Data harmonization and federated learning for multi-cohort dementia research using the OMOP common data model: A Netherlands consortium of dementia cohorts case study.

BACKGROUND: Establishing collaborations between cohort studies has been fundamental for progress in health research. However, such collaborations are hampered by heterogeneous data representations across cohorts and legal constraints to data sharing. The first arises from a lack of consensus in standards of data collection and representation across cohort studies and is usually tackled by applying data harmonization processes. The second is increasingly important due to raised awareness for privacy protection and stricter regulations, such as the GDPR. Federated learning has emerged as a privacy-preserving alternative to transferring data between institutions through analyzing data in a decentralized manner. METHODS: In this study, we set up a federated learning infrastructure for a consortium of nine Dutch cohorts with appropriate data available to the etiology of dementia, including an extract, transform, and load (ETL) pipeline for data harmonization. Additionally, we assessed the challenges of transforming and standardizing cohort data using the Observational Medical Outcomes Partnership (OMOP) common data model (CDM) and evaluated our tool in one of the cohorts employing federated algorithms. RESULTS: We successfully applied our ETL tool and observed a complete coverage of the cohorts' data by the OMOP CDM. The OMOP CDM facilitated the data representation and standardization, but we identified limitations for cohort-specific data fields and in the scope of the vocabularies available. Specific challenges arise in a multi-cohort federated collaboration due to technical constraints in local environments, data heterogeneity, and lack of direct access to the data. CONCLUSION: In this article, we describe the solutions to these challenges and limitations encountered in our study. Our study shows the potential of federated learning as a privacy-preserving solution for multi-cohort studies that enhance reproducibility and reuse of both data and analyses.

LncRNA LITATS1 suppresses TGF-ß-induced EMT and cancer cell plasticity by potentiating TßRI degradation.

Epithelial cells acquire mesenchymal phenotypes through epithelial-mesenchymal transition (EMT) during cancer progression. However, how epithelial cells retain their epithelial traits and prevent malignant transformation is not well understood. Here, we report that the long noncoding RNA LITATS1 (LINC01137, ZC3H12A-DT) is an epithelial gatekeeper in normal epithelial cells and inhibits EMT in breast and non-small cell lung cancer cells. Transcriptome analysis identified LITATS1 as a TGF-ß target gene. LITATS1 expression is reduced in lung adenocarcinoma tissues compared with adjacent normal tissues and correlates with a favorable prognosis in breast and non-small cell lung cancer patients. LITATS1 depletion promotes TGF-ß-induced EMT, migration, and extravasation in cancer cells. Unbiased pathway analysis demonstrated that LITATS1 knockdown potently and selectively potentiates TGF-ß/SMAD signaling. Mechanistically, LITATS1 enhances the polyubiquitination and proteasomal degradation of TGF-ß type I receptor (TßRI). LITATS1 interacts with TßRI and the E3 ligase SMURF2, promoting the cytoplasmic retention of SMURF2. Our findings highlight a protective function of LITATS1 in epithelial integrity maintenance through the attenuation of TGF-ß/SMAD signaling and EMT.

A transcriptome atlas of leg muscles from healthy human volunteers reveals molecular and cellular signatures associated with muscle location.

Skeletal muscles support the stability and mobility of the skeleton but differ in biomechanical properties and physiological functions. The intrinsic factors that regulate muscle-specific characteristics are poorly understood. To study these, we constructed a large atlas of RNA-seq profiles from six leg muscles and two locations from one muscle, using biopsies from 20 healthy young males. We identified differential expression patterns and cellular composition across the seven tissues using three bioinformatics approaches confirmed by large-scale newly developed quantitative immune-histology procedures. With all three procedures, the muscle samples clustered into three groups congruent with their anatomical location. Concomitant with genes marking oxidative metabolism, genes marking fast- or slow-twitch myofibers differed between the three groups. The groups of muscles with higher expression of slow-twitch genes were enriched in endothelial cells and showed higher capillary content. In addition, expression profiles of Homeobox (HOX) transcription factors differed between the three groups and were confirmed by spatial RNA hybridization. We created an open-source graphical interface to explore and visualize the leg muscle atlas (https://tabbassidaloii.shinyapps.io/muscleAtlasShinyApp/). Our study reveals the molecular specialization of human leg muscles, and provides a novel resource to study muscle-specific molecular features, which could be linked with (patho)physiological processes.

