bootRanges: flexible generation of null sets of genomic ranges for hypothesis testing.

MOTIVATION: Enrichment analysis is a widely utilized technique in genomic analysis that aims to determine if there is a statistically significant association between two sets of genomic features. To conduct this type of hypothesis testing, an appropriate null model is typically required. However, the null distribution that is commonly used can be overly simplistic and may result in inaccurate conclusions. RESULTS: bootRanges provides fast functions for generation of block bootstrapped genomic ranges representing the null hypothesis in enrichment analysis. As part of a modular workflow, bootRanges offers greater flexibility for computing various test statistics leveraging other Bioconductor packages. We show that shuffling or permutation schemes may result in overly narrow test statistic null distributions and over-estimation of statistical significance, while creating new range sets with a block bootstrap preserves local genomic correlation structure and generates more reliable null distributions. It can also be used in more complex analyses, such as accessing correlations between cis-regulatory elements (CREs) and genes across cell types or providing optimized thresholds, e.g. log fold change (logFC) from differential analysis. AVAILABILITY AND IMPLEMENTATION: bootRanges is freely available in the R/Bioconductor package nullranges hosted at https://bioconductor.org/packages/nullranges.

matchRanges: generating null hypothesis genomic ranges via covariate-matched sampling.

MOTIVATION: Deriving biological insights from genomic data commonly requires comparing attributes of selected genomic loci to a null set of loci. The selection of this null set is non-trivial, as it requires careful consideration of potential covariates, a problem that is exacerbated by the non-uniform distribution of genomic features including genes, enhancers, and transcription factor binding sites. Propensity score-based covariate matching methods allow the selection of null sets from a pool of possible items while controlling for multiple covariates; however, existing packages do not operate on genomic data classes and can be slow for large data sets making them difficult to integrate into genomic workflows. RESULTS: To address this, we developed matchRanges, a propensity score-based covariate matching method for the efficient and convenient generation of matched null ranges from a set of background ranges within the Bioconductor framework. AVAILABILITY AND IMPLEMENTATION: Package: https://bioconductor.org/packages/nullranges, Code: https://github.com/nullranges, Documentation: https://nullranges.github.io/nullranges.

excluderanges: exclusion sets for T2T-CHM13, GRCm39, and other genome assemblies.

SUMMARY: Exclusion regions are sections of reference genomes with abnormal pileups of short sequencing reads. Removing reads overlapping them improves biological signal, and these benefits are most pronounced in differential analysis settings. Several labs created exclusion region sets, available primarily through ENCODE and Github. However, the variety of exclusion sets creates uncertainty which sets to use. Furthermore, gap regions (e.g. centromeres, telomeres, short arms) create additional considerations in generating exclusion sets. We generated exclusion sets for the latest human T2T-CHM13 and mouse GRCm39 genomes and systematically assembled and annotated these and other sets in the excluderanges R/Bioconductor data package, also accessible via the BEDbase.org API. The package provides unified access to 82 GenomicRanges objects covering six organisms, multiple genome assemblies, and types of exclusion regions. For human hg38 genome assembly, we recommend hg38.Kundaje.GRCh38_unified_blacklist as the most well-curated and annotated, and sets generated by the Blacklist tool for other organisms. AVAILABILITY AND IMPLEMENTATION: https://bioconductor.org/packages/excluderanges/. Package website: https://dozmorovlab.github.io/excluderanges/.

Melanoma differentiation associated gene-9/syndecan binding protein promotes hepatocellular carcinoma.

