Epigenetic and Transcriptional Shifts in Human Neural Stem Cells after Reprogramming into Induced Pluripotent Stem Cells and Subsequent Redifferentiation.

Induced pluripotent stem cells (iPSCs) and their derivatives have been described to display epigenetic memory of their founder cells, as well as de novo reprogramming-associated alterations. In order to selectively explore changes due to the reprogramming process and not to heterologous somatic memory, we devised a circular reprogramming approach where somatic stem cells are used to generate iPSCs, which are subsequently re-differentiated into their original fate. As somatic founder cells, we employed human embryonic stem cell-derived neural stem cells (NSCs) and compared them to iPSC-derived NSCs derived thereof. Global transcription profiling of this isogenic circular system revealed remarkably similar transcriptomes of both NSC populations, with the exception of 36 transcripts. Amongst these we detected a disproportionately large fraction of X chromosomal genes, all of which were upregulated in iPSC-NSCs. Concurrently, we detected differential methylation of X chromosomal sites spatially coinciding with regions harboring differentially expressed genes. While our data point to a pronounced overall reinstallation of autosomal transcriptomic and methylation signatures when a defined somatic lineage is propagated through pluripotency, they also indicate that X chromosomal genes may partially escape this reinstallation process. Considering the broad application of iPSCs in disease modeling and regenerative approaches, such reprogramming-associated alterations in X chromosomal gene expression and DNA methylation deserve particular attention.

Application of transfer learning to predict drug-induced human in vivo gene expression changes using rat in vitro and in vivo data.

The liver is the primary site for the metabolism and detoxification of many compounds, including pharmaceuticals. Consequently, it is also the primary location for many adverse reactions. As the liver is not readily accessible for sampling in humans; rodent or cell line models are often used to evaluate potential toxic effects of a novel compound or candidate drug. However, relating the results of animal and in vitro studies to relevant clinical outcomes for the human in vivo situation still proves challenging. In this study, we incorporate principles of transfer learning within a deep artificial neural network allowing us to leverage the relative abundance of rat in vitro and in vivo exposure data from the Open TG-GATEs data set to train a model to predict the expected pattern of human in vivo gene expression following an exposure given measured human in vitro gene expression. We show that domain adaptation has been successfully achieved, with the rat and human in vitro data no longer being separable in the common latent space generated by the network. The network produces physiologically plausible predictions of human in vivo gene expression pattern following an exposure to a previously unseen compound. Moreover, we show the integration of the human in vitro data in the training of the domain adaptation network significantly improves the temporal accuracy of the predicted rat in vivo gene expression pattern following an exposure to a previously unseen compound. In this way, we demonstrate the improvements in prediction accuracy that can be achieved by combining data from distinct domains.

Unsupervised clustering of venous thromboembolism patients by clinical features at presentation identifies novel endotypes that improve prognostic stratification.

BACKGROUND: Individuals with acute venous thromboembolism (VTE) constitute a heterogeneous group of patients with diverse clinical characteristics and outcome. OBJECTIVES: To identify endotypes of individuals with acute VTE based on clinical characteristics at presentation through unsupervised cluster analysis and to evaluate their molecular proteomic profile and clinical outcome. METHODS: Data from 591 individuals from the Genotyping and Molecular phenotyping of Venous thromboembolism (GMP-VTE) project were explored. Hierarchical clustering was applied to 58 variables to define VTE endotypes. Clinical characteristics, three-year incidence of thromboembolic events or death, and acute-phase plasma proteomics were assessed. RESULTS: Four endotypes were identified, exhibiting different patterns of clinical characteristics and clinical course. Endotype 1 (n = 300), comprising older individuals with comorbidities, had the highest incidence of thromboembolic events or death (HR [95 % CI]: 3.76 [1.96-7.19]), followed by endotype 4 (n = 127) (HR [95 % CI]: 2.55 [1.26-5.16]), characterised by men with history of VTE and provoking risk factors, and endotype 3 (n = 57) (HR [95 % CI]: 1.57 [0.63-3.87]), composed of young women with provoking risk factors, vs. reference endotype 2 (n = 107). The reference endotype was constituted by individuals diagnosed with PE without comorbidities, who had the lowest incidence of the investigated endpoint. Differentially expressed proteins associated with the endotypes were related to distinct biological processes, supporting differences in molecular pathophysiology. The endotypes had superior prognostic ability compared to existing risk stratifications such as provoked vs unprovoked VTE and D-dimer levels. CONCLUSION: Four endotypes of VTE were identified by unsupervised phenotype-based clustering that diverge in clinical outcome and plasmatic protein signature. This approach might support the future development of individualized treatment in VTE.

