RNA sequencing-based single sample predictors of molecular subtype and risk of recurrence for clinical assessment of early-stage breast cancer.

Multigene assays for molecular subtypes and biomarkers can aid management of early invasive breast cancer. Using RNA-sequencing we aimed to develop single-sample predictor (SSP) models for clinical markers, subtypes, and risk of recurrence (ROR). A cohort of 7743 patients was divided into training and test set. We trained SSPs for subtypes and ROR assigned by nearest-centroid (NC) methods and SSPs for biomarkers from histopathology. Classifications were compared with Prosigna in two external cohorts (ABiM, n = 100 and OSLO2-EMIT0, n = 103). Prognostic value was assessed using distant recurrence-free interval. Agreement between SSP and NC for PAM50 (five subtypes) was high (85%, Kappa = 0.78) for Subtype (four subtypes) very high (90%, Kappa = 0.84) and for ROR risk category high (84%, Kappa = 0.75, weighted Kappa = 0.90). Prognostic value was assessed as equivalent and clinically relevant. Agreement with histopathology was very high or high for receptor status, while moderate for Ki67 status and poor for Nottingham histological grade. SSP and Prosigna concordance was high for subtype (OSLO-EMIT0 83%, Kappa = 0.73 and ABiM 80%, Kappa = 0.72) and moderate and high for ROR risk category (68 and 84%, Kappa = 0.50 and 0.70, weighted Kappa = 0.70 and 0.78). Pooled concordance for emulated treatment recommendation dichotomized for chemotherapy was high (85%, Kappa = 0.66). Retrospective evaluation suggested that SSP application could change chemotherapy recommendations for up to 17% of postmenopausal ER+/HER2-/N0 patients with balanced escalation and de-escalation. Results suggest that NC and SSP models are interchangeable on a group-level and nearly so on a patient level and that SSP models can be derived to closely match clinical tests.

Comparative analysis of the molecular subtype landscape in canine and human mammary gland tumors.

Breast cancers in humans belong to one of several intrinsic molecular subtypes each with different tumor biology and different clinical impact. Mammary gland tumors in dogs are proposed as a relevant comparative model for human breast cancer; however, it is still unclear whether the intrinsic molecular subtypes have the same significance in dogs and humans. Using publicly available data, we analyzed gene expression and whole-exome sequencing data from 158 canine mammary gland tumors. We performed molecular subtyping using the PAM50 method followed by subtype-specific comparisons of gene expression characteristics, mutation patterns and copy number profiles between canine tumors and human breast tumors from The Cancer Genome Atlas (TCGA) breast cancer cohort (n = 1097). We found that luminal A canine tumors greatly resemble luminal A human tumors both in gene expression characteristics, mutations and copy number profiles. Also, the basal-like canine and human tumors were relatively similar, with low expression of luminal epithelial markers and high expression of genes involved in cell proliferation. There were, however, distinct differences in immune-related gene expression patterns in basal-like tumors between the two species. Characteristic HER2-enriched and luminal B subtypes were not present in the canine cohort, and we found no tumors with high-level ERBB2 amplifications. Benign and malignant canine tumors displayed similar PAM50 subtype characteristics. Our findings indicate that deeper understanding of the different molecular subtypes in canine mammary gland tumors will further improve the value of canines as comparative models for human breast cancer.

Sample Preparation Approach Influences PAM50 Risk of Recurrence Score in Early Breast Cancer.

The PAM50 gene expression subtypes and the associated risk of recurrence (ROR) score are used to predict the risk of recurrence and the benefits of adjuvant therapy in early-stage breast cancer. The Prosigna assay includes the PAM50 subtypes along with their clinicopathological features, and is approved for treatment recommendations for adjuvant hormonal therapy and chemotherapy in hormone-receptor-positive early breast cancer. The Prosigna test utilizes RNA extracted from macrodissected tumor cells obtained from formalin-fixed, paraffin-embedded (FFPE) tissue sections. However, RNA extracted from fresh-frozen (FF) bulk tissue without macrodissection is widely used for research purposes, and yields high-quality RNA for downstream analyses. To investigate the impact of the sample preparation approach on ROR scores, we analyzed 94 breast carcinomas included in an observational study that had available gene expression data from macrodissected FFPE tissue and FF bulk tumor tissue, along with the clinically approved Prosigna scores for the node-negative, hormone-receptor-positive, HER2-negative cases (n = 54). ROR scores were calculated in R; the resulting two sets of scores from FFPE and FF samples were compared, and treatment recommendations were evaluated. Overall, ROR scores calculated based on the macrodissected FFPE tissue were consistent with the Prosigna scores. However, analyses from bulk tissue yielded a higher proportion of cases classified as normal-like; these were samples with relatively low tumor cellularity, leading to lower ROR scores. When comparing ROR scores (low, intermediate, and high), discordant cases between the two preparation approaches were revealed among the luminal tumors; the recommended treatment would have changed in a minority of cases.