Identification and functional characterization of imbalanced osteoarthritis-associated fibronectin splice variants.

OBJECTIVE: To identify FN1 transcripts associated with OA pathophysiology and investigate the downstream effects of modulating FN1 expression and relative transcript ratio. METHODS: FN1 transcriptomic data was obtained from our previously assessed RNA-seq dataset of lesioned and preserved OA cartilage samples from the Research osteoArthritis Articular Cartilage (RAAK) study. Differential transcript expression analysis was performed on all 27 FN1 transcripts annotated in the Ensembl database. Human primary chondrocytes were transduced with lentiviral particles containing short hairpin RNA (shRNA) targeting full-length FN1 transcripts or non-targeting shRNA. Subsequently, matrix deposition was induced in our 3D in vitro neo-cartilage model. Effects of changes in the FN1 transcript ratio on sulphated glycosaminoglycan (sGAG) deposition were investigated by Alcian blue staining and dimethylmethylene blue assay. Moreover, gene expression levels of 17 cartilage-relevant markers were determined by reverse transcription quantitative polymerase chain reaction. RESULTS: We identified 16 FN1 transcripts differentially expressed between lesioned and preserved cartilage. FN1-208, encoding migration-stimulating factor, was the most significantly differentially expressed protein coding transcript. Downregulation of full-length FN1 and a concomitant increased FN1-208 ratio resulted in decreased sGAG deposition as well as decreased ACAN and COL2A1 and increased ADAMTS-5, ITGB1 and ITGB5 gene expression levels. CONCLUSION: We show that full-length FN1 downregulation and concomitant relative FN1-208 upregulation was unbeneficial for deposition of cartilage matrix, likely due to decreased availability of the classical RGD (Arg-Gly-Asp) integrin-binding site of fibronectin.

Long non-coding RNA expression profiling of subchondral bone reveals AC005165.1 modifying FRZB expression during osteoarthritis.

OBJECTIVE: To gain insight in the expression profile of long non-coding RNAs (lncRNAs) in OA subchondral bone. METHODS: RNA sequencing data of macroscopically preserved and lesioned OA subchondral bone of patients that underwent joint replacement surgery due to OA (N = 22 pairs; 5 hips, 17 knees, Research osteoArthrits Articular Tissue (RAAK study) was run through an in-house pipeline to detect expression of lncRNAs. Differential expression analysis between preserved and lesioned bone was performed. Spearman correlations were calculated between differentially expressed lncRNAs and differentially expressed mRNAs identified previously in the same samples. Primary osteogenic cells were transfected with locked nucleic acid (LNA) GapmeRs targeting AC005165.1 lncRNA, to functionally investigate its potential mRNA targets. RESULTS: In total, 2816 lncRNAs were well-expressed in subchondral bone and we identified 233 lncRNAs exclusively expressed in knee and 307 lncRNAs exclusively in hip. Differential expression analysis, using all samples (N = 22 pairs; 5 hips, 17 knees), resulted in 21 differentially expressed lncRNAs [false discovery rate (FDR) < 0.05, fold change (FC) range 1.19-7.39], including long intergenic non-protein coding RNA (LINC) 1411 (LINC01411, FC = 7.39, FDR = 2.20 × 10-8), AC005165.1 (FC = 0.44, FDR = 2.37 × 10-6) and empty spiracles homeobox 2 opposite strand RNA (EMX2OS, FC = 0.41, FDR = 7.64 × 10-3). Among the differentially expressed lncRNAs, five were also differentially expressed in articular cartilage, including AC005165.1, showing similar direction of effect. Downregulation of AC005165.1 in primary osteogenic cells resulted in consistent downregulation of highly correlated frizzled related protein (FRZB). CONCLUSION: The current study identified a novel lncRNA, AC005165.1, being dysregulated in OA articular cartilage and subchondral bone. Downregulation of AC005165.1 caused a decreased expression of OA risk gene FRZB, an important member of the wnt pathway, suggesting that AC005165.1 could be an attractive potential therapeutic target with effects in articular cartilage and subchondral bone.

