) or same tumors (12 adenoma tumors and 6 LUAD tumors, intramouse or intratumoral; i) or from different mice (intermouse or intertumoral; j). P values from two-sided Wilcoxon signed-rank tests versus AT2 with matched autosomes are shown. In b, e, f, i and j, horizontal lines of each box represent the 75th percentile, median and 25th percentile. Whiskers indicate nonoutlier maximum and nonoutlier minimum, with outlier thresholds at 1.5× interquartile range beyond box limits. FC, fold change. Source data

Fig. 3. Changes of cancer state-specific 3D genome features during lung cancer progression.

a, Demixing scores (s.d. of normalized mean interloci distances) of the active chrX (X_a, n = 95) and inactive chrX (X_i, n = 95) in human IMR90 cells show a reduction (increased intermixing) in X_i, as previously described using other analyses. P value from two-sided Levene’s test is shown. b, Distributions of demixing scores of each autosome (n = 19) in each cell state. n = 4,806, 791, 1,603, 1,941 and 17,711 for WT AT2, AdenomaR, AdenomaY, AdenomaG and LUAD cells, respectively. Cell numbers of each cell state are identical in b–i. c, log₂(FC) of the demixing score of each autosome, comparing each cancer state to WT AT2 cell state. An asterisk indicates FDR < 0.05, two-sided Levene’s test. d, Distribution of polarization indices of each autosome (n = 19) in each cell state. e, Distribution of interchromosomal distances in each cell state. f, log₂(FC) of mean interloci distances for AdenomaR, AdenomaY and AdenomaG relative to AT2 cells. Yellow lines highlight the boundaries of chromosomes. g, Distributions of radial scores of chr13, chr7 and chr19 in each cell state. h, Mean log₂(FC) of radial scores of each chromosome, comparing each cancer state to the AT2 cell state. An asterisk indicates FDR < 0.05, two-sided Wilcoxon signed-rank test. i, log₂(FC) of mean interloci distances for LUAD relative to AT2 cells. Yellow lines highlight the boundaries of chromosomes. j, Nuclear convex hull volume of WT AT2, AdenomaR, AdenomaY, AdenomaG and LUAD cells. P values of two-sided Wilcoxon rank-sum tests are shown. n = 3,410, 157, 689, 878 and 8,834 for WT AT2, AdenomaR, AdenomaG, AdenomaY and LUAD cells, respectively (cells with at least 10 traces). In b, d, e and g, P values of two-sided Wilcoxon signed-rank test comparing each cancer state to WT AT2 state with matched autosomes (b,d) or matched genomic loci (e,g) are displayed. In b, d, e, g and j, horizontal lines of each box represent the 75th percentile, median and 25th percentile. Whiskers indicate nonoutlier maximum and nonoutlier minimum, with outlier thresholds at 1.5× interquartile range beyond box limits. Source data

Fig. 4. The single-cell 3D genome distinguishes and encodes cancer progression states.

a, Cartoon illustration of scA/B score calculation (top-left). t-SNE (top-right), UMAP (bottom-left) and PaCMAP (bottom-right) plots of single-cell 3D genome conformations. n = 3,410, 157, 689, 878 and 8,834 for WT AT2, AdenomaR, AdenomaG, AdenomaY and LUAD cells, respectively. Cell numbers of each cell state are identical in a–c, g and h. b, Confusion matrix of supervised machine learning in mouse lung cells. The number in each matrix element represents the precision in each predicted state. c, ROC curves of the machine learning model in mouse lung cells. The AUC values are shown. d, PCA plot of single-cell 3D genome conformations. n = 1,103, 191 and 268 for normal duct, PanIN and PDAC cells, respectively. Cell numbers of each cell state are identical in d–f. e, Confusion matrix of supervised machine learning in mouse pancreas cells. The number in each matrix element represents the precision in each predicted state. f, ROC curves of the machine learning model in mouse pancreas cells. The AUC values are shown. g, PCA plot of single-cell 3D genome conformations of adenoma and LUAD cells (left). Leiden clustering separates Adenoma-like and LUAD-like clusters (right). h, Percentages of cells with adenoma-like or LUAD-like 3D genome conformations in g, in each of the AdenomaR/AdenomaY, AdenomaG and LUAD states. The adenoma-like or LUAD-like conformation state for each cell is assigned based on a Leiden clustering approach. P values from two-sided Fisher’s exact test are shown. t-SNE, t-distributed stochastic neighbor embedding; UMAP, uniform manifold approximation and projection; PaCMAP, pairwise controlled manifold approximation; ROC, receiver operating characteristic; AUC, area under the curve. Source data