BACKGROUND AND AIMS: The oncogene Melanoma differentiation associated gene-9/syndecan binding protein (MDA-9/SDCBP) is overexpressed in many cancers, promoting aggressive, metastatic disease. However, the role of MDA-9 in regulating hepatocellular carcinoma (HCC) has not been well studied. APPROACH AND RESULTS: To unravel the function of MDA-9 in HCC, we generated and characterized a transgenic mouse with hepatocyte-specific overexpression of MDA-9 (Alb/MDA-9). Compared with wild-type (WT) littermates, Alb/MDA-9 mice demonstrated significantly higher incidence of N-nitrosodiethylamine/phenobarbital-induced HCC, with marked activation and infiltration of macrophages. RNA sequencing (RNA-seq) in naive WT and Alb/MDA-9 hepatocytes identified activation of signaling pathways associated with invasion, angiogenesis, and inflammation, especially NF-κB and integrin-linked kinase signaling pathways. In nonparenchymal cells purified from naive livers, single-cell RNA-seq showed activation of Kupffer cells and macrophages in Alb/MDA-9 mice versus WT mice. A robust increase in the expression of Secreted phosphoprotein 1 (Spp1/osteopontin) was observed upon overexpression of MDA-9. Inhibition of NF-κB pathway blocked MDA-9-induced Spp1 induction, and knock down of Spp1 resulted in inhibition of MDA-9-induced macrophage migration, as well as angiogenesis. CONCLUSIONS: Alb/MDA-9 is a mouse model with MDA-9 overexpression in any tissue type. Our findings unravel an HCC-promoting role of MDA-9 mediated by NF-κB and Spp1 and support the rationale of using MDA-9 inhibitors as a potential treatment for aggressive HCC.

Astrocyte Elevated Gene-1 Cys75 S-Palmitoylation by ZDHHC6 Regulates Its Biological Activity.

Nonalcoholic fatty liver disease is a major risk factor for hepatocellular carcinoma (HCC). Astrocyte elevated gene-1/Metadherin (AEG-1/MTDH) augments lipid accumulation (steatosis), inflammation, and tumorigenesis, thereby promoting the whole spectrum of this disease process. Targeting AEG-1 is a potential interventional strategy for nonalcoholic steatohepatitis (NASH) and HCC. Thus, proper understanding of the regulation of this molecule is essential. We found that AEG-1 is palmitoylated at residue cysteine 75 (Cys75). Mutation of Cys75 to serine (Ser) completely abolished AEG-1 palmitoylation. We identified ZDHHC6 as a palmitoyltransferase catalyzing the process in HEK293T cells. To obtain insight into how palmitoylation regulates AEG-1 function, we generated knock-in mice by CRISPR/Cas9 in which Cys75 of AEG-1 was mutated to Ser (AEG-1-C75S). No developmental or anatomical abnormality was observed between AEG-1-wild type (AEG-1-WT) and AEG-1-C75S littermates. However, global gene expression analysis by RNA-sequencing unraveled that signaling pathways and upstream regulators, which contribute to cell proliferation, motility, inflammation, angiogenesis, and lipid accumulation, were activated in AEG-1-C75S hepatocytes compared to AEG-1-WT. These findings suggest that AEG-1-C75S functions as dominant positive and that palmitoylation restricts oncogenic and NASH-promoting functions of AEG-1. We thus identify a previously unknown regulatory mechanism of AEG-1, which might help design new therapeutic strategies for NASH and HCC.

Yin Yang 1 controls cerebellar astrocyte maturation.

Diverse subpopulations of astrocytes tile different brain regions to accommodate local requirements of neurons and associated neuronal circuits. Nevertheless, molecular mechanisms governing astrocyte diversity remain mostly unknown. We explored the role of a zinc finger transcription factor Yin Yang 1 (YY1) that is expressed in astrocytes. We found that specific deletion of YY1 from astrocytes causes severe motor deficits in mice, induces Bergmann gliosis, and results in simultaneous loss of GFAP expression in velate and fibrous cerebellar astrocytes. Single cell RNA-seq analysis showed that YY1 exerts specific effects on gene expression in subpopulations of cerebellar astrocytes. We found that although YY1 is dispensable for the initial stages of astrocyte development, it regulates subtype-specific gene expression during astrocyte maturation. Moreover, YY1 is continuously needed to maintain mature astrocytes in the adult cerebellum. Our findings suggest that YY1 plays critical roles regulating cerebellar astrocyte maturation during development and maintaining a mature phenotype of astrocytes in the adult cerebellum.

preciseTAD: a transfer learning framework for 3D domain boundary prediction at base-pair resolution.

