Lydia Hopp

A Global Genome Segmentation Method for Exploration of Epigenetic Patterns

Current genome-wide ChIP-seq experiments on different epigenetic marks aim at unraveling the interplay between their regulation mechanisms. Published evaluation tools, however, allow testing for predefined hypotheses only. Here, we present a novel method for annotation-independent exploration of epigenetic data and their inter-correlation with other genome-wide features. Our method is based on a combinatorial genome segmentation solely using information on combinations of epigenetic marks. It does not require prior knowledge about the data (e.g. gene positions), but allows integrating the data in a straightforward manner. Thereby, it combines compression, clustering and visualization of the data in a single tool. Our method provides intuitive maps of epigenetic patterns across multiple levels of organization, e.g. of the co-occurrence of different epigenetic marks in different cell types. Thus, it facilitates the formulation of new hypotheses on the principles of epigenetic regulation. We apply our method to histone modification data on trimethylation of histone H3 at lysine 4, 9 and 27 in multi-potent and lineage-primed mouse cells, analyzing their combinatorial modification pattern as well as differentiation-related changes of single modifications. We demonstrate that our method is capable of reproducing recent findings of gene centered approaches, e.g. correlations between CpG-density and the analyzed histone modifications. Moreover, combining the clustered epigenetic data with information on the expression status of associated genes we classify differences in epigenetic status of e.g. house-keeping genes versus differentiation-related genes. Visualizing the distribution of modification states on the chromosomes, we discover strong patterns for chromosome X. For example, exclusively H3K9me3 marked segments are enriched, while poised and active states are rare. Hence, our method also provides new insights into chromosome-specific epigenetic patterns, opening up new questions how “epigenetic computation” is distributed over the genome in space and time.

Mapping heterogeneity in patient-derived melanoma cultures by single-cell RNA-seq

Recent technological advances in single-cell genomics make it possible to analyze cellular heterogeneity of tumor samples. Here, we applied single-cell RNA-seq to measure the transcriptomes of 307 single cells cultured from three biopsies of three different patients with a BRAF/NRAS wild type, BRAF mutant/NRAS wild type and BRAF wild type/NRAS mutant melanoma metastasis, respectively. Analysis based on self-organizing maps identified sub-populations defined by multiple gene expression modules involved in proliferation, oxidative phosphorylation, pigmentation and cellular stroma. Gene expression modules had prognostic relevance when compared with gene expression data from published melanoma samples and patient survival data. We surveyed kinome expression patterns across sub-populations of the BRAF/NRAS wild type sample and found that CDK4 and CDK2 were consistently highly expressed in the majority of cells, suggesting that these kinases might be involved in melanoma progression. Treatment of cells with the CDK4 inhibitor palbociclib restricted cell proliferation to a similar, and in some cases greater, extent than MAPK inhibitors. Finally, we identified a low abundant sub-population in this sample that highly expressed a module containing ABC transporter ABCB5, surface markers CD271 and CD133, and multiple aldehyde dehydrogenases (ALDHs). Patient-derived cultures of the BRAF mutant/NRAS wild type and BRAF wild type/NRAS mutant metastases showed more homogeneous single-cell gene expression patterns with gene expression modules for proliferation and ABC transporters. Taken together, our results describe an intertumor and intratumor heterogeneity in melanoma short-term cultures which might be relevant for patient survival, and suggest promising targets for new treatment approaches in melanoma therapy.

Epigenetic Heterogeneity of B-Cell Lymphoma: DNA Methylation, Gene Expression and Chromatin States

Mature B-cell lymphoma is a clinically and biologically highly diverse disease. Its diagnosis and prognosis is a challenge due to its molecular heterogeneity and diverse regimes of biological dysfunctions, which are partly driven by epigenetic mechanisms. We here present an integrative analysis of DNA methylation and gene expression data of several lymphoma subtypes. Our study confirms previous results about the role of stemness genes during development and maturation of B-cells and their dysfunction in lymphoma locking in more proliferative or immune-reactive states referring to B-cell functionalities in the dark and light zone of the germinal center and also in plasma cells. These dysfunctions are governed by widespread epigenetic effects altering the promoter methylation of the involved genes, their activity status as moderated by histone modifications and also by chromatin remodeling. We identified four groups of genes showing characteristic expression and methylation signatures among Burkitt’s lymphoma, diffuse large B cell lymphoma, follicular lymphoma and multiple myeloma. These signatures are associated with epigenetic effects such as remodeling from transcriptionally inactive into active chromatin states, differential promoter methylation and the enrichment of targets of transcription factors such as EZH2 and SUZ12.