Deregulation of JAK2 signaling underlies primary cutaneous CD8+ aggressive epidermotropic cytotoxic T-cell lymphoma.

Primary cutaneous CD8+ aggressive epidermotropic cytotoxic T-cell lymphoma (pcAECyTCL) is a rare variant of cutaneous T-cell lymphoma with an aggressive clinical course and a very poor prognosis. Until now, neither a systematic characterization of genetic alterations driving pcAECyTCL has been performed, nor effective therapeutic regimes for patients have been defined. Here, we present the first highresolution genetic characterization of pcAECyTCL by using wholegenome and RNA sequencing. Our study provides a comprehensive description of genetic alterations (i.e., genomic rearrangements, copy number alterations and small-scale mutations) with pathogenic relevance in this lymphoma, including events that recurrently impact genes with important roles in the cell cycle, chromatin regulation and the JAKSTAT pathway. In particular, we show that mutually exclusive structural alterations involving JAK2 and SH2B3 predominantly underlie pcAECyTCL. In line with the genomic data, transcriptome analysis uncovered upregulation of the cell cycle, JAK2 signaling, NF-κB signaling and a high inflammatory response in this cancer. Functional studies confirmed oncogenicity of JAK2 fusions identified in pcAECyTCL and their sensitivity to JAK inhibitor treatment. Our findings strongly suggest that overactive JAK2 signaling is a central driver of pcAECyTCL, and consequently, patients with this neoplasm would likely benefit from therapy with JAK2 inhibitors such as Food and Drug Adminstration-approved ruxolitinib.

Improved Sézary cell detection and novel insights into immunophenotypic and molecular heterogeneity in Sézary syndrome.

Sézary syndrome (SS) is an aggressive leukemic form of cutaneous T-cell lymphoma with neoplastic CD4+ T cells present in skin, lymph nodes, and blood. Despite advances in therapy, prognosis remains poor, with a 5-year overall survival of 30%. The immunophenotype of Sézary cells is diverse, which hampers efficient diagnosis, sensitive disease monitoring, and accurate assessment of treatment response. Comprehensive immunophenotypic profiling of Sézary cells with an in-depth analysis of maturation and functional subsets has not been performed thus far. We immunophenotypically profiled 24 patients with SS using standardized and sensitive EuroFlow-based multiparameter flow cytometry. We accurately identified and quantified Sézary cells in blood and performed an in-depth assessment of their phenotypic characteristics in comparison with their normal counterparts in the blood CD4+ T-cell compartment. We observed inter- and intrapatient heterogeneity and phenotypic changes over time. Sézary cells exhibited phenotypes corresponding with classical and nonclassical T helper subsets with different maturation phenotypes. We combined multiparameter flow cytometry analyses with fluorescence-activated cell sorting and performed RNA sequencing studies on purified subsets of malignant Sézary cells and normal CD4+ T cells of the same patients. We confirmed pure monoclonality in Sézary subsets, compared transcriptomes of phenotypically distinct Sézary subsets, and identified novel downregulated genes, most remarkably THEMIS and LAIR1, which discriminate Sézary cells from normal residual CD4+ T cells. Together, these findings further unravel the heterogeneity of Sézary cell subpopulations within and between patients. These new data will support improved blood staging and more accurate disease monitoring.