MOTIVATION: Chromosome conformation capture technologies (Hi-C) revealed extensive DNA folding into discrete 3D domains, such as Topologically Associating Domains and chromatin loops. The correct binding of CTCF and cohesin at domain boundaries is integral in maintaining the proper structure and function of these 3D domains. 3D domains have been mapped at the resolutions of 1 kilobase and above. However, it has not been possible to define their boundaries at the resolution of boundary-forming proteins. RESULTS: To predict domain boundaries at base-pair resolution, we developed preciseTAD, an optimized transfer learning framework trained on high-resolution genome annotation data. In contrast to current TAD/loop callers, preciseTAD-predicted boundaries are strongly supported by experimental evidence. Importantly, this approach can accurately delineate boundaries in cells without Hi-C data. preciseTAD provides a powerful framework to improve our understanding of how genomic regulators are shaping the 3D structure of the genome at base-pair resolution. AVAILABILITY AND IMPLEMENTATION: preciseTAD is an R/Bioconductor package available at https://bioconductor.org/packages/preciseTAD/. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

PDGFRα signaling drives adipose tissue fibrosis by targeting progenitor cell plasticity.

Fibrosis is a common disease process in which profibrotic cells disturb organ function by secreting disorganized extracellular matrix (ECM). Adipose tissue fibrosis occurs during obesity and is associated with metabolic dysfunction, but how profibrotic cells originate is still being elucidated. Here, we use a developmental model to investigate perivascular cells in white adipose tissue (WAT) and their potential to cause organ fibrosis. We show that a Nestin-Cre transgene targets perivascular cells (adventitial cells and pericyte-like cells) in WAT, and Nestin-GFP specifically labels pericyte-like cells. Activation of PDGFRα signaling in perivascular cells causes them to transition into ECM-synthesizing profibrotic cells. Before this transition occurs, PDGFRα signaling up-regulates mTOR signaling and ribosome biogenesis pathways and perturbs the expression of a network of epigenetically imprinted genes that have been implicated in cell growth and tissue homeostasis. Isolated Nestin-GFP(+) cells differentiate into adipocytes ex vivo and form WAT when transplanted into recipient mice. However, PDGFRα signaling opposes adipogenesis and generates profibrotic cells instead, which leads to fibrotic WAT in transplant experiments. These results identify perivascular cells as fibro/adipogenic progenitors in WAT and show that PDGFRα targets progenitor cell plasticity as a profibrotic mechanism.

Histone acetylation at the sulfotransferase 1a1 gene is associated with its hepatic expression in normal aging.

OBJECTIVES: Phase II drug metabolism is poorly studied in advanced age and older adults may exhibit significant variability in their expression of phase II enzymes. We hypothesized that age-related changes to epigenetic regulation of genes involved in phase II drug metabolism may contribute to these effects. METHODS: We examined published epigenome-wide studies of human blood and identified the SULT1A1 and UGT1A6 genes as the top loci showing epigenetic changes with age. To assess possible functional alterations with age in the liver, we assayed DNA methylation (5mC) and histone acetylation changes around the mouse homologs Sult1a1 and Ugt1a6 in liver tissue from mice aged 4-32 months. RESULTS: Our sample shows a significant loss of 5mC at Sult1a1 (ß = -1.08, 95% CI [-1.8, -0.2], SE = 0.38, P = 0.011), mirroring the loss of 5mC with age observed in human blood DNA at the same locus. We also detected increased histone 3 lysine 9 acetylation (H3K9ac) with age at Sult1a1 (ß = 0.11, 95% CI [0.002, 0.22], SE = 0.05, P = 0.04), but no change to histone 3 lysine 27 acetylation (H3K27ac). Sult1a1 gene expression is significantly positively associated with H3K9ac levels, accounting for 23% of the variation in expression. We did not detect any significant effects at Ugt1a6. CONCLUSIONS: Sult1a1 expression is under epigenetic influence in normal aging and this influence is more pronounced for H3K9ac than DNA methylation or H3K27ac in this study. More generally, our findings support the relevance of epigenetics in regulating key drug-metabolizing pathways. In the future, epigenetic biomarkers could prove useful to inform dosing in older adults.