Tailored graphical lasso for data integration in gene network reconstruction.

BACKGROUND: Identifying gene interactions is a topic of great importance in genomics, and approaches based on network models provide a powerful tool for studying these. Assuming a Gaussian graphical model, a gene association network may be estimated from multiomic data based on the non-zero entries of the inverse covariance matrix. Inferring such biological networks is challenging because of the high dimensionality of the problem, making traditional estimators unsuitable. The graphical lasso is constructed for the estimation of sparse inverse covariance matrices in such situations, using [Formula: see text]-penalization on the matrix entries. The weighted graphical lasso is an extension in which prior biological information from other sources is integrated into the model. There are however issues with this approach, as it naïvely forces the prior information into the network estimation, even if it is misleading or does not agree with the data at hand. Further, if an associated network based on other data is used as the prior, the method often fails to utilize the information effectively. RESULTS: We propose a novel graphical lasso approach, the tailored graphical lasso, that aims to handle prior information of unknown accuracy more effectively. We provide an R package implementing the method, tailoredGlasso. Applying the method to both simulated and real multiomic data sets, we find that it outperforms the unweighted and weighted graphical lasso in terms of all performance measures we consider. In fact, the graphical lasso and weighted graphical lasso can be considered special cases of the tailored graphical lasso, and a parameter determined by the data measures the usefulness of the prior information. We also find that among a larger set of methods, the tailored graphical is the most suitable for network inference from high-dimensional data with prior information of unknown accuracy. With our method, mRNA data are demonstrated to provide highly useful prior information for protein-protein interaction networks. CONCLUSIONS: The method we introduce utilizes useful prior information more effectively without involving any risk of loss of accuracy should the prior information be misleading.

Multi-Omics Marker Analysis Enables Early Prediction of Breast Tumor Progression.

Ductal carcinoma in situ (DCIS) is a preinvasive form of breast cancer with a highly variable potential of becoming invasive and affecting mortality of the patients. Due to the lack of accurate markers of disease progression, many women with detected DCIS are currently overtreated. To distinguish those DCIS cases who are likely to require therapy from those who should be left untreated, there is a need for robust and predictive biomarkers extracted from molecular or genetic profiles. We developed a supervised machine learning approach that implements multi-omics feature selection and model regularization for the identification of biomarker combinations that could be used to distinguish low-risk DCIS lesions from those with a higher likelihood of progression. To investigate the genetic heterogeneity of disease progression, we applied this approach to 40 pure DCIS and 259 invasive breast cancer (IBC) samples profiled with genome-wide transcriptomics, DNA methylation, and DNA copy number variation. Feature selection using the multi-omics Lasso-regularized algorithm identified both known genes involved in breast cancer development, as well as novel markers for early detection. Even though the gene expression-based model features led to the highest classification accuracy alone, methylation data provided a complementary source of features and improved especially the sensitivity of correctly classifying DCIS cases. We also identified a number of repeatedly misclassified DCIS cases when using either the expression or methylation markers. A small panel of 10 gene markers was able to distinguish DCIS and IBC cases with high accuracy in nested cross-validation (AU-ROC = 0.99). The marker panel was not specific to any of the established breast cancer subtypes, suggesting that the 10-gene signature may provide a subtype-agnostic and cost-effective approach for breast cancer detection and patient stratification. We further confirmed high accuracy of the 10-gene signature in an external validation cohort (AU-ROC = 0.95), profiled using distinct transcriptomic assay, hence demonstrating robustness of the risk signature.

Contrasting DCIS and invasive breast cancer by subtype suggests basal-like DCIS as distinct lesions.