Fig. 5. Single-cell 3D genomics nominates prognostic genes and genetic dependencies.

a, Normalized scA/B score changes of genes with decreased (n = 94 genes) or increased (n = 120 genes) expression from AdenomaG to LUAD. P values were from one-sided one-sample t tests. b, Heatmap of CPD gene expression comparing AdenomaG and LUAD. c,d, Gene expression homogeneity of LUAD cells in the K-MADM-Trp53 (c) and Kras^LSL-G12D/WT; Trp53^flox/flox (KP) (d) models comparing CPDs (expression up + scA/B up) with control genes (expression up + scA/B unchanged). P values were from two-sided Wilcoxon rank-sum tests. e, Decoded RNA MERFISH image (top) with RNA molecules pseudo-colored by gene identities in the SA6082inf LUAD cell line. Gene expression homogeneity (bottom) across single cells comparing CPDs with control genes. P values were from two-sided Wilcoxon rank-sum tests. f, Kaplan–Meier survival curves (top) comparing TCGA LUAD patients with gene expression profiles most or least (top versus bottom 20%) correlated with CPDs and controls (expression up + scA/B unchanged; expression up only). A total of 21 genes are included per panel. n = 96 tumors per group. P values were from two-sided log-rank tests. A multivariate Cox regression model (bottom) analyzing CPD and control signatures in predicting patient survival after controlling for clinical covariates across the TCGA LUAD dataset (n = 478). g, Dependency scores (DEMETER2) of LUAD cell lines (n = 57) comparing CPDs and controls (expression up + scA/B unchanged or expression up only). n = 19–20 genes per group with available RNAi screen data in Cancer Dependency Map. Lower score = more dependent. P values were from the Kruskal–Wallis test with Dunn’s post hoc test. h, Cell viability (mean ± s.d., normalized to NTC, n = 3 replicates per hairpin, 3 hairpins per gene) of arrayed RNAi screen targeting CPDs in KP (top) and SA6082inf (bottom) cells. NTC (green), positive controls (red) and CPDs with substantial phenotypes (≥2 hairpins with <80% of NTC cell count, purple) are indicated. In a, c–e and g, horizontal lines of each box represent the 75th percentile, median and 25th percentile. Whiskers in a and c–e indicate nonoutlier maximum and nonoutlier minimum, with outlier thresholds at 1.5× interquartile range beyond box limits. Whiskers in g indicate maximum and minimum. NTC, nontargeting control. Source data

Fig. 6. Rnf2 partially regulates 3D genome organization changes during the adenoma-to-LUAD transition.

a, Western blot analysis of Rnf2 protein levels following Rnf2 knockdown with three independent hairpins in the KP cell line. Control cells were constructed with an NTC shRNA sequence (shNTC). HSP90 is loading control. b, Cell viability (mean ± s.d., normalized to mean of shNTC) following knockdown with three Rnf2 shRNAs, n = 3 replicates per hairpin. P values from two-sided two-sample t tests are shown. c, scA/B score changes from shRnf2 to shNTC versus those from AdenomaG to LUAD show a substantial positive correlation, n = 473 target genomic regions. Spearman correlation coefficients and P values from two-sided t tests are shown. d, Rnf2 peak densities in A (n = 209 regions) and B (n = 264 regions) compartments in shNTC. The horizontal lines of each box from top to bottom represent the 75th percentile, median and 25th percentile. Whiskers extend to the nonoutlier maximum and nonoutlier minimum. Outliers are defined as values at least 1.5× interquartile range away from the top or bottom of the box. P value of two-sided Wilcoxon rank-sum test is shown. e, CUT&RUN read density heatmaps of Rnf2, H3K4me3, H3K27me3, RNA polymerase II with phosphorylated S5 modification, mono-ubiquitination of lysine 119 of histone H2A (H2AK119ub) and BMI1 in shNTC KP cells. Rnf2 peak regions (−5 kb, +5 kb) of all target genomic regions are shown and are categorized as active (H3K4me3⁺, H3K27me3⁻), bivalent (H3K4me3⁺, H3K27me3⁺), repressed (H3K4me3⁻, H3K27me3⁺) or other (H3K4me3⁻, H3K27me3⁻) based on chromatin marks. f, scA/B score changes from shRnf2 to shNTC versus those from AdenomaG to LUAD show a stronger or similarly positive correlation using target genomic regions with only active Rnf2 peaks, n = 113 target genomic regions. Spearman correlation coefficients and P values from two-sided t tests are shown. Source data