Disease classification: from phenotypic similarity to integrative genomics and beyond.

A fundamental challenge of modern biomedical research is understanding how diseases that are similar on the phenotypic level are similar on the molecular level. Integration of various genomic data sets with the traditionally used phenotypic disease similarity revealed novel genetic and molecular mechanisms and blurred the distinction between monogenic (Mendelian) and complex diseases. Network-based medicine has emerged as a complementary approach for identifying disease-causing genes, genetic mediators, disruptions in the underlying cellular functions and for drug repositioning. The recent development of machine and deep learning methods allow for leveraging real-life information about diseases to refine genetic and phenotypic disease relationships. This review describes the historical development and recent methodological advancements for studying disease classification (nosology).

Coamplification of miR-4728 protects HER2-amplified breast cancers from targeted therapy.

HER2 (ERBB2) amplification is a driving oncogenic event in breast cancer. Clinical trials have consistently shown the benefit of HER2 inhibitors (HER2i) in treating patients with both local and advanced HER2+ breast cancer. Despite this benefit, their efficacy as single agents is limited, unlike the robust responses to other receptor tyrosine kinase inhibitors like EGFR inhibitors in EGFR-mutant lung cancer. Interestingly, the lack of HER2i efficacy occurs despite sufficient intracellular signaling shutdown following HER2i treatment. Exploring possible intrinsic causes for this lack of response, we uncovered remarkably depressed levels of NOXA, an endogenous inhibitor of the antiapoptotic MCL-1, in HER2-amplified breast cancer. Upon investigation of the mechanism leading to low NOXA, we identified a micro-RNA encoded in an intron of HER2, termed miR-4728, that targets the mRNA of the Estrogen Receptor α (ESR1). Reduced ESR1 expression in turn prevents ERα-mediated transcription of NOXA, mitigating apoptosis following treatment with the HER2i lapatinib. Importantly, resistance can be overcome with pharmacological inhibition of MCL-1. More generally, while many cancers like EGFR-mutant lung cancer are driven by activated kinases that when drugged lead to robust monotherapeutic responses, we demonstrate that the efficacy of targeted therapies directed against oncogenes active through focal amplification may be mitigated by coamplified genes.

SpectralTAD: an R package for defining a hierarchy of topologically associated domains using spectral clustering.

BACKGROUND: The three-dimensional (3D) structure of the genome plays a crucial role in gene expression regulation. Chromatin conformation capture technologies (Hi-C) have revealed that the genome is organized in a hierarchy of topologically associated domains (TADs), sub-TADs, and chromatin loops. Identifying such hierarchical structures is a critical step in understanding genome regulation. Existing tools for TAD calling are frequently sensitive to biases in Hi-C data, depend on tunable parameters, and are computationally inefficient. METHODS: To address these challenges, we developed a novel sliding window-based spectral clustering framework that uses gaps between consecutive eigenvectors for TAD boundary identification. RESULTS: Our method, implemented in an R package, SpectralTAD, detects hierarchical, biologically relevant TADs, has automatic parameter selection, is robust to sequencing depth, resolution, and sparsity of Hi-C data. SpectralTAD outperforms four state-of-the-art TAD callers in simulated and experimental settings. We demonstrate that TAD boundaries shared among multiple levels of the TAD hierarchy were more enriched in classical boundary marks and more conserved across cell lines and tissues. In contrast, boundaries of TADs that cannot be split into sub-TADs showed less enrichment and conservation, suggesting their more dynamic role in genome regulation. CONCLUSION: SpectralTAD is available on Bioconductor, http://bioconductor.org/packages/SpectralTAD/ .

Correction to: SpectralTAD: an R package for defining a hierarchy of topologically associated domains using spectral clustering.

An amendment to this paper has been published and can be accessed via the original article.

A method for estimating coherence of molecular mechanisms in major human disease and traits.