Ductal carcinoma in situ (DCIS) is a non-invasive type of breast cancer with highly variable potential of becoming invasive and affecting mortality. Currently, many patients with DCIS are overtreated due to the lack of specific biomarkers that distinguish low risk lesions from those with a higher risk of progression. In this study, we analyzed 57 pure DCIS and 313 invasive breast cancers (IBC) from different patients. Three levels of genomic data were obtained; gene expression, DNA methylation, and DNA copy number. We performed subtype stratified analyses and identified key differences between DCIS and IBC that suggest subtype specific progression. Prominent differences were found in tumors of the basal-like subtype: Basal-like DCIS were less proliferative and showed a higher degree of differentiation than basal-like IBC. Also, core basal tumors (characterized by high correlation to the basal-like centroid) were not identified amongst DCIS as opposed to IBC. At the copy number level, basal-like DCIS exhibited fewer copy number aberrations compared with basal-like IBC. An intriguing finding through analysis of the methylome was hypermethylation of multiple protocadherin genes in basal-like IBC compared with basal-like DCIS and normal tissue, possibly caused by long range epigenetic silencing. This points to silencing of cell adhesion-related genes specifically in IBC of the basal-like subtype. Our work confirms that subtype stratification is essential when studying progression from DCIS to IBC, and we provide evidence that basal-like DCIS show less aggressive characteristics and question the assumption that basal-like DCIS is a direct precursor of basal-like invasive breast cancer.

Immune phenotype of tumor microenvironment predicts response to bevacizumab in neoadjuvant treatment of ER-positive breast cancer.

Antiangiogenic drugs are potentially a useful supplement to neoadjuvant chemotherapy for a subgroup of patients with human epidermal growth factor receptor 2 (HER2) negative breast cancer, but reliable biomarkers for improved response are lacking. Here, we report on a randomized phase II clinical trial to study the added effect of bevacizumab in neoadjuvant chemotherapy with FEC100 (5-fluorouracil, epirubicin and cyclophosphamide) and taxanes (n = 132 patients). Gene expression from the tumors was obtained before neoadjuvant treatment, and treatment response was evaluated by residual cancer burden (RCB) at time of surgery. Bevacizumab increased the proportion of complete responders (RCB class 0) from 5% to 20% among patients with estrogen receptor (ER) positive tumors (P = .02). Treatment with bevacizumab was associated with improved 8-year disease-free survival (P = .03) among the good responders (RCB class 0 or I). Patients treated with paclitaxel (n = 45) responded better than those treated with docetaxel (n = 21; P = .03). Improved treatment response was associated with higher proliferation rate and an immune phenotype characterized by high presence of classically activated M1 macrophages, activated NK cells and memory activated CD4 T cells. Treatment with bevacizumab increased the number of adverse events, including hemorrhage, hypertension, infection and febrile neutropenia, but despite this, the ECOG status was not affected.

An independent poor-prognosis subtype of breast cancer defined by a distinct tumor immune microenvironment.

How mixtures of immune cells associate with cancer cell phenotype and affect pathogenesis is still unclear. In 15 breast cancer gene expression datasets, we invariably identify three clusters of patients with gradual levels of immune infiltration. The intermediate immune infiltration cluster (Cluster B) is associated with a worse prognosis independently of known clinicopathological features. Furthermore, immune clusters are associated with response to neoadjuvant chemotherapy. In silico dissection of the immune contexture of the clusters identified Cluster A as immune cold, Cluster C as immune hot while Cluster B has a pro-tumorigenic immune infiltration. Through phenotypical analysis, we find epithelial mesenchymal transition and proliferation associated with the immune clusters and mutually exclusive in breast cancers. Here, we describe immune clusters which improve the prognostic accuracy of immune contexture in breast cancer. Our discovery of a novel independent prognostic factor in breast cancer highlights a correlation between tumor phenotype and immune contexture.

A Bayesian two-way latent structure model for genomic data integration reveals few pan-genomic cluster subtypes in a breast cancer cohort.