Omixer: multivariate and reproducible sample randomization to proactively counter batch effects in omics studies.

MOTIVATION: Batch effects heavily impact results in omics studies, causing bias and false positive results, but software to control them preemptively is lacking. Sample randomization prior to measurement is vital for minimizing these effects, but current approaches are often ad hoc, poorly documented and ill-equipped to handle multiple batches and outcomes. RESULTS: We developed Omixer-a Bioconductor package implementing multivariate and reproducible sample randomization for omics studies. It proactively counters correlations between technical factors and biological variables of interest by optimizing sample distribution across batches. AVAILABILITYAND IMPLEMENTATION: Omixer is available from Bioconductor at http://bioconductor.org/packages/release/bioc/html/Omixer.html. Scripts and data used to generate figures available upon request. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

CACTUS: integrating clonal architecture with genomic clustering and transcriptome profiling of single tumor cells.

BACKGROUND: Drawing genotype-to-phenotype maps in tumors is of paramount importance for understanding tumor heterogeneity. Assignment of single cells to their tumor clones of origin can be approached by matching the genotypes of the clones to the mutations found in RNA sequencing of the cells. The confidence of the cell-to-clone mapping can be increased by accounting for additional measurements. Follicular lymphoma, a malignancy of mature B cells that continuously acquire mutations in parallel in the exome and in B cell receptor loci, presents a unique opportunity to join exome-derived mutations with B cell receptor sequences as independent sources of evidence for clonal evolution. METHODS: Here, we propose CACTUS, a probabilistic model that leverages the information from an independent genomic clustering of cells and exploits the scarce single cell RNA sequencing data to map single cells to given imperfect genotypes of tumor clones. RESULTS: We apply CACTUS to two follicular lymphoma patient samples, integrating three measurements: whole exome, single-cell RNA, and B cell receptor sequencing. CACTUS outperforms a predecessor model by confidently assigning cells and B cell receptor-based clusters to the tumor clones. CONCLUSIONS: The integration of independent measurements increases model certainty and is the key to improving model performance in the challenging task of charting the genotype-to-phenotype maps in tumors. CACTUS opens the avenue to study the functional implications of tumor heterogeneity, and origins of resistance to targeted therapies. CACTUS is written in R and source code, along with all supporting files, are available on GitHub ( https://github.com/LUMC/CACTUS ).

Gene Regulatory Network Analysis and Engineering Directs Development and Vascularization of Multilineage Human Liver Organoids.

Pluripotent stem cell (PSC)-derived organoids have emerged as novel multicellular models of human tissue development but display immature phenotypes, aberrant tissue fates, and a limited subset of cells. Here, we demonstrate that integrated analysis and engineering of gene regulatory networks (GRNs) in PSC-derived multilineage human liver organoids direct maturation and vascular morphogenesis in vitro. Overexpression of PROX1 and ATF5, combined with targeted CRISPR-based transcriptional activation of endogenous CYP3A4, reprograms tissue GRNs and improves native liver functions, such as FXR signaling, CYP3A4 enzymatic activity, and stromal cell reactivity. The engineered tissues possess superior liver identity when compared with other PSC-derived liver organoids and show the presence of hepatocyte, biliary, endothelial, and stellate-like cell populations in single-cell RNA-seq analysis. Finally, they show hepatic functions when studied in vivo. Collectively, our approach provides an experimental framework to direct organogenesis in vitro by systematically probing molecular pathways and transcriptional networks that promote tissue development.

Metabolic Age Based on the BBMRI-NL 1H-NMR Metabolomics Repository as Biomarker of Age-related Disease.