A targeted proteomics investigation of the obesity paradox in venous thromboembolism.

The obesity paradox, the controversial finding that obesity promotes disease development but protects against sequelae in patients, has been observed in venous thromboembolism (VTE). The aim of this investigation was to identify a body mass-related proteomic signature in VTE patients and to evaluate whether this signature mediates the obesity paradox in VTE patients. Data from the Genotyping and Molecular Phenotyping in Venous ThromboEmbolism Project, a prospective cohort study of 693 VTE patients, were analyzed. A combined end point of recurrent VTE or all-cause death was used. Relative quantification of 444 proteins was performed using high-throughput targeted proteomics technology. Measurements were performed in samples collected during the acute VTE event and at 12-month follow-up. An 11-protein signature (CLEC4C, FABP4, FLT3LG, IL-17C, LEP, LYVE1, MASP1, ST2, THBS2, THBS4, TSLP) for body mass in VTE patients was identified. The signature did not significantly mediate the obesity paradox (change in hazard ratio [HR]: 0.04; likelihood ratio test of nested models = 7.7; P = .74), but its main constituent protein, leptin, was inversely associated with recurrent VTE or death (adjusted HR [95% confidence interval] per standard deviation increase: 0.66 [0.46-0.94]). This relationship was significantly (P = .007) modified by markers of leptin resistance (ie, high body mass index and high circulating matrix metalloproteinase-2 levels). Although the signature did not substantially explain the obesity paradox, leptin appears to be protective against disease recurrence and death in VTE patients. This protective effect was abrogated under conditions of leptin resistance and hence was unrelated to the obesity paradox.

Assessing the Contribution of Relative Macrophage Frequencies to Subcutaneous Adipose Tissue.

Background: Macrophages play an important role in regulating adipose tissue function, while their frequencies in adipose tissue vary between individuals. Adipose tissue infiltration by high frequencies of macrophages has been linked to changes in adipokine levels and low-grade inflammation, frequently associated with the progression of obesity. The objective of this project was to assess the contribution of relative macrophage frequencies to the overall subcutaneous adipose tissue gene expression using publicly available datasets. Methods: Seven publicly available microarray gene expression datasets from human subcutaneous adipose tissue biopsies (n = 519) were used together with TissueDecoder to determine the adipose tissue cell-type composition of each sample. We divided the subjects in four groups based on their relative macrophage frequencies. Differential gene expression analysis between the high and low relative macrophage frequencies groups was performed, adjusting for sex and study. Finally, biological processes were identified using pathway enrichment and network analysis. Results: We observed lower frequencies of adipocytes and higher frequencies of adipose stem cells in individuals characterized by high macrophage frequencies. We additionally studied whether, within subcutaneous adipose tissue, interindividual differences in the relative frequencies of macrophages were reflected in transcriptional differences in metabolic and inflammatory pathways. Adipose tissue of individuals with high macrophage frequencies had a higher expression of genes involved in complement activation, chemotaxis, focal adhesion, and oxidative stress. Similarly, we observed a lower expression of genes involved in lipid metabolism, fatty acid synthesis, and oxidation and mitochondrial respiration. Conclusion: We present an approach that combines publicly available subcutaneous adipose tissue gene expression datasets with a deconvolution algorithm to calculate subcutaneous adipose tissue cell-type composition. The results showed the expected increased inflammation gene expression profile accompanied by decreased gene expression in pathways related to lipid metabolism and mitochondrial respiration in subcutaneous adipose tissue in individuals characterized by high macrophage frequencies. This approach demonstrates the hidden strength of reusing publicly available data to gain cell-type-specific insights into adipose tissue function.

Protein expression profiling suggests relevance of noncanonical pathways in isolated pulmonary embolism.

