Dynamics of broad H3K4me3 domains uncover an epigenetic switch between cell identity and cancer-related genes.

Broad domains of H3K4 methylation have been associated with consistent expression of tissue-specific, cell identity, and tumor suppressor genes. Here, we identified broad domain-associated genes in healthy human thymic T cell populations and a collection of T cell acute lymphoblastic leukemia (T-ALL) primary samples and cell lines. We found that broad domains are highly dynamic throughout T cell differentiation, and their varying breadth allows the distinction between normal and neoplastic cells. Although broad domains preferentially associate with cell identity and tumor suppressor genes in normal thymocytes, they flag key oncogenes in T-ALL samples. Moreover, the expression of broad domain-associated genes, both coding and noncoding, is frequently deregulated in T-ALL. Using two distinct leukemic models, we showed that the ectopic expression of T-ALL oncogenic transcription factor preferentially impacts the expression of broad domain-associated genes in preleukemic cells. Finally, an H3K4me3 demethylase inhibitor differentially targets T-ALL cell lines depending on the extent and number of broad domains. Our results show that the regulation of broad H3K4me3 domains is associated with leukemogenesis, and suggest that the presence of these structures might be used for epigenetic prioritization of cancer-relevant genes, including long noncoding RNAs.

Histone Variant H2A.L.2 Guides Transition Protein-Dependent Protamine Assembly in Male Germ Cells.

Histone replacement by transition proteins (TPs) and protamines (Prms) constitutes an essential step for the successful production of functional male gametes, yet nothing is known on the underlying functional interplay between histones, TPs, and Prms. Here, by studying spermatogenesis in the absence of a spermatid-specific histone variant, H2A.L.2, we discover a fundamental mechanism involved in the transformation of nucleosomes into nucleoprotamines. H2A.L.2 is synthesized at the same time as TPs and enables their loading onto the nucleosomes. TPs do not displace histones but rather drive the recruitment and processing of Prms, which are themselves responsible for histone eviction. Altogether, the incorporation of H2A.L.2 initiates and orchestrates a series of successive transitional states that ultimately shift to the fully compacted genome of the mature spermatozoa. Hence, the current view of histone-to-nucleoprotamine transition should be revisited and include an additional step with H2A.L.2 assembly prior to the action of TPs and Prms.

Dynamic Competing Histone H4 K5K8 Acetylation and Butyrylation Are Hallmarks of Highly Active Gene Promoters.

Recently discovered histone lysine acylation marks increase the functional diversity of nucleosomes well beyond acetylation. Here, we focus on histone butyrylation in the context of sperm cell differentiation. Specifically, we investigate the butyrylation of histone H4 lysine 5 and 8 at gene promoters where acetylation guides the binding of Brdt, a bromodomain-containing protein, thereby mediating stage-specific gene expression programs and post-meiotic chromatin reorganization. Genome-wide mapping data show that highly active Brdt-bound gene promoters systematically harbor competing histone acetylation and butyrylation marks at H4 K5 and H4 K8. Despite acting as a direct stimulator of transcription, histone butyrylation competes with acetylation, especially at H4 K5, to prevent Brdt binding. Additionally, H4 K5K8 butyrylation also marks retarded histone removal during late spermatogenesis. Hence, alternating H4 acetylation and butyrylation, while sustaining direct gene activation and dynamic bromodomain binding, could impact the final male epigenome features.

Widespread Enhancer Activity from Core Promoters.

Gene expression in higher eukaryotes is precisely regulated in time and space through the interplay between promoters and gene-distal regulatory regions, known as enhancers. The original definition of enhancers implies the ability to activate gene expression remotely, while promoters entail the capability to locally induce gene expression. Despite the conventional distinction between them, promoters and enhancers share many genomic and epigenomic features. One intriguing finding in the gene regulation field comes from the observation that many core promoter regions display enhancer activity. Recent high-throughput reporter assays along with clustered regularly interspaced short palindromic repeats (CRISPR)/Cas9-related approaches have indicated that this phenomenon is common and might have a strong impact on our global understanding of genome organisation and gene expression regulation.