Extended Data Fig. 1. Genome-scale chromatin tracing visualizes 3D genome organization in vivo.

a, Pearson correlation coefficients of mean cis-chromosomal interloci distances between WT datasets. b, Mean cis-chromosomal interloci distances of WT dataset 3 versus mean cis-chromosomal interloci distances of WT dataset 2. The red line is a fitted linear regression line. The black line is the y = x line. c, Mean cis-chromosomal interloci spatial distance versus genomic distance for all pairs of genomic loci on each autosome in AT2 cells. Different pseudocolors represent different autosomes. n = 6,039 intrachromosomal interloci pairs in b–d. n = 4,806 WT AT2 cells in c and d. d, Power-law scaling of all 19 mouse autosomes (chr1–19) in WT mouse lung. e, Power-law scaling of chr19 in E14.5 mouse fetal liver. Data were re-analyzed from ref. . f, Power-law scaling of all 20 mouse chromosomes (chr1–19, chrX) in the mouse brain inhibitory neurons expressing Vip. Data were re-analyzed from ref. . g, Schematic illustration of the experimental procedure. The schematic in panel g was created with BioRender.com. h, Whole-section fluorescence images of WT mouse lung and K-MADM-Trp53 mouse lungs containing adenomas or LUAD. Results are representative of 5 biological replicates in WT (4 mice) and adenoma (4 mice) states, and 6 biological replicates in the LUAD (5 mice) state. i, Detection efficiency of each target genomic region along detected chromosome traces in each cancer state. j, Distribution of the trace length across individual chromosomes in cells of each cancer state. k, Distribution of detected loci number per cell. n = 26,852 cells combining the five cell states. The horizontal lines of the box from top to bottom represent the 75th percentile, the median and the 25th percentile. Whiskers extend to the nonoutlier maximum and nonoutlier minimum. Outliers are defined as values at least 1.5 times interquartile range away from the top or bottom of the box. l, A and B compartment score profiles of example autosomes measured by Hi-C and chromatin tracing in mouse LUAD. m, Fraction of genomic regions with concordant A and B compartment assignment between Hi-C and chromatin tracing. For d–f, lines are fitted power-law functions, and S is the scaling factor. Source data

Extended Data Fig. 2. Systematic changes of intrachromosomal compaction and chromatin conformation heterogeneity during lung cancer progression.

a,b, Mean interloci distances (a) and coefficients of variation (COV) of interloci distances (b) of all autosomes in WT AT2 (n = 4,806), Trp53^+/+ adenoma (AdenomaR, n = 791), Trp53^+/− adenoma (AdenomaY, n = 1,603), Trp53^−/− adenoma (AdenomaG, n = 1,941) and Trp53^−/− LUAD (n = 17,711) cells. Source data

Extended Data Fig. 3. Subtraction maps showing systematic differences of intrachromosomal compaction and chromatin conformation heterogeneity during lung cancer progression.

a,b, Subtraction maps of mean interloci distances (a) and coefficients of variation (COV) of interloci distances (b) in all autosomes of Trp53^+/+ adenoma (AdenomaR, n = 791), Trp53^+/− adenoma (AdenomaY, n = 1,603), Trp53^−/− adenoma (AdenomaG, n = 1,941) and Trp53^−/− LUAD (n = 17,711) cells minus those of WT AT2 (n = 4,806) cells. Source data