Genome-wide DNA promoter methylation and transcriptome analysis in human adipose tissue unravels novel candidate genes for obesity

Objective/methods DNA methylation plays an important role in obesity and related metabolic complications. We examined genome-wide DNA promoter methylation along with mRNA profiles in paired samples of human subcutaneous adipose tissue (SAT) and omental visceral adipose tissue (OVAT) from non-obese vs. obese individuals. Results We identified negatively correlated methylation and expression of several obesity-associated genes in our discovery dataset and in silico replicated ETV6 in two independent cohorts. Further, we identified six adipose tissue depot-specific genes (HAND2, HOXC6, PPARG, SORBS2, CD36, and CLDN1). The effects were further supported in additional independent cohorts. Our top hits might play a role in adipogenesis and differentiation, obesity, lipid metabolism, and adipose tissue expandability. Finally, we show that in vitro methylation of SORBS2 directly represses gene expression. Conclusions Taken together, our data show distinct tissue specific epigenetic alterations which associate with obesity.

Portraying the expression landscapes of cancer subtypes

Self-organizing maps (SOM) portray molecular phenotypes with individual resolution. We present an analysis pipeline based on SOM machine learning which allows the comprehensive study of large scale clinical data. The potency of the method is demonstrated in selected applications studying the diversity of gene expression in Glioblastoma Multiforme (GBM) and prostate cancer progression. Our method characterizes relationships between the samples, disentangles the expression patterns into well separated groups of co-regulated genes, extracts their functional contexts using enrichment techniques, and enables the detection of contaminations and outliers in the samples. We found that the four GBM subtypes can be divided into two “localized” and two “intermediate” ones. The localized subtypes are characterized by the antagonistic activation of processes related to immune response and cell division, commonly observed also in other cancers. In contrast, each of the “intermediate” subtypes forms a heterogeneous continuum of expression states linking the “localized” subtypes. Both “intermediate” subtypes are characterized by distinct expression patterns related to translational activity and innate immunity as well as nervous tissue and cell function. We show that SOM portraits provide a comprehensive framework for the description of the diversity of expression landscapes using concepts of molecular function.

Portraying the Expression Landscapes of B-CellLymphoma-Intuitive Detection of Outlier Samples and of Molecular Subtypes

We present an analytic framework based on Self-Organizing Map (SOM) machine learning to study large scale patient data sets. The potency of the approach is demonstrated in a case study using gene expression data of more than 200 mature aggressive B-cell lymphoma patients. The method portrays each sample with individual resolution, characterizes the subtypes, disentangles the expression patterns into distinct modules, extracts their functional context using enrichment techniques and enables investigation of the similarity relations between the samples. The method also allows to detect and to correct outliers caused by contaminations. Based on our analysis, we propose a refined classification of B-cell Lymphoma into four molecular subtypes which are characterized by differential functional and clinical characteristics.

Genomic and transcriptomic heterogeneity of colorectal tumors arising in Lynch Syndrome