Anomaly detection in genomic catalogues using unsupervised multi-view autoencoders.

BACKGROUND: Accurate identification of Transcriptional Regulator binding locations is essential for analysis of genomic regions, including Cis Regulatory Elements. The customary NGS approaches, predominantly ChIP-Seq, can be obscured by data anomalies and biases which are difficult to detect without supervision. RESULTS: Here, we develop a method to leverage the usual combinations between many experimental series to mark such atypical peaks. We use deep learning to perform a lossy compression of the genomic regions' representations with multiview convolutions. Using artificial data, we show that our method correctly identifies groups of correlating series and evaluates CRE according to group completeness. It is then applied to the ReMap database's large volume of curated ChIP-seq data. We show that peaks lacking known biological correlators are singled out and less confirmed in real data. We propose normalization approaches useful in interpreting black-box models. CONCLUSION: Our approach detects peaks that are less corroborated than average. It can be extended to other similar problems, and can be interpreted to identify correlation groups. It is implemented in an open-source tool called atyPeak.

Genetic Predisposition to the Mortality in Septic Shock Patients: From GWAS to the Identification of a Regulatory Variant Modulating the Activity of a CISH Enhancer.

The high mortality rate in septic shock patients is likely due to environmental and genetic factors, which influence the host response to infection. Two genome-wide association studies (GWAS) on 832 septic shock patients were performed. We used integrative bioinformatic approaches to annotate and prioritize the sepsis-associated single nucleotide polymorphisms (SNPs). An association of 139 SNPs with death based on a false discovery rate of 5% was detected. The most significant SNPs were within the CISH gene involved in cytokine regulation. Among the 139 SNPs associated with death and the 1311 SNPs in strong linkage disequilibrium with them, we investigated 1439 SNPs within non-coding regions to identify regulatory variants. The highest integrative weighted score (IW-score) was obtained for rs143356980, indicating that this SNP is a robust regulatory candidate. The rs143356980 region is located in a non-coding region close to the CISH gene. A CRISPR-Cas9-mediated deletion of this region and specific luciferase assays in K562 cells showed that rs143356980 modulates the enhancer activity in K562 cells. These analyses allowed us to identify several genes associated with death in patients with septic shock. They suggest that genetic variations in key genes, such as CISH, perturb relevant pathways, increasing the risk of death in sepsis patients.

A demonstration of the H3 trimethylation ChIP-seq analysis of galline follicular mesenchymal cells and male germ cells.

OBJECTIVE: Trimethylation of histone 3 (H3) at 4th lysine N-termini (H3K4me3) in gene promoter region was the universal marker of active genes specific to cell lineage. On the contrary, coexistence of trimethylation at 27th lysine (H3K27me3) in the same loci-the bivalent H3K4m3/H3K27me3 was known to suspend the gene transcription in germ cells, and could also be inherited to the developed stem cell. In galline species, throughout example of H3K4m3 and H3K27me3 ChIP-seq analysis was still not provided. We therefore designed and demonstrated such procedures using ChIP-seq and mRNA-seq data of chicken follicular mesenchymal cells and male germ cells. METHODS: Analytical workflow was designed and provided in this study. ChIP-seq and RNA-seq datasets of follicular mesenchymal cells and male germ cells were acquired and properly preprocessed. Peak calling by Model-based analysis of ChIP-seq 2 was performed to identify H3K4m3 or H3K27me3 enriched regions (Fold-change≥2, FDR≤0.01) in gene promoter regions. Integrative genomics viewer was utilized for cellular retinoic acid binding protein 1 (CRABP1), growth differentiation factor 10 (GDF10), and gremlin 1 (GREM1) gene explorations. RESULTS: The acquired results indicated that follicular mesenchymal cells and germ cells shared several unique gene promoter regions enriched with H3K4me3 (5,704 peaks) and also unique regions of bivalent H3K4m3/H3K27me3 shared between all cell types and germ cells (1,909 peaks). Subsequent observation of follicular mesenchyme-specific genes-CRABP1, GDF10, and GREM1 correctly revealed vigorous transcriptions of these genes in follicular mesenchymal cells. As expected, bivalent H3K4m3/H3K27me3 pattern was manifested in gene promoter regions of germ cells, and thus suspended their transcriptions. CONCLUSION: According the results, an example of chicken H3K4m3/H3K27me3 ChIP-seq data analysis was successfully demonstrated in this study. Hopefully, the provided methodology should hereby be useful for galline ChIP-seq data analysis in the future.