BACKGROUND: Phenotypes such as height and intelligence, are thought to be a product of the collective effects of multiple phenotype-associated genes and interactions among their protein products. High/low degree of interactions is suggestive of coherent/random molecular mechanisms, respectively. Comparing the degree of interactions may help to better understand the coherence of phenotype-specific molecular mechanisms and the potential for therapeutic intervention. However, direct comparison of the degree of interactions is difficult due to different sizes and configurations of phenotype-associated gene networks. METHODS: We introduce a metric for measuring coherence of molecular-interaction networks as a slope of internal versus external distributions of the degree of interactions. The internal degree distribution is defined by interaction counts within a phenotype-specific gene network, while the external degree distribution counts interactions with other genes in the whole protein-protein interaction (PPI) network. We present a novel method for normalizing the coherence estimates, making them directly comparable. RESULTS: Using STRING and BioGrid PPI databases, we compared the coherence of 116 phenotype-associated gene sets from GWAScatalog against size-matched KEGG pathways (the reference for high coherence) and random networks (the lower limit of coherence). We observed a range of coherence estimates for each category of phenotypes. Metabolic traits and diseases were the most coherent, while psychiatric disorders and intelligence-related traits were the least coherent. We demonstrate that coherence and modularity measures capture distinct network properties. CONCLUSIONS: We present a general-purpose method for estimating and comparing the coherence of molecular-interaction gene networks that accounts for the network size and shape differences. Our results highlight gaps in our current knowledge of genetics and molecular mechanisms of complex phenotypes and suggest priorities for future GWASs.

Anillin regulates breast cancer cell migration, growth, and metastasis by non-canonical mechanisms involving control of cell stemness and differentiation.

BACKGROUND: Breast cancer metastasis is driven by a profound remodeling of the cytoskeleton that enables efficient cell migration and invasion. Anillin is a unique scaffolding protein regulating major cytoskeletal structures, such as actin filaments, microtubules, and septin polymers. It is markedly overexpressed in breast cancer, and high anillin expression is associated with poor prognosis. The aim of this study was to investigate the role of anillin in breast cancer cell migration, growth, and metastasis. METHODS: CRISPR/Cas9 technology was used to deplete anillin in highly metastatic MDA-MB-231 and BT549 cells and to overexpress it in poorly invasive MCF10AneoT cells. The effects of anillin depletion and overexpression on breast cancer cell motility in vitro were examined by wound healing and Matrigel invasion assays. Assembly of the actin cytoskeleton and matrix adhesion were evaluated by immunofluorescence labeling and confocal microscopy. In vitro tumor development was monitored by soft agar growth assays, whereas cancer stem cells were examined using a mammosphere formation assay and flow cytometry. The effects of anillin knockout on tumor growth and metastasis in vivo were determined by injecting control and anillin-depleted breast cancer cells into NSG mice. RESULTS: Loss-of-function and gain-of-function studies demonstrated that anillin is necessary and sufficient to accelerate migration, invasion, and anchorage-independent growth of breast cancer cells in vitro. Furthermore, loss of anillin markedly attenuated primary tumor growth and metastasis of breast cancer in vivo. In breast cancer cells, anillin was localized in the nucleus; however, knockout of this protein affected the cytoplasmic/cortical events, e.g., the organization of actin cytoskeleton and cell-matrix adhesions. Furthermore, we observed a global transcriptional reprogramming of anillin-depleted breast cancer cells that resulted in suppression of their stemness and induction of the mesenchymal to epithelial trans-differentiation. Such trans-differentiation was manifested by the upregulation of basal keratins along with the increased expression of E-cadherin and P-cadherin. Knockdown of E-cadherin restored the impaired migration and invasion of anillin-deficient breast cancer cells. CONCLUSION: Our study demonstrates that anillin plays essential roles in promoting breast cancer growth and metastatic dissemination in vitro and in vivo and unravels novel functions of anillin in regulating breast cancer stemness and differentiation.

multiHiCcompare: joint normalization and comparative analysis of complex Hi-C experiments.