Patients with isolated pulmonary embolism (PE) have a distinct clinical profile from those with deep vein thrombosis (DVT)-associated PE, with more pulmonary conditions and atherosclerosis. These findings suggest a distinct molecular pathophysiology and the potential involvement of alternative pathways in isolated PE. To test this hypothesis, data from 532 individuals from the Genotyping and Molecular Phenotyping of Venous ThromboEmbolism Project, a multicenter prospective cohort study with extensive biobanking, were analyzed. Targeted, high-throughput proteomics, machine learning, and bioinformatic methods were applied to contrast the acute-phase plasma proteomes of isolated PE patients (n = 96) against those of patients with DVT-associated PE (n = 276) or isolated DVT (n = 160). This resulted in the identification of shared molecular processes between PE phenotypes, as well as an isolated PE-specific protein signature. Shared processes included upregulation of inflammation, response to oxidative stress, and the loss of pulmonary surfactant. The isolated PE-specific signature consisted of 5 proteins: interferon-γ, glial cell line-derived neurotrophic growth factor, polypeptide N-acetylgalactosaminyltransferase 3, peptidyl arginine deiminase type-2, and interleukin-15 receptor subunit α. These proteins were orthogonally validated using cis protein quantitative trait loci. External replication in an independent population-based cohort (n = 5778) further validated the proteomic results and showed that they were prognostic for incident primary isolated PE in individuals without history of VTE (median time to event: 2.9 years; interquartile range: 1.6-4.2 years), supporting their possible involvement in the early pathogenesis. This study has identified molecular overlaps and differences between VTE phenotypes. In particular, the results implicate noncanonical pathways more commonly associated with respiratory and atherosclerotic disease in the acute pathophysiology of isolated PE.

Missing value imputation in proximity extension assay-based targeted proteomics data.

Targeted proteomics utilizing antibody-based proximity extension assays provides sensitive and highly specific quantifications of plasma protein levels. Multivariate analysis of this data is hampered by frequent missing values (random or left censored), calling for imputation approaches. While appropriate missing-value imputation methods exist, benchmarks of their performance in targeted proteomics data are lacking. Here, we assessed the performance of two methods for imputation of values missing completely at random, the previously top-benchmarked 'missForest' and the recently published 'GSimp' method. Evaluation was accomplished by comparing imputed with remeasured relative concentrations of 91 inflammation related circulating proteins in 86 samples from a cohort of 645 patients with venous thromboembolism. The median Pearson correlation between imputed and remeasured protein expression values was 69.0% for missForest and 71.6% for GSimp (p = 5.8e-4). Imputation with missForest resulted in stronger reduction of variance compared to GSimp (median relative variance of 25.3% vs. 68.6%, p = 2.4e-16) and undesired larger bias in downstream analyses. Irrespective of the imputation method used, the 91 imputed proteins revealed large variations in imputation accuracy, driven by differences in signal to noise ratio and information overlap between proteins. In summary, GSimp outperformed missForest, while both methods show good overall imputation accuracy with large variations between proteins.

The Impact of Amino Acids on Postprandial Glucose and Insulin Kinetics in Humans: A Quantitative Overview.

Different amino acids (AAs) may exert distinct effects on postprandial glucose and insulin concentrations. A quantitative comparison of the effects of AAs on glucose and insulin kinetics in humans is currently lacking. PubMed was queried to identify intervention studies reporting glucose and insulin concentrations after acute ingestion and/or intravenous infusion of AAs in healthy adults and those living with obesity and/or type 2 diabetes (T2DM). The systematic literature search identified 55 studies that examined the effects of l-leucine, l-isoleucine, l-alanine, l-glutamine, l-arginine, l-lysine, glycine, l-proline, l-phenylalanine, l-glutamate, branched-chain AAs (i.e., l-leucine, l-isoleucine, and l-valine), and multiple individual l-AAs on glucose and insulin concentrations. Oral ingestion of most individual AAs induced an insulin response, but did not alter glucose concentrations in healthy participants. Specific AAs (i.e., leucine and isoleucine) co-ingested with glucose exerted a synergistic effect on the postprandial insulin response and attenuated the glucose response compared to glucose intake alone in healthy participants. Oral AA ingestion as well as intravenous AA infusion was able to stimulate an insulin response and decrease glucose concentrations in T2DM and obese individuals. The extracted information is publicly available and can serve multiple purposes such as computational modeling.

Use of deep learning methods to translate drug-induced gene expression changes from rat to human primary hepatocytes.

In clinical trials, animal and cell line models are often used to evaluate the potential toxic effects of a novel compound or candidate drug before progressing to human trials. However, relating the results of animal and in vitro model exposures to relevant clinical outcomes in the human in vivo system still proves challenging, relying on often putative orthologs. In recent years, multiple studies have demonstrated that the repeated dose rodent bioassay, the current gold standard in the field, lacks sufficient sensitivity and specificity in predicting toxic effects of pharmaceuticals in humans. In this study, we evaluate the potential of deep learning techniques to translate the pattern of gene expression measured following an exposure in rodents to humans, circumventing the current reliance on orthologs, and also from in vitro to in vivo experimental designs. Of the applied deep learning architectures applied in this study the convolutional neural network (CNN) and a deep artificial neural network with bottleneck architecture significantly outperform classical machine learning techniques in predicting the time series of gene expression in primary human hepatocytes given a measured time series of gene expression from primary rat hepatocytes following exposure in vitro to a previously unseen compound across multiple toxicologically relevant gene sets. With a reduction in average mean absolute error across 76 genes that have been shown to be predictive for identifying carcinogenicity from 0.0172 for a random regression forest to 0.0166 for the CNN model (p < 0.05). These deep learning architecture also perform well when applied to predict time series of in vivo gene expression given measured time series of in vitro gene expression for rats.