MOTIVATION: Unsupervised clustering is important in disease subtyping, among having other genomic applications. As genomic data has become more multifaceted, how to cluster across data sources for more precise subtyping is an ever more important area of research. Many of the methods proposed so far, including iCluster and Cluster of Cluster Assignments (COCAs), make an unreasonable assumption of a common clustering across all data sources, and those that do not are fewer and tend to be computationally intensive. RESULTS: We propose a Bayesian parametric model for integrative, unsupervised clustering across data sources. In our two-way latent structure model, samples are clustered in relation to each specific data source, distinguishing it from methods like COCAs and iCluster, but cluster labels have across-dataset meaning, allowing cluster information to be shared between data sources. A common scaling across data sources is not required, and inference is obtained by a Gibbs Sampler, which we improve with a warm start strategy and modified density functions to robustify and speed convergence. Posterior interpretation allows for inference on common clusterings occurring among subsets of data sources. An interesting statistical formulation of the model results in sampling from closed-form posteriors despite incorporation of a complex latent structure. We fit the model with Gaussian and more general densities, which influences the degree of across-dataset cluster label sharing. Uniquely among integrative clustering models, our formulation makes no nestedness assumptions of samples across data sources so that a sample missing data from one genomic source can be clustered according to its existing data sources. We apply our model to a Norwegian breast cancer cohort of ductal carcinoma in situ and invasive tumors, comprised of somatic copy-number alteration, methylation and expression datasets. We find enrichment in the Her2 subtype and ductal carcinoma among those observations exhibiting greater cluster correspondence across expression and CNA data. In general, there are few pan-genomic clusterings, suggesting that models assuming a common clustering across genomic data sources might yield misleading results. AVAILABILITY AND IMPLEMENTATION: The model is implemented in an R package called twl ('two-way latent'), available on CRAN. Data for analysis are available within the R package. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

A Longitudinal Study of the Association between Mammographic Density and Gene Expression in Normal Breast Tissue.

High mammographic density (MD) is associated with a 4-6 times increase in breast cancer risk. For post-menopausal women, MD often decreases over time, but little is known about the underlying biological mechanisms. MD reflects breast tissue composition, and may be associated with microenvironment subtypes previously identified in tumor-adjacent normal tissue. Currently, these subtypes have not been explored in normal breast tissue. We obtained biopsies from breasts of healthy women at two different time points several years apart and performed microarray gene expression analysis. At time point 1, 65 samples with both MD and gene expression were available. At time point 2, gene expression and MD data were available from 17 women, of which 11 also had gene expression data available from the first time point. We validated findings from our previous study; negative correlation between RBL1 and MD in post-menopausal women, indicating involvement of the TGFß pathway. We also found that breast tissue samples from women with a large decrease in MD sustained higher expression of genes in the histone family H4. In addition, we explored the previously defined active and inactive microenvironment subtypes and demonstrated that normal breast samples of the active subtype had characteristics similar to the claudin-low breast cancer subtype. Breast biopsies from healthy women are challenging to obtain, but despite a limited sample size, we have identified possible mechanisms relevant for changes in breast biology and MD over time that may be of importance for breast cancer risk and tumor initiation.

Integrated analysis of DNA-methylation and gene expression using high-dimensional penalized regression: a cohort study on bone mineral density in postmenopausal women.

BACKGROUND: Using high-dimensional penalized regression we studied genome-wide DNA-methylation in bone biopsies of 80 postmenopausal women in relation to their bone mineral density (BMD). The women showed BMD varying from severely osteoporotic to normal. Global gene expression data from the same individuals was available, and since DNA-methylation often affects gene expression, the overall aim of this paper was to include both of these omics data sets into an integrated analysis. METHODS: The classical penalized regression uses one penalty, but we incorporated individual penalties for each of the DNA-methylation sites. These individual penalties were guided by the strength of association between DNA-methylations and gene transcript levels. DNA-methylations that were highly associated to one or more transcripts got lower penalties and were therefore favored compared to DNA-methylations showing less association to expression. Because of the complex pathways and interactions among genes, we investigated both the association between DNA-methylations and their corresponding cis gene, as well as the association between DNA-methylations and trans-located genes. Two integrating penalized methods were used: first, an adaptive group-regularized ridge regression, and secondly, variable selection was performed through a modified version of the weighted lasso. RESULTS: When information from gene expressions was integrated, predictive performance was considerably improved, in terms of predictive mean square error, compared to classical penalized regression without data integration. We found a 14.7% improvement in the ridge regression case and a 17% improvement for the lasso case. Our version of the weighted lasso with data integration found a list of 22 interesting methylation sites. Several corresponded to genes that are known to be important in bone formation. Using BMD as response and these 22 methylation sites as covariates, least square regression analyses resulted in R2=0.726, comparable to an average R2=0.438 for 10000 randomly selected groups of DNA-methylations with group size 22. CONCLUSIONS: Two recent types of penalized regression methods were adapted to integrate DNA-methylation and their association to gene expression in the analysis of bone mineral density. In both cases predictions clearly benefit from including the additional information on gene expressions.