An update on recent methods applied for deciphering the diversity of the noncoding RNA genome structure and function.

The explosion of high throughput sequencing technologies marked a turn in our way of understanding the complexity and diversity of the transcriptome, including noncoding transcription dependent on RNA polymerase II. Many new ncRNA populations were described in recent years, including for example TSS RNAs, lincRNAs, eRNAs, PROMPTS and several others. Besides the advances in the average depth coverage of RNA-seq experiments, various additional protocols are now available that can be used to address qualitative and quantitative aspects of the noncoding transcriptome complexity and function. In this review, we will focus on methods allowing isolation and characterization of complex RNA populations using sequencing based approaches, including conventional strategies already used for coding genome and more specific developments allowing, for example, the study of nascent strand transcription, protein-bound or structured RNAs.

[Functions of lncRNA in development and diseases]. / Rôle des longs ARN non codants dans le développement normal et pathologique.

The transcription of essentially the entire eukaryotic genome generates a myriad of non-coding RNA species that show complex overlapping patterns of expression and regulation. In the last decade, several large scale genomic analyses have shed light on the widespread existence of long non-coding RNAs (lncRNAs) in mammals. Although the function of most lncRNAs remains unknown, many of them have been suggested to play important roles in the regulation of gene expression during normal development and diseases, including cancers. Indeed, functional studies have demonstrated that lncRNAs participate in various biological processes, including reprogramming of pluripotent stem cells, oncogenic progression and cell cycle regulation. In this review, we summarize recent findings about the biology of lncRNAs and their functions in normal and pathological development in mammals.

Differential transcript usage across mammalian oocytes at the germinal vesicle and metaphase II stages.

Ongoing progress in mRNA-Sequencing technologies has significantly contributed to the refinement of assisted reproductive technologies. However, the prior investigations have predominantly concentrated on alterations in overall gene expression levels, thereby leaving a considerable gap in our understanding of the influence of transcript isoform expression on fundamental cellular mechanisms of oocytes. Given the efficacy of differential transcript usage (DTU) analysis to address such knowledge, we conducted comprehensive DTU analysis utilizing mRNA-Seq datasets of germinal vesicle (GV) and metaphase II (MII) oocytes across six mammalian species from the SRA database, including cow, donkey, horse, human, mouse, and pig. To further illuminate the roles of these genes, we also conducted a rigorous Gene Ontology (GO) term enrichment analysis. While the DTU analysis of each species exhibited several genes with alterations in their transcript isoform usage, referred to as DTU genes, this study focused on only ten cross-species DTU genes sharing among a minimum of five distinct species (FDR≤0.05). These cross-species DTU genes were as follows: ABCF1, CDC6, CFAP36, CNOT10, DNM3, IWS1, NBN, NDEL1, RAD50 and ZCCHC17. GO term enrichment analysis unveiled the alignment of these cross-species DTU gene functions with RNA and cell-cycle control mechanisms across diverse mammalian species, thereby suggesting their vital roles during oocyte maturation. Further exploration of the transcript isoforms of these genes hence bore the potential to uncover novel transcript isoform markers for future reproductive technologies in both human and animal contexts.