Extended Data Fig. 4. Cancer state-independent 3D genome features during lung cancer progression.

a–c, Normalized trans-chromosomal proximity frequency between genomic loci in AT2 (a), adenoma (b) and LUAD (c) cells. The proximity frequency between each pair of trans-chromosomal genomic regions were normalized to the mean proximity frequency of all loci pairs of the two corresponding chromosomes. A cutoff distance of 800 nm was used for defining proximity. The genomic regions were re-ordered so that A compartment loci were grouped separately from B compartment loci. d–f, Distribution of normalized trans-chromosomal proximity frequencies of pairs of A loci (A–A), pairs of B loci (B–B) and pairs of A and B loci (A–B) in AT2 (d), adenoma (e) and LUAD (f) cells. g–i, Normalized trans-chromosomal proximity frequencies of A–A, B–B and A–B loci pairs in AT2 (g), adenoma (h) and LUAD (i) cells. The horizontal lines of each box from top to bottom represent the 75th percentile, the median and the 25th percentile. Whiskers extend to the nonoutlier maximum and nonoutlier minimum. Outliers are defined as values at least 1.5 times interquartile range away from the top or bottom of the box. P values from two-sided Wilcoxon rank-sum tests are shown. j–l, The proximity frequency between each pair of cis-chromosomal genomic regions as a function of their genomic distances in AT2 (j), adenoma (k) and LUAD (l) cells. m, Radial scores versus A–B compartment scores of genomic loci in the five cell states. The lines are fitted linear regression lines. Pearson correlation coefficients (R) and P values from two-sided t-tests are shown. n = 4,806, 791, 1,603, 1,941 and 17,711 for WT AT2, AdenomaR, AdenomaY, AdenomaG and LUAD cells, respectively. Cell numbers are the same in all panels. Source data

Extended Data Fig. 5. Exome and RNA sequencing analyses of lung tumors from K-MADM-Trp53 mice.

a, Representative large GFP+ (green) tumor for whole-exome and RNA sequencing. Scale bar = 2.5 mm. Image is representative of 6 mice. b, Frequency of nonsynonymous single nucleotide variants (SNVs) per megabase (Mb) with variant allele fraction (VAF) > 5% in AdenomaG and LUAD. c,d, Heatmap (c) and distribution (d) of mean copy number of each autosome across AdenomaG and LUAD (n = 6 tumors per group). Black lines (d) represent the median. e, Schematic illustration of snRNA-seq in K-MADM-Trp53 mice. The schematic in panel e was created with BioRender.com. f,g, UMAP of single-cell gene expression profiles in K-MADM-Trp53 lung tumors (n = 6,300 cells). Different cell-type clusters were labeled with different colors (f) based on gene set enrichment patterns of adenoma (n = 1,631) and LUAD (n = 4,669)-specific genes (g). h, Distribution of coefficient of variation of mean inferred copy numbers of all genes on each autosome (n = 19 autosomes) in adenoma and LUAD. P value from two-sided Wilcoxon signed-rank test. Horizontal lines of each box represent the 75th percentile, median and 25th percentile. Whiskers indicate nonoutlier maximum and nonoutlier minimum, with outlier thresholds at 1.5 times interquartile range beyond box limits. i,j, Unsupervised hierarchical clustering of all expressed genes (i) segregates dissected green tumors into two clusters (n = 6 tumors per cluster) identified as AdenomaG and LUAD based on known histologic progression markers (ref. ; j). Row-normalized expression counts are shown in heatmap (i). Horizontal lines in each violin in j represent the 75th percentile, median and 25th percentile. P values were from two-sided Wilcoxon rank-sum tests. k, Gene set enrichment analysis using MSigDB Hallmarks (H1) shows top enriched gene sets for upregulated (red, log₂ fold change > 2 and FDR < 0.05) and downregulated (blue, log₂ fold change <−1 and FDR < 0.05) genes in LUAD over AdenomaG. P values were from one-sided hypergeometric tests. Source data