Prognostic value of PAM50 and risk of recurrence score in patients with early-stage breast cancer with long-term follow-up.

BACKGROUND: The aim of this study was to investigate the prognostic value of the PAM50 intrinsic subtypes and risk of recurrence (ROR) score in patients with early breast cancer and long-term follow-up. A special focus was placed on hormone receptor-positive/human epidermal growth factor receptor 2-negative (HR+/HER2-) pN0 patients not treated with chemotherapy. METHODS: Patients with early breast cancer (n = 653) enrolled in the observational Oslo1 study (1995-1998) were followed for distant recurrence and breast cancer death. Clinicopathological parameters were collected from hospital records. The primary tumors were analyzed using the Prosigna® PAM50 assay to determine the prognostic value of the intrinsic subtypes and ROR score in comparison with pathological characteristics. The primary endpoints were distant disease-free survival (DDFS) and breast cancer-specific survival (BCSS). RESULTS: Of 653 tumors, 52.2% were classified as luminal A, 26.5% as luminal B, 10.6% as HER2-enriched, and 10.7% as basal-like. Among the HR+/HER2- patients (n = 476), 37.8% were categorized as low risk by ROR score, 22.7% as intermediate risk, and 39.5% as high risk. Median follow-up durations for BCSS and DDFS were 16.6 and 7.1 years, respectively. Multivariate analysis showed that intrinsic subtypes (all patients) and ROR risk classification (HR+/HER2- patients) yielded strong prognostic information. Among the HR+/HER2- pN0 patients with no adjuvant treatment (n = 231), 53.7% of patients had a low ROR, and their prognosis at 15 years was excellent (15-year BCSS 96.3%). Patients with intermediate risk had reduced survival compared with those with low risk (p = 0.005). In contrast, no difference in survival between the low- and intermediate-risk groups was seen for HR+/HER2- pN0 patients who received tamoxifen only. Ki-67 protein, grade, and ROR score were analyzed in the unselected, untreated pT1pN0 HR+/HER2- population (n = 171). In multivariate analysis, ROR score outperformed both Ki-67 and grade. Furthermore, 55% of patients who according to the PREDICT tool ( http://www.predict.nhs.uk/ ) would be considered chemotherapy candidates were ROR low risk (33%) or luminal A ROR intermediate risk (22%). CONCLUSIONS: The PAM50 intrinsic subtype classification and ROR score improve classification of patients with breast cancer into prognostic groups, allowing for a more precise identification of future recurrence risk and providing an improved basis for adjuvant treatment decisions. Node-negative patients with low ROR scores had an excellent outcome at 15 years even in the absence of adjuvant therapy.

Distinct DNA methylation profiles in bone and blood of osteoporotic and healthy postmenopausal women.

DNA methylation affects expression of associated genes and may contribute to the missing genetic effects from genome-wide association studies of osteoporosis. To improve insight into the mechanisms of postmenopausal osteoporosis, we combined transcript profiling with DNA methylation analyses in bone. RNA and DNA were isolated from 84 bone biopsies of postmenopausal donors varying markedly in bone mineral density (BMD). In all, 2529 CpGs in the top 100 genes most significantly associated with BMD were analyzed. The methylation levels at 63 CpGs differed significantly between healthy and osteoporotic women at 10% false discovery rate (FDR). Five of these CpGs at 5% FDR could explain 14% of BMD variation. To test whether blood DNA methylation reflect the situation in bone (as shown for other tissues), an independent cohort was selected and BMD association was demonstrated in blood for 13 of the 63 CpGs. Four transcripts representing inhibitors of bone metabolism-MEPE, SOST, WIF1, and DKK1-showed correlation to a high number of methylated CpGs, at 5% FDR. Our results link DNA methylation to the genetic influence modifying the skeleton, and the data suggest a complex interaction between CpG methylation and gene regulation. This is the first study in the hitherto largest number of postmenopausal women to demonstrate a strong association among bone CpG methylation, transcript levels, and BMD/fracture. This new insight may have implications for evaluation of osteoporosis stage and susceptibility.

Hi-C-constrained physical models of human chromosomes recover functionally-related properties of genome organization.