Divergent transcription is associated with promoters of transcriptional regulators.

BACKGROUND: Divergent transcription is a wide-spread phenomenon in mammals. For instance, short bidirectional transcripts are a hallmark of active promoters, while longer transcripts can be detected antisense from active genes in conditions where the RNA degradation machinery is inhibited. Moreover, many described long non-coding RNAs (lncRNAs) are transcribed antisense from coding gene promoters. However, the general significance of divergent lncRNA/mRNA gene pair transcription is still poorly understood. Here, we used strand-specific RNA-seq with high sequencing depth to thoroughly identify antisense transcripts from coding gene promoters in primary mouse tissues. RESULTS: We found that a substantial fraction of coding-gene promoters sustain divergent transcription of long non-coding RNA (lncRNA)/mRNA gene pairs. Strikingly, upstream antisense transcription is significantly associated with genes related to transcriptional regulation and development. Their promoters share several characteristics with those of transcriptional developmental genes, including very large CpG islands, high degree of conservation and epigenetic regulation in ES cells. In-depth analysis revealed a unique GC skew profile at these promoter regions, while the associated coding genes were found to have large first exons, two genomic features that might enforce bidirectional transcription. Finally, genes associated with antisense transcription harbor specific H3K79me2 epigenetic marking and RNA polymerase II enrichment profiles linked to an intensified rate of early transcriptional elongation. CONCLUSIONS: We concluded that promoters of a class of transcription regulators are characterized by a specialized transcriptional control mechanism, which is directly coupled to relaxed bidirectional transcription.

Cross-species analysis of differential transcript usage in humans and chickens with fatty liver disease.

Background and Aim: Fatty liver disease is a common condition, characterized by excess fat accumulation in the liver. It can contribute to more severe liver-related health issues, making it a critical concern in avian and human medicine. Apart from modifying the gene expression of liver cells, the disease also alters the expression of specific transcript isoforms, which might serve as new biological markers for both species. This study aimed to identify cross-species genes displaying differential expressions in their transcript isoforms in humans and chickens with fatty liver disease. Materials and Methods: We performed differential gene expression and differential transcript usage (DTU) analyses on messenger RNA datasets from the livers of both chickens and humans with fatty liver disease. Using appropriate cross-species gene identification methods, we reviewed the acquired candidate genes and their transcript isoforms to determine their potential role in fatty liver disease's pathogenesis. Results: We identified seven genes - ALG5, BRD7, DIABLO, RSU1, SFXN5, STIMATE, TJP3, and VDAC2 - and their corresponding transcript isoforms as potential candidates (false discovery rate ≤0.05). Our findings showed that these genes most likely contribute to fatty disease development and progression. Conclusion: This study successfully identified novel human-chicken DTU genes in fatty liver disease. Further research is encouraged to verify the functions and regulations of these transcript isoforms as potential diagnostic markers for fatty liver disease in humans and chickens.

The Contribution of Multiplexing Single Cell RNA Sequencing in Acute Myeloid Leukemia.