Extended Data Fig. 6. Genome-wide chromatin tracing of pancreatic ductal adenocarcinoma progression.

a,b, Whole-section fluorescence images of a K-MADM-Trp53 pancreas with PanINs (a) and PDAC (b). c,d, Immunofluorescence staining (top) of the duct cell marker CK19 and DAPI staining in a field of view, each in a and b. Fluorescent protein imaging (bottom) of GFP, tdTomato and DAPI staining in the corresponding fields. Images are from two independent experiments. The images are maximum-intensity z-projections from 10-μm z-stacks. e, Matrix of mean interloci distances between all genomic loci in normal duct cells. n = 1,529 cells. f, Mean interloci spatial distance versus genomic distance for all pairs of genomic loci on each autosome in duct cells. Different pseudocolors represent different autosomes. n = 6,039 intrachromosomal interloci pairs. n = 1,529 cells. g–i, Distributions of the decompaction, heterogeneity and demixing scores of each autosome (n = 19 autosomes) in each pancreatic cell state. n = 1,529, 123, 189, 361 and 475 for normal duct, PanIN R, PanIN Y, PanIN G and PDAC cells, respectively. Cell numbers of each cell state are identical in g–k. j, log₂ fold change of mean interchromosomal distances, comparing each cancer state to normal duct cells. k, Distribution of the mean log₂ fold change of radial scores of chr17, chr1 and chr19, comparing each cancer state to normal duct cells. P values of two-sided Wilcoxon signed-rank tests comparing each cancer state to WT duct cells with matched genomic loci (g, h, j and k) or matched autosomes (i) are shown. In g–k, the horizontal lines of each box represent the 75th percentile, median and 25th percentile. Whiskers extend to the nonoutlier maximum and nonoutlier minimum. Outliers are defined as values away from the top or bottom of the box by more than 1.5 times interquartile range. Source data

Extended Data Fig. 7. Subsampling analysis of chromatin folding changes during lung cancer progression.

a–c, Distributions of the decompaction (mean interloci distance), heterogeneity (COV of interloci distance) and demixing scores of each autosome (n = 19) in a randomly subsampled population of 100 cells per cell state. d, Distribution of interchromosomal distances in a randomly subsampled population of 100 cells per cell state. e, Distribution of the polarization indices of A and B compartments of each autosome (n = 19) in a randomly subsampled population of 100 cells per cell state. f, Distribution of the radial scores of chr15, chr2 and chr10 in a randomly subsampled population of 100 cells per cell state. P values of two-sided Wilcoxon signed-rank test (a–c, e and f) and two-sided Wilcoxon rank-sum test (d) are displayed. In a–f, the horizontal lines of each box from top to bottom represent the 75th percentile, median and 25th percentile. Whiskers extend to the nonoutlier maximum and nonoutlier minimum. Outliers are defined as values at least 1.5 times interquartile range away from the top or bottom of the box. Source data

Extended Data Fig. 8. Fine-scale tracing of candidate drivers of cancer progression.

a, Schematic illustration of high-resolution chromatin tracing targeting the cis-regulatory regions of 15 genes, including CPD, Kras and Myc. Distribution of detected loci number per cell is displayed. n = 12,661 cells combining the five cell states. b, Distribution of the decompaction (mean interloci distance) scores of all target genes. n = 2,328, 654, 558 and 9,121 for normal AT2, AdenomaR/Y, AdenomaG and LUAD cells, respectively. Cell numbers in each cell state are identical in b–f. P values of two-sided Wilcoxon signed-rank tests are displayed. c, Pileup heatmap of normalized E–P distances centered around each E–P loop. E–P loops of CPD genes are called in LUAD cells. d, Distribution of the log₂ fold change of the normalized distances between promoter and putative enhancers of CPD genes, comparing each cancer state to AT2 cells. P value of two-sided Wilcoxon signed-rank test is displayed. e, Normalized interloci distance matrices of the target genomic regions surrounding the Kras, Myc and Foxa3 (CPD) genes in AT2, adenoma and LUAD cell states. The green circles designate putative enhancer–promoter contacts. Putative enhancer and gene annotation tracks are aligned to the target regions. f, Distribution of the normalized distances between promoter and putative enhancers of the Kras and Myc genes. P values of two-sided Wilcoxon rank-sum tests are displayed. g, Distribution of the normalized scA/B scores of Myc gene-containing target genomic region in AT2 and adenoma cells. P value of two-sided Wilcoxon rank-sum test is displayed. In a, b, d, f and g, the horizontal lines of each box from top to bottom represent the 75th percentile, median and 25th percentile. Whiskers extend to the nonoutlier maximum and nonoutlier minimum. Outliers are defined as values at least 1.5 times interquartile range away from the top or bottom of the box. Source data