Adipose tissue in health and disease through the lens of its building blocks.

Understanding adipose tissue cellular heterogeneity and homeostasis is essential to comprehend the cell type dynamics in metabolic diseases. Cellular subpopulations in the adipose tissue have been related to disease development, but efforts towards characterizing the adipose tissue cell type composition are limited. Here, we identify the cell type composition of the adipose tissue by using gene expression deconvolution of large amounts of publicly available transcriptomics level data. The proposed approach allows to present a comprehensive study of adipose tissue cell type composition, determining the relative amounts of 21 different cell types in 1282 adipose tissue samples detailing differences across four adipose tissue depots, between genders, across ranges of BMI and in different stages of type-2 diabetes. We compare our results to previous marker-based studies by conducting a literature review of adipose tissue cell type composition and propose candidate cellular markers to distinguish different cell types within the adipose tissue. This analysis reveals gender-specific differences in CD4+ and CD8+ T cell subsets; identifies adipose tissue as rich source of multipotent stem/stromal cells; and highlights a strongly increased immune cell content in epicardial and pericardial adipose tissue compared to subcutaneous and omental depots. Overall, this systematic analysis provides comprehensive insights into adipose tissue cell-type heterogeneity in health and disease.

Isolated Pulmonary Embolism Is Associated With a High Risk of Arterial Thrombotic Disease: Results From the VTEval Study.

BACKGROUND: Isolated PE is associated with a higher burden of atherosclerotic disease than other manifestations of VTE. RESEARCH QUESTION: We hypothesized that the presence of isolated PE may signal a chronically elevated risk of arterial thrombotic disease. STUDY DESIGN AND METHODS: Data from the VTEval Study, a prospective cohort study enrolling individuals with clinical suspicion and imaging-based diagnosis or exclusion of VTE, were analyzed. Patients with PE received whole-leg ultrasonography to assess presence of DVT. Regularized logistic regression identified features that discriminate between isolated PE and other VTE phenotypes at clinical presentation. Survival analyses were performed to evaluate the crude and adjusted 3-year risks of arterial thrombotic disease, recurrent VTE, and death. RESULTS: The sample comprised 510 patients. Isolated PE patients (n = 63) had a distinct clinical profile from patients with other VTE phenotypes (n = 447). COPD, peripheral artery disease, atrial fibrillation, and coronary artery disease were significantly more prevalent among patients with isolated PE. Isolated PE patients had significantly higher risk (incidence rate ratio vs DVT-associated PE, 3.7 (95% CI, 1.3-10.8, P = .009); vs isolated DVT, 4.8 (1.7-14.3, P = .001) of arterial thrombotic events (ie, myocardial infarction, stroke/transient ischemic attack). After adjustment for clinical profile and medication intake, the risk of arterial thrombotic events for patients with isolated PE remained quadruple that of other VTE phenotypes (hazard ratio [HR], 3.8 [1.3-10.9], P = .01). INTERPRETATION: Patients with isolated PE are at higher risk for arterial thrombosis and may require screening for arterial disease and development of novel therapeutic strategies. CLINICAL TRIAL REGISTRATION: NCT02156401.

EFMviz: A COBRA Toolbox extension to visualize Elementary Flux Modes in Genome-Scale Metabolic Models.

Elementary Flux Modes (EFMs) are a tool for constraint-based modeling and metabolic network analysis. However, systematic and automated visualization of EFMs, capable of integrating various data types is still a challenge. In this study, we developed an extension for the widely adopted COBRA Toolbox, EFMviz, for analysis and graphical visualization of EFMs as networks of reactions, metabolites and genes. The analysis workflow offers a platform for EFM visualization to improve EFM interpretability by connecting COBRA toolbox with the network analysis and visualization software Cytoscape. The biological applicability of EFMviz is demonstrated in two use cases on medium (Escherichia coli, iAF1260) and large (human, Recon 2.2) genome-scale metabolic models. EFMviz is open-source and integrated into COBRA Toolbox. The analysis workflows used for the two use cases are detailed in the two tutorials provided with EFMviz along with the data used in this study.