Decades ago, the treatment for acute myeloid leukemia relied on cytarabine and anthracycline. However, advancements in medical research have introduced targeted therapies, initially employing monoclonal antibodies such as ant-CD52 and anti-CD123, and subsequently utilizing specific inhibitors that target molecular mutations like anti-IDH1, IDH2, or FLT3. The challenge lies in determining the role of these therapeutic options, considering the inherent tumor heterogeneity associated with leukemia diagnosis and the clonal drift that this type of tumor can undergo. Targeted drugs necessitate an examination of various therapeutic targets at the individual cell level rather than assessing the entire population. It is crucial to differentiate between the prognostic value and therapeutic potential of a specific molecular target, depending on whether it is found in a terminally differentiated cell with limited proliferative potential or a stem cell with robust capabilities for both proliferation and self-renewal. However, this cell-by-cell analysis is accompanied by several challenges. Firstly, the scientific aspect poses difficulties in comparing different single cell analysis experiments despite efforts to standardize the results through various techniques. Secondly, there are practical obstacles as each individual cell experiment incurs significant financial costs and consumes a substantial amount of time. A viable solution lies in the ability to process multiple samples simultaneously, which is a distinctive feature of the cell hashing technique. In this study, we demonstrate the applicability of the cell hashing technique for analyzing acute myeloid leukemia cells. By comparing it to standard single cell analysis, we establish a strong correlation in various parameters such as quality control, gene expression, and the analysis of leukemic blast markers in patients. Consequently, this technique holds the potential to become an integral part of the biological assessment of acute myeloid leukemia, contributing to the personalized and optimized management of the disease, particularly in the context of employing targeted therapies.

Single-Cell Transcriptome Analysis of Acute Myeloid Leukemia Cells Using Methanol Fixation and Cryopreservation.

INTRODUCTION: The application of single-cell RNA sequencing has greatly improved our understanding of various cellular and molecular mechanisms involved in physiological and pathophysiological processes. However, obtaining living cells for this technique can be difficult under certain conditions. To solve this problem, the methanol fixation method appeared as a promising alternative for routine clinical use. MATERIALS AND METHODS: In this study, we selected two AML samples that had been fixed in methanol for 12-18 months. Once the cells were rehydrated, these samples were subjected to single-cell RNA sequencing. We then compared the results obtained from these samples with those obtained from the same samples cryopreserved in DMSO. RESULTS: We used a previously validated methanol fixation protocol to perform scRNA-seq on DMSO cryopreserved cells and cells fixed in methanol for more than one year. Preliminary results show that methanol fixation induces some genetic and transcriptional modification compared with DMSO cryopreservation but remains a valuable method for single-cell analysis of primary human leukemia cells. CONCLUSIONS: The initial findings from this study highlight certain resemblances in methanol fixation over a 12-month period and cryopreservation with DMSO, along with associated transcriptional level modifications. However, we observed genetic degradation in the fixation condition when extending beyond one year. Despite certain study limitations, it is evident that short-term methanol fixation can be effectively used for leukemia blast samples. Its ease of implementation holds the potential to simplify the integration of this technique into routine clinical practice.

Nucleoside diphosphate kinases 1 and 2 regulate a protective liver response to a high-fat diet.

The synthesis of fatty acids from acetyl-coenzyme A (AcCoA) is deregulated in diverse pathologies, including cancer. Here, we report that fatty acid accumulation is negatively regulated by nucleoside diphosphate kinases 1 and 2 (NME1/2), housekeeping enzymes involved in nucleotide homeostasis that were recently found to bind CoA. We show that NME1 additionally binds AcCoA and that ligand recognition involves a unique binding mode dependent on the CoA/AcCoA 3' phosphate. We report that Nme2 knockout mice fed a high-fat diet (HFD) exhibit excessive triglyceride synthesis and liver steatosis. In liver cells, NME2 mediates a gene transcriptional response to HFD leading to the repression of fatty acid accumulation and activation of a protective gene expression program via targeted histone acetylation. Our findings implicate NME1/2 in the epigenetic regulation of a protective liver response to HFD and suggest a potential role in controlling AcCoA usage between the competing paths of histone acetylation and fatty acid synthesis.

TranscriptomeBrowser 3.0: introducing a new compendium of molecular interactions and a new visualization tool for the study of gene regulatory networks.