Extended Data Fig. 9. 3D genome organization in cancer cells is largely independent of spatial proximity to immune cells.

a–c, Mean log₂ fold change of decompaction (a), heterogeneity (b) and demixing (c) scores of each chromosome, comparing AT2/cancer cells near (less than 10 μm) versus far (more than 10 μm) from CD45+ immune cells. The P values (a,b) or FDR (c) with significance (<0.05) from two-sided Wilcoxon signed-rank test (a,b) or two-sided Levene’s test (c) are displayed. d, Distribution of polarization indices of A–B compartments in AT2/cancer cells near or far from immune cells. Two-sided Wilcoxon rank-sum test yielded no significant P values (P < 0.05). e, t-SNE plots of single-cell 3D chromatin conformations in AT2, adenoma and LUAD cells show no distinct clusters based on spatial proximity to immune cells. In a, b and d, the horizontal lines of each box from top to bottom represent the 75th percentile, median and 25th percentile. Whiskers extend to the nonoutlier maximum and nonoutlier minimum. Outliers are defined as values at least 1.5 times interquartile range away from the top or bottom of the box. In a-d, n = 3,806, 758, 1,563, 1,882 and 17,711 for WT AT2, AdenomaR, AdenomaY, AdenomaG and LUAD cells, respectively (excluding early AT2 and adenoma replicates without CD45 labeling). In e, n = 2,719, 1,710 and 8,834 for WT AT2, Adenoma and LUAD cells, respectively (cells with at least 10 traces). Source data

Extended Data Fig. 10. Rnf2 regulates 3D genome organization through a ubiquitin ligase-independent activity.

a, Western blot analysis of Rnf2 protein levels following rescue of shRNA knockdown (shRnf2-3) with stable transduction of Rnf2 WT, ubiquitin ligase dead Rnf2 I53S mutant or empty vector (EV) control. HSP90 is loading control. Results were repeated in three biological replicates. b, scA/B score changes from shRnf2-3 to shRnf2-3 + WT Rnf2 (left) and to shRnf2-3 + I53S Rnf2 (right) versus those from AdenomaG to LUAD. n = 473 target genomic regions. Spearman correlation coefficients and P values from two-sided t-tests are shown. c, Western blot of Rnf2 and H2AK119ub in KP LUAD cells with Rnf2-dTAG after treatment with the dTAG-13 ligand or negative control ligand (a diastereomer of dTAG-13) at the designated times (0, 0.5 or 12 h). Results were repeated in three biological replicates. d, Immunofluorescence images of Rnf2 and H2AK119ub in KP LUAD cells with Rnf2-dTAG after treatment with dTAG-13 ligand or negative control at designated times. e, Quantification of the immunofluorescence intensities in d. n = 461, 534, 177, 1,040, 214, 602, 251 and 1,168 cells from left to right. The horizontal lines of each box from top to bottom represent the 75th percentile, median and 25th percentile. Whiskers extend to the nonoutlier maximum and nonoutlier minimum. Outliers are defined as values at least 1.5 times interquartile range away from the top or bottom of the box. P value from two-sided Wilcoxon rank-sum test is shown. f, scA/B score changes from Rnf2-degraded cells (dTAG-13) to Rnf2 nondegraded cells (dTAG-13 negative control) versus those from AdenomaG to LUAD and those from shRnf2 to shNTC, with n = 473 target genomic regions. Spearman correlation coefficients and P values from two-sided t-tests are shown. Source data