Colorectal cancer (CRC) arising in Lynch syndrome (LS) comprises tumours with constitutional mutations in DNA mismatch repair genes. There is still a lack of whole‐genome and transcriptome studies of LS‐CRC to address questions about similarities and differences in mutation and gene expression characteristics between LS‐CRC and sporadic CRC, about the molecular heterogeneity of LS‐CRC, and about specific mechanisms of LS‐CRC genesis linked to dysfunctional mismatch repair in LS colonic mucosa and the possible role of immune editing. Here, we provide a first molecular characterization of LS tumours and of matched tumour‐distant reference colonic mucosa based on whole‐genome DNA‐sequencing and RNA‐sequencing analyses. Our data support two subgroups of LS‐CRCs, G1 and G2, whereby G1 tumours show a higher number of somatic mutations, a higher amount of microsatellite slippage, and a different mutation spectrum. The gene expression phenotypes support this difference. Reference mucosa of G1 shows a strong immune response associated with the expression of HLA and immune checkpoint genes and the invasion of CD4+ T cells. Such an immune response is not observed in LS tumours, G2 reference and normal (non‐Lynch) mucosa, and sporadic CRC. We hypothesize that G1 tumours are edited for escape from a highly immunogenic microenvironment via loss of HLA presentation and T‐cell exhaustion. In contrast, G2 tumours seem to develop in a less immunogenic microenvironment where tumour‐promoting inflammation parallels tumourigenesis. Larger studies on non‐neoplastic mucosa tissue of mutation carriers are required to better understand the early phases of emerging tumours. Copyright

Bistable Epigenetic States Explain Age‐Dependent Decline in Mesenchymal Stem Cell Heterogeneity

The molecular mechanisms by which heterogeneity, a major characteristic of stem cells, is achieved are yet unclear. We here study the expression of the membrane stem cell antigen‐1 (Sca‐1) in mouse bone marrow mesenchymal stem cell (MSC) clones. We show that subpopulations with varying Sca‐1 expression profiles regenerate the Sca‐1 profile of the mother population within a few days. However, after extensive replication in vitro, the expression profiles shift to lower values and the regeneration time increases. Study of the promoter of Ly6a unravels that the expression level of Sca‐1 is related to the promoter occupancy by the activating histone mark H3K4me3. We demonstrate that these findings can be consistently explained by a computational model that considers positive feedback between promoter H3K4me3 modification and gene transcription. This feedback implicates bistable epigenetic states which the cells occupy with an age‐dependent frequency due to persistent histone (de‐)modification. Our results provide evidence that MSC heterogeneity, and presumably that of other stem cells, is associated with bistable epigenetic states and suggest that MSCs are subject to permanent state fluctuations. Stem Cells 2017;35:694–704

Analysis of MicroRNA Expression Using Machine Learning

The systematic analysis of miRNA expression and its potential mRNA targets constitutes a basal objective in miRNA research in addition to miRNA gene detection and miRNA target prediction. In this chapter we address methodical issues of miRNA expression analysis using self-organizing maps (SOM), a neural network machine learning algorithm with strong visualization and second-level analysis capabilities widely used to categorize large-scale, high-dimensional data. We shortly review selected experimental and theoretical aspects of miRNA expression analysis. Then, the protocol of our SOM method is outlined with special emphasis on miRNA/mRNA coexpression. The method allows extracting differentially expressed RNA transcripts, their functional context, and also characterization of global properties of expression states and profiles. In addition to the separate study of miRNA and mRNA expression landscapes, we propose the combined analysis of both entities using a covariance SOM.

MicroRNA Expression Landscapes in Stem Cells, Tissues, and Cancer

MicroRNAs play critical roles in the regulation of gene expression with two major functions: marking mRNA for degradation in a sequence-specific manner or repressing translation. Publicly available data sets on miRNA and mRNA expression in embryonal and induced stem cells, human tissues, and solid tumors are analyzed in this case study using self-organizing maps (SOMs) to characterize miRNA expression landscapes in the context of cell fate commitment, tissue-specific differentiation, and its dysfunction in cancer. The SOM portraits of the individual samples clearly reveal groups of miRNA specifically overexpressed without the need of additional pairwise comparisons between the different systems. Sets of miRNA differentially over- and underexpressed in different systems have been detected in this study. The individual portraits of the expression landscapes enable a very intuitive, image-based perception which clearly promotes the discovery of qualitative relationships between the systems studied. We see perspectives for broad applications of this method in standard analysis to many kinds of high-throughput data of single miRNA and especially combined miRNA/mRNA data sets.