BACKGROUND: Deciphering gene regulatory networks by in silico approaches is a crucial step in the study of the molecular perturbations that occur in diseases. The development of regulatory maps is a tedious process requiring the comprehensive integration of various evidences scattered over biological databases. Thus, the research community would greatly benefit from having a unified database storing known and predicted molecular interactions. Furthermore, given the intrinsic complexity of the data, the development of new tools offering integrated and meaningful visualizations of molecular interactions is necessary to help users drawing new hypotheses without being overwhelmed by the density of the subsequent graph. RESULTS: We extend the previously developed TranscriptomeBrowser database with a set of tables containing 1,594,978 human and mouse molecular interactions. The database includes: (i) predicted regulatory interactions (computed by scanning vertebrate alignments with a set of 1,213 position weight matrices), (ii) potential regulatory interactions inferred from systematic analysis of ChIP-seq experiments, (iii) regulatory interactions curated from the literature, (iv) predicted post-transcriptional regulation by micro-RNA, (v) protein kinase-substrate interactions and (vi) physical protein-protein interactions. In order to easily retrieve and efficiently analyze these interactions, we developed In-teractomeBrowser, a graph-based knowledge browser that comes as a plug-in for Transcriptome-Browser. The first objective of InteractomeBrowser is to provide a user-friendly tool to get new insight into any gene list by providing a context-specific display of putative regulatory and physical interactions. To achieve this, InteractomeBrowser relies on a "cell compartments-based layout" that makes use of a subset of the Gene Ontology to map gene products onto relevant cell compartments. This layout is particularly powerful for visual integration of heterogeneous biological information and is a productive avenue in generating new hypotheses. The second objective of InteractomeBrowser is to fill the gap between interaction databases and dynamic modeling. It is thus compatible with the network analysis software Cytoscape and with the Gene Interaction Network simulation software (GINsim). We provide examples underlying the benefits of this visualization tool for large gene set analysis related to thymocyte differentiation. CONCLUSIONS: The InteractomeBrowser plugin is a powerful tool to get quick access to a knowledge database that includes both predicted and validated molecular interactions. InteractomeBrowser is available through the TranscriptomeBrowser framework and can be found at: http://tagc.univ-mrs.fr/tbrowser/. Our database is updated on a regular basis.

A novel cross-species differential tumor classification method based on exosome-derived microRNA biomarkers established by human-dog lymphoid and mammary tumor cell lines' transcription profiles.

Background and Aim: Exosome-derived microRNA (miRNA) has been widely studied as a non-invasive candidate biomarker for tumor diagnosis in humans and dogs. Its application, however, was primarily focused on intraspecies usage for individual tumor type diagnosis. This study aimed to gain insight into its application as a cross-species differential tumor diagnostic tool; we demonstrated the process of identifying and using exosome-derived miRNA as biomarkers for the classification of lymphoid and mammary tumor cell lines in humans and dogs. Materials and Methods: Exosome-derived miRNA sequencing data from B-cell lymphoid tumor cell lines (n=13), mammary tumor cell lines (n=8), and normal mammary epithelium cultures (n=4) were pre-processed in humans and dogs. F-test and rank product (RP) analyses were used to select candidate miRNA orthologs for tumor cell line classification. The classification was carried out using an optimized support vector machine (SVM) with various kernel classifiers, including linear SVM, polynomial SVM, and radial basis function SVM. The receiver operating characteristic and precision-recall curves were used to assess the performance of all models. Results: MIR10B, MIR21, and MIR30E were chosen as the candidate orthologs from a total of 236 human-dog miRNA orthologs (p≤0.01, F-test score ≥10, and RP score ≤10). Their use of polynomial SVM provided the best performance in classifying samples from various tumor cell lines and normal epithelial culture. Conclusion: The study successfully demonstrated a method for identifying and utilizing candidate human-dog exosome-derived miRNA orthologs for differential tumor cell line classification. Such findings shed light on a novel non-invasive tumor diagnostic tool that could be used in both human and veterinary medicine in the future.

Epigenetic regulation of transcription factor binding motifs promotes Th1 response in Chagas disease cardiomyopathy.