BACKGROUND: The blood metabolome incorporates cues from the environment and the host's genetic background, potentially offering a holistic view of an individual's health status. METHODS: We have compiled a vast resource of proton nuclear magnetic resonance metabolomics and phenotypic data encompassing over 25 000 samples derived from 26 community and hospital-based cohorts. RESULTS: Using this resource, we constructed a metabolomics-based age predictor (metaboAge) to calculate an individual's biological age. Exploration in independent cohorts demonstrates that being judged older by one's metabolome, as compared with one's chronological age, confers an increased risk on future cardiovascular disease, mortality, and functionality in older individuals. A web-based tool for calculating metaboAge (metaboage.researchlumc.nl) allows easy incorporation in other epidemiological studies. Access to data can be requested at bbmri.nl/samples-images-data. CONCLUSIONS: In summary, we present a vast resource of metabolomics data and illustrate its merit by constructing a metabolomics-based score for biological age that captures aspects of current and future cardiometabolic health.

Elucidating Epigenetic Regulation by Identifying Functional cis-Acting Long Noncoding RNAs and Their Targets in Osteoarthritic Articular Cartilage.

OBJECTIVE: To identify robustly differentially expressed long noncoding RNAs (lncRNAs) with osteoarthritis (OA) pathophysiology in cartilage and to explore potential target messenger RNA (mRNA) by establishing coexpression networks, followed by functional validation. METHODS: RNA sequencing was performed on macroscopically lesioned and preserved OA cartilage from patients who underwent joint replacement surgery due to OA (n = 98). Differential expression analysis was performed on lncRNAs that were annotated in GENCODE and Ensembl databases. To identify potential interactions, correlations were calculated between the identified differentially expressed lncRNAs and the previously reported differentially expressed protein-coding genes in the same samples. Modulation of chondrocyte lncRNA expression was achieved using locked nucleic acid GapmeRs. RESULTS: By applying our in-house pipeline, we identified 5,053 lncRNAs that were robustly expressed, of which 191 were significantly differentially expressed (according to false discovery rate) between lesioned and preserved OA cartilage. Upon integrating mRNA sequencing data, we showed that intergenic and antisense differentially expressed lncRNAs demonstrate high, positive correlations with their respective flanking sense genes. To functionally validate this observation, we selected P3H2-AS1, which was down-regulated in primary chondrocytes, resulting in the down-regulation of P3H2 gene expression levels. As such, we can confirm that P3H2-AS1 regulates its sense gene P3H2. CONCLUSION: By applying an improved detection strategy, robustly differentially expressed lncRNAs in OA cartilage were detected. Integration of these lncRNAs with differential mRNA expression levels in the same samples provided insight into their regulatory networks. Our data indicates that intergenic and antisense lncRNAs play an important role in regulating the pathophysiology of OA.

Whole-genome analysis uncovers recurrent IKZF1 inactivation and aberrant cell adhesion in blastic plasmacytoid dendritic cell neoplasm.

Blastic plasmacytoid dendritic cell neoplasm (BPDCN) is a rare and highly aggressive hematological malignancy with a poorly understood pathobiology and no effective therapeutic options. Despite a few recurrent genetic defects (eg, single nucleotide changes, indels, large chromosomal aberrations) have been identified in BPDCN, none are disease-specific, and more importantly, none explain its genesis or clinical behavior. In this study, we performed the first high resolution whole-genome analysis of BPDCN with a special focus on structural genomic alterations by using whole-genome sequencing and RNA sequencing. Our study, the first to characterize the landscape of genomic rearrangements and copy number alterations of BPDCN at nucleotide-level resolution, revealed that IKZF1, a gene encoding a transcription factor required for the differentiation of plasmacytoid dendritic cell precursors, is focally inactivated through recurrent structural alterations in this neoplasm. In concordance with the genomic data, transcriptome analysis revealed that conserved IKZF1 target genes display a loss-of-IKZF1 expression pattern. Furthermore, up-regulation of cellular processes responsible for cell-cell and cell-ECM interactions, which is a hallmark of IKZF1 deficiency, was prominent in BPDCN. Our findings suggest that IKZF1 inactivation plays a central role in the pathobiology of the disease, and consequently, therapeutic approaches directed at reestablishing the function of this gene might be beneficial for patients.