MOTIVATION: With the development of chromatin conformation capture technology and its high-throughput derivative Hi-C sequencing, studies of the three-dimensional interactome of the genome that involve multiple Hi-C datasets are becoming available. To account for the technology-driven biases unique to each dataset, there is a distinct need for methods to jointly normalize multiple Hi-C datasets. Previous attempts at removing biases from Hi-C data have made use of techniques which normalize individual Hi-C datasets, or, at best, jointly normalize two datasets. RESULTS: Here, we present multiHiCcompare, a cyclic loess regression-based joint normalization technique for removing biases across multiple Hi-C datasets. In contrast to other normalization techniques, it properly handles the Hi-C-specific decay of chromatin interaction frequencies with the increasing distance between interacting regions. multiHiCcompare uses the general linear model framework for comparative analysis of multiple Hi-C datasets, adapted for the Hi-C-specific decay of chromatin interaction frequencies. multiHiCcompare outperforms other methods when detecting a priori known chromatin interaction differences from jointly normalized datasets. Applied to the analysis of auxin-treated versus untreated experiments, and CTCF depletion experiments, multiHiCcompare was able to recover the expected epigenetic and gene expression signatures of loss of chromatin interactions and reveal novel insights. AVAILABILITY AND IMPLEMENTATION: multiHiCcompare is freely available on GitHub and as a Bioconductor R package https://bioconductor.org/packages/multiHiCcompare. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Identification of breast cancer associated variants that modulate transcription factor binding.

Genome-wide association studies (GWAS) have discovered thousands loci associated with disease risk and quantitative traits, yet most of the variants responsible for risk remain uncharacterized. The majority of GWAS-identified loci are enriched for non-coding single-nucleotide polymorphisms (SNPs) and defining the molecular mechanism of risk is challenging. Many non-coding causal SNPs are hypothesized to alter transcription factor (TF) binding sites as the mechanism by which they affect organismal phenotypes. We employed an integrative genomics approach to identify candidate TF binding motifs that confer breast cancer-specific phenotypes identified by GWAS. We performed de novo motif analysis of regulatory elements, analyzed evolutionary conservation of identified motifs, and assayed TF footprinting data to identify sequence elements that recruit TFs and maintain chromatin landscape in breast cancer-relevant tissue and cell lines. We identified candidate causal SNPs that are predicted to alter TF binding within breast cancer-relevant regulatory regions that are in strong linkage disequilibrium with significantly associated GWAS SNPs. We confirm that the TFs bind with predicted allele-specific preferences using CTCF ChIP-seq data. We used The Cancer Genome Atlas breast cancer patient data to identify ANKLE1 and ZNF404 as the target genes of candidate TF binding site SNPs in the 19p13.11 and 19q13.31 GWAS-identified loci. These SNPs are associated with the expression of ZNF404 and ANKLE1 in breast tissue. This integrative analysis pipeline is a general framework to identify candidate causal variants within regulatory regions and TF binding sites that confer phenotypic variation and disease risk.

Separation of breast cancer and organ microenvironment transcriptomes in metastases.

BACKGROUND: The seed and soil hypothesis was proposed over a century ago to describe why cancer cells (seeds) grow in certain organs (soil). Since then, the genetic properties that define the cancer cells have been heavily investigated; however, genomic mediators within the organ microenvironment that mediate successful metastatic growth are less understood. These studies sought to identify cancer- and organ-specific genomic programs that mediate metastasis. METHODS: In these studies, a set of 14 human breast cancer patient-derived xenograft (PDX) metastasis models was developed and then tested for metastatic tropism with two approaches: spontaneous metastases from mammary tumors and intravenous injection of PDX cells. The transcriptomes of the cancer cells when growing as tumors or metastases were separated from the transcriptomes of the microenvironment via species-specific separation of the genomes. Drug treatment of PDX spheroids was performed to determine if genes activated in metastases may identify targetable mediators of viability. RESULTS: The experimental approaches that generated metastases in PDX models were identified. RNA sequencing of 134 tumors, metastases, and normal non-metastatic organs identified cancer- and organ-specific genomic properties that mediated metastasis. A common genomic response of the liver microenvironment was found to occur in reaction to the invading PDX cells. Genes within the cancer cells were found to be either transiently regulated by the microenvironment or permanently altered due to clonal selection of metastatic sublines. Gene Set Enrichment Analyses identified more than 400 gene signatures that were commonly activated in metastases across basal-like PDXs. A Src signaling signature was found to be extensively upregulated in metastases, and Src inhibitors were found to be cytotoxic to PDX spheroids. CONCLUSIONS: These studies identified that during the growth of breast cancer metastases, there were genomic changes that occurred within both the cancer cells and the organ microenvironment. We hypothesize that pathways upregulated in metastases are mediators of viability and that simultaneously targeting changes within different cancer cell pathways and/or different tissue compartments may be needed for inhibition of disease progression.