Chagas disease, caused by the protozoan Trypanosoma cruzi, is an endemic parasitic disease of Latin America, affecting 7 million people. Although most patients are asymptomatic, 30% develop complications, including the often-fatal Chronic Chagasic Cardiomyopathy (CCC). Although previous studies have demonstrated some genetic deregulations associated with CCCs, the causes of their deregulations remain poorly described. Based on bulk RNA-seq and whole genome DNA methylation data, we investigated the genetic and epigenetic deregulations present in the moderate and severe stages of CCC. Analysis of heart tissue gene expression profile allowed us to identify 1407 differentially expressed transcripts (DEGs) specific from CCC patients. A tissue DNA methylation analysis done on the same tissue has permitted the identification of 92 regulatory Differentially Methylated Regions (DMR) localized in the promoter of DEGs. An in-depth study of the transcription factors binding sites (TFBS) in the DMRs corroborated the importance of TFBS's DNA methylation for gene expression in CCC myocardium. TBX21, RUNX3 and EBF1 are the transcription factors whose binding motif appears to be affected by DNA methylation in the largest number of genes. By combining both transcriptomic and methylomic analysis on heart tissue, and methylomic analysis on blood, 4 biological processes affected by severe CCC have been identified, including immune response, ion transport, cardiac muscle processes and nervous system. An additional study on blood methylation of moderate CCC samples put forward the importance of ion transport and nervous system in the development of the disease.

An evaluation of the role of gene expression in the prediction and diagnosis of ventilator-associated pneumonia.

BACKGROUND: The SepsiChip project explored transcriptional modulation associated with ventilator-associated pneumonia (VAP) in patients admitted to the intensive care unit for trauma. Genome-wide expression analysis may help to identify potential diagnostic markers for diseases. The current study examined the changes in blood transcriptome during VAP. METHODS: The authors prospectively included 165 trauma patients, and 41 developed VAP. Whole blood samples were collected at admission and at VAP. To predict VAP, the admission samples were compared by microarray in patients who did or did not develop VAP. To identify diagnosis markers, paired samples of 35 patients who developed VAP were analyzed. Using NanoString (Seattle, WA), the results were confirmed in the patients who developed VAP. Trauma patients who did not develop VAP served as controls to eliminate a time effect. RESULTS: The injury severity scores of the patients who did or did not develop VAP were 36 and 29, respectively. No predictive biomarker was identified. For patients who developed VAP, a transcriptional signature was identified between the two sampling times. However, this signature was a generalized pattern related to trauma, independent of the infectious process. Genes involved in the proinflammatory response were down-regulated in the patients who developed VAP, but this difference was not statistically significant. CONCLUSIONS: In contrast to clinical assessment, transcriptional analysis of whole blood samples cannot predict or diagnose VAP in trauma patients. Differentiating infection from inflammation seems challenging.

OLOGRAM-MODL: mining enriched n-wise combinations of genomic features with Monte Carlo and dictionary learning.

Most epigenetic marks, such as Transcriptional Regulators or histone marks, are biological objects known to work together in n-wise complexes. A suitable way to infer such functional associations between them is to study the overlaps of the corresponding genomic regions. However, the problem of the statistical significance of n-wise overlaps of genomic features is seldom tackled, which prevent rigorous studies of n-wise interactions. We introduce OLOGRAM-MODL, which considers overlaps between n ≥ 2 sets of genomic regions, and computes their statistical mutual enrichment by Monte Carlo fitting of a Negative Binomial distribution, resulting in more resolutive P-values. An optional machine learning method is proposed to find complexes of interest, using a new itemset mining algorithm based on dictionary learning which is resistant to noise inherent to biological assays. The overall approach is implemented through an easy-to-use CLI interface for workflow integration, and a visual tree-based representation of the results suited for explicability. The viability of the method is experimentally studied using both artificial and biological data. This approach is accessible through the command line interface of the pygtftk toolkit, available on Bioconda and from https://github.com/dputhier/pygtftk.