A comparison of automatic cell identification methods for single-cell RNA sequencing data.

BACKGROUND: Single-cell transcriptomics is rapidly advancing our understanding of the cellular composition of complex tissues and organisms. A major limitation in most analysis pipelines is the reliance on manual annotations to determine cell identities, which are time-consuming and irreproducible. The exponential growth in the number of cells and samples has prompted the adaptation and development of supervised classification methods for automatic cell identification. RESULTS: Here, we benchmarked 22 classification methods that automatically assign cell identities including single-cell-specific and general-purpose classifiers. The performance of the methods is evaluated using 27 publicly available single-cell RNA sequencing datasets of different sizes, technologies, species, and levels of complexity. We use 2 experimental setups to evaluate the performance of each method for within dataset predictions (intra-dataset) and across datasets (inter-dataset) based on accuracy, percentage of unclassified cells, and computation time. We further evaluate the methods' sensitivity to the input features, number of cells per population, and their performance across different annotation levels and datasets. We find that most classifiers perform well on a variety of datasets with decreased accuracy for complex datasets with overlapping classes or deep annotations. The general-purpose support vector machine classifier has overall the best performance across the different experiments. CONCLUSIONS: We present a comprehensive evaluation of automatic cell identification methods for single-cell RNA sequencing data. All the code used for the evaluation is available on GitHub ( https://github.com/tabdelaal/scRNAseq_Benchmark ). Additionally, we provide a Snakemake workflow to facilitate the benchmarking and to support the extension of new methods and new datasets.

Genomic analysis reveals recurrent deletion of JAK-STAT signaling inhibitors HNRNPK and SOCS1 in mycosis fungoides.

Mycosis fungoides (MF) is the most common cutaneous T-cell lymphoma (CTCL). Causative genetic alterations in MF are unknown. The low recurrence of pathogenic small-scale mutations (ie, nucleotide substitutions, indels) in the disease, calls for the study of additional aspects of MF genetics. Here, we investigated structural genomic alterations in tumor-stage MF by integrating whole-genome sequencing and RNA-sequencing. Multiple genes with roles in cell physiology (n = 113) and metabolism (n = 92) were found to be impacted by genomic rearrangements, including 47 genes currently implicated in cancer. Fusion transcripts involving genes of interest such as DOT1L, KDM6A, LIFR, TP53, and TP63 were also observed. Additionally, we identified recurrent deletions of genes involved in cell cycle control, chromatin regulation, the JAK-STAT pathway, and the PI-3-K pathway. Remarkably, many of these deletions result from genomic rearrangements. Deletion of tumor suppressors HNRNPK and SOCS1 were the most frequent genetic alterations in MF after deletion of CDKN2A. Notably, SOCS1 deletion could be detected in early-stage MF. In agreement with the observed genomic alterations, transcriptome analysis revealed up-regulation of the cell cycle, JAK-STAT, PI-3-K and developmental pathways. Our results position inactivation of HNRNPK and SOCS1 as potential driver events in MF development.

omicsPrint: detection of data linkage errors in multiple omics studies.

Summary: OmicsPrint is a versatile method for the detection of data linkage errors in multiple omics studies encompassing genetic, transcriptome and/or methylome data. OmicsPrint evaluates data linkage within and between omics data types using genotype calls from SNP arrays, DNA- or RNA-sequencing data and includes an algorithm to infer genotypes from Illumina DNA methylation array data. The method uses classification to verify assumed relationships and detect any data linkage errors, e.g. arising from sample mix-ups and mislabeling. Graphical and text output is provided to inspect and resolve putative data linkage errors. If sufficient genotype calls are available, first degree family relations also are revealed which can be used to check parent-offspring relations or zygosity in twin studies. Availability and implementation: omicsPrint is available from BioConductor; http://bioconductor.org/packages/omicsPrint. Supplementary information: Supplementary data are available at Bioinformatics online.