HiCcompare: an R-package for joint normalization and comparison of HI-C datasets.

BACKGROUND: Changes in spatial chromatin interactions are now emerging as a unifying mechanism orchestrating the regulation of gene expression. Hi-C sequencing technology allows insight into chromatin interactions on a genome-wide scale. However, Hi-C data contains many DNA sequence- and technology-driven biases. These biases prevent effective comparison of chromatin interactions aimed at identifying genomic regions differentially interacting between, e.g., disease-normal states or different cell types. Several methods have been developed for normalizing individual Hi-C datasets. However, they fail to account for biases between two or more Hi-C datasets, hindering comparative analysis of chromatin interactions. RESULTS: We developed a simple and effective method, HiCcompare, for the joint normalization and differential analysis of multiple Hi-C datasets. The method introduces a distance-centric analysis and visualization of the differences between two Hi-C datasets on a single plot that allows for a data-driven normalization of biases using locally weighted linear regression (loess). HiCcompare outperforms methods for normalizing individual Hi-C datasets and methods for differential analysis (diffHiC, FIND) in detecting a priori known chromatin interaction differences while preserving the detection of genomic structures, such as A/B compartments. CONCLUSIONS: HiCcompare is able to remove between-dataset bias present in Hi-C matrices. It also provides a user-friendly tool to allow the scientific community to perform direct comparisons between the growing number of pre-processed Hi-C datasets available at online repositories. HiCcompare is freely available as a Bioconductor R package https://bioconductor.org/packages/HiCcompare/ .

Characterizing the efficacy of cancer therapeutics in patient-derived xenograft models of metastatic breast cancer.

PURPOSE: Basal-like breast cancers are aggressive and often metastasize to vital organs. Treatment is largely limited to chemotherapy. This study aims to characterize the efficacy of cancer therapeutics in vitro and in vivo within the primary tumor and metastatic setting, using patient-derived xenograft (PDX) models. METHODS: We employed two basal-like, triple-negative PDX models, WHIM2 and WHIM30. PDX cells, obtained from mammary tumors grown in mice, were treated with twelve cancer therapeutics to evaluate their cytotoxicity in vitro. Four of the effective drugs-carboplatin, cyclophosphamide, bortezomib, and dacarbazine-were tested in vivo for their efficacy in treating mammary tumors, and metastases generated by intracardiac injection of tumor cells. RESULTS: RNA sequencing showed that global gene expression of PDX cells grown in the mammary gland was similar to those tested in culture. In vitro, carboplatin was cytotoxic to WHIM30 but not WHIM2, whereas bortezomib, dacarbazine, and cyclophosphamide were cytotoxic to both lines. Yet, these drugs were ineffective in treating both primary and metastatic WHIM2 tumors in vivo. Carboplatin and cyclophosphamide were effective in treating WHIM30 mammary tumors and reducing metastatic burden in the brain, liver, and lungs. WHIM2 and WHIM30 metastases showed distinct patterns of cytokeratin and vimentin expression, regardless of treatment, suggesting that different tumor cell subpopulations may preferentially seed in different organs. CONCLUSIONS: This study highlights the utility of PDX models for studying the efficacy of therapeutics in reducing metastatic burden in specific organs. The differential treatment responses between two PDX models of the same intrinsic subtype, in both the primary and metastatic setting, recapitulates the challenges faced in treating cancer patients and highlights the need for combination therapies and predictive biomarkers.