Combining genome-wide structural models with phenomenological data is at the forefront of efforts to understand the organizational principles regulating the human genome. Here, we use chromosome-chromosome contact data as knowledge-based constraints for large-scale three-dimensional models of the human diploid genome. The resulting models remain minimally entangled and acquire several functional features that are observed in vivo and that were never used as input for the model. We find, for instance, that gene-rich, active regions are drawn towards the nuclear center, while gene poor and lamina associated domains are pushed to the periphery. These and other properties persist upon adding local contact constraints, suggesting their compatibility with non-local constraints for the genome organization. The results show that suitable combinations of data analysis and physical modelling can expose the unexpectedly rich functionally-related properties implicit in chromosome-chromosome contact data. Specific directions are suggested for further developments based on combining experimental data analysis and genomic structural modelling.

Better prediction by use of co-data: adaptive group-regularized ridge regression.

For many high-dimensional studies, additional information on the variables, like (genomic) annotation or external p-values, is available. In the context of binary and continuous prediction, we develop a method for adaptive group-regularized (logistic) ridge regression, which makes structural use of such 'co-data'. Here, 'groups' refer to a partition of the variables according to the co-data. We derive empirical Bayes estimates of group-specific penalties, which possess several nice properties: (i) They are analytical. (ii) They adapt to the informativeness of the co-data for the data at hand. (iii) Only one global penalty parameter requires tuning by cross-validation. In addition, the method allows use of multiple types of co-data at little extra computational effort. We show that the group-specific penalties may lead to a larger distinction between 'near-zero' and relatively large regression parameters, which facilitates post hoc variable selection. The method, termed GRridge, is implemented in an easy-to-use R-package. It is demonstrated on two cancer genomics studies, which both concern the discrimination of precancerous cervical lesions from normal cervix tissues using methylation microarray data. For both examples, GRridge clearly improves the predictive performances of ordinary logistic ridge regression and the group lasso. In addition, we show that for the second study, the relatively good predictive performance is maintained when selecting only 42 variables.

HiBrowse: multi-purpose statistical analysis of genome-wide chromatin 3D organization.

UNLABELLED: Recently developed methods that couple next-generation sequencing with chromosome conformation capture-based techniques, such as Hi-C and ChIA-PET, allow for characterization of genome-wide chromatin 3D structure. Understanding the organization of chromatin in three dimensions is a crucial next step in the unraveling of global gene regulation, and methods for analyzing such data are needed. We have developed HiBrowse, a user-friendly web-tool consisting of a range of hypothesis-based and descriptive statistics, using realistic assumptions in null-models. AVAILABILITY AND IMPLEMENTATION: HiBrowse is supported by all major browsers, and is freely available at http://hyperbrowser.uio.no/3d. Software is implemented in Python, and source code is available for download by following instructions on the main site.

The Genomic HyperBrowser: an analysis web server for genome-scale data.

The immense increase in availability of genomic scale datasets, such as those provided by the ENCODE and Roadmap Epigenomics projects, presents unprecedented opportunities for individual researchers to pose novel falsifiable biological questions. With this opportunity, however, researchers are faced with the challenge of how to best analyze and interpret their genome-scale datasets. A powerful way of representing genome-scale data is as feature-specific coordinates relative to reference genome assemblies, i.e. as genomic tracks. The Genomic HyperBrowser (http://hyperbrowser.uio.no) is an open-ended web server for the analysis of genomic track data. Through the provision of several highly customizable components for processing and statistical analysis of genomic tracks, the HyperBrowser opens for a range of genomic investigations, related to, e.g., gene regulation, disease association or epigenetic modifications of the genome.

Handling realistic assumptions in hypothesis testing of 3D co-localization of genomic elements.

The study of chromatin 3D structure has recently gained much focus owing to novel techniques for detecting genome-wide chromatin contacts using next-generation sequencing. A deeper understanding of the architecture of the DNA inside the nucleus is crucial for gaining insight into fundamental processes such as transcriptional regulation, genome dynamics and genome stability. Chromatin conformation capture-based methods, such as Hi-C and ChIA-PET, are now paving the way for routine genome-wide studies of chromatin 3D structure in a range of organisms and tissues. However, appropriate methods for analyzing such data are lacking. Here, we propose a hypothesis test and an enrichment score of 3D co-localization of genomic elements that handles intra- or interchromosomal interactions, both separately and jointly, and that adjusts for biases caused by structural dependencies in the 3D data. We show that maintaining structural properties during resampling is essential to obtain valid estimation of P-values. We apply the method on chromatin states and a set of mutated regions in leukemia cells, and find significant co-localization of these elements, with varying enrichment scores, supporting the role of chromatin 3D structure in shaping the landscape of somatic mutations in cancer.