A computational model of postprandial adipose tissue lipid metabolism derived using human arteriovenous stable isotope tracer data.

Given the association of disturbances in non-esterified fatty acid (NEFA) metabolism with the development of Type 2 Diabetes and Non-Alcoholic Fatty Liver Disease, computational models of glucose-insulin dynamics have been extended to account for the interplay with NEFA. In this study, we use arteriovenous measurement across the subcutaneous adipose tissue during a mixed meal challenge test to evaluate the performance and underlying assumptions of three existing models of adipose tissue metabolism and construct a new, refined model of adipose tissue metabolism. Our model introduces new terms, explicitly accounting for the conversion of glucose to glyceraldehye-3-phosphate, the postprandial influx of glycerol into the adipose tissue, and several physiologically relevant delays in insulin signalling in order to better describe the measured adipose tissues fluxes. We then applied our refined model to human adipose tissue flux data collected before and after a diet intervention as part of the Yoyo study, to quantify the effects of caloric restriction on postprandial adipose tissue metabolism. Significant increases were observed in the model parameters describing the rate of uptake and release of both glycerol and NEFA. Additionally, decreases in the model's delay in insulin signalling parameters indicates there is an improvement in adipose tissue insulin sensitivity following caloric restriction.

Characterization of disease-specific cellular abundance profiles of chronic inflammatory skin conditions from deconvolution of biopsy samples.

BACKGROUND: Psoriasis and atopic dermatitis are two inflammatory skin diseases with a high prevalence and a significant burden on the patients. Underlying molecular mechanisms include chronic inflammation and abnormal proliferation. However, the cell types contributing to these molecular mechanisms are much less understood. Recently, deconvolution methodologies have allowed the digital quantification of cell types in bulk tissue based on mRNA expression data from biopsies. Using these methods to study the cellular composition of the skin enables the rapid enumeration of multiple cell types, providing insight into the numerical changes of cell types associated with chronic inflammatory skin conditions. Here, we use deconvolution to enumerate the cellular composition of the skin and estimate changes related to onset, progress, and treatment of these skin diseases. METHODS: A novel signature matrix, i.e. DerM22, containing expression data from 22 reference cell types, is used, in combination with the CIBERSORT algorithm, to identify and quantify the cellular subsets within whole skin biopsy samples. We apply the approach to public microarray mRNA expression data from the skin layers and 648 samples from healthy subjects and patients with psoriasis or atopic dermatitis. The methodology is validated by comparison to experimental results from flow cytometry and immunohistochemistry studies, and the deconvolution of independent data from isolated cell types. RESULTS: We derived the relative abundance of cell types from healthy, lesional, and non-lesional skin and observed a marked increase in the abundance of keratinocytes and leukocytes in the lesions of both inflammatory dermatological conditions. The relative fraction of these cells varied from healthy to diseased skin and from non-lesional to lesional skin. We show that changes in the relative abundance of skin-related cell types can be used to distinguish between mild and severe cases of psoriasis and atopic dermatitis, and trace the effect of treatment. CONCLUSIONS: Our analysis demonstrates the value of this new resource in interpreting skin-derived transcriptomics data by enabling the direct quantification of cell types in a skin sample and the characterization of pathological changes in tissue composition.

Termites manipulate moisture content of wood to maximize foraging resources.

Animals use cues to find their food, in microhabitats within their physiological tolerances. Termites build and modify their microhabitat, to transform hostile environments into benign ones, which raises questions about the relative importance of cues. Termites are desiccation intolerant and foraging termites are attracted to water, so most research has considered moisture to be a cue. However, termites can also transport water to food, and so moisture may play other roles than previously considered. To examine the role of moisture, we compared Coptotermes acinaciformis termite foraging decisions in laboratory experiments when they were offered dry and moist wood, with and without load. Without load, termites preferred moist wood and ate it without any building, whereas they moistened dry wood after wrapping it in a layer of clay. For the 'With load' units, termites substituted some of the wood for load-bearing clay walls, and kept the wood drier than on the unloaded units. As drier wood has higher compressive strength and higher rigidity, it allows more of the wood to be consumed. These results suggest that moisture plays a more important role in termite ecology than previously thought. Termites manipulate the moisture content according to the situational context and use it for multiple purposes: increased moisture levels soften the fibre, which facilitates foraging, yet keeping the wood dry provides higher structural stability against buckling which is especially important when foraging on wood under load.

Improved quantification of muscle insulin sensitivity using oral glucose tolerance test data: the MISI Calculator.

The Muscle Insulin Sensitivity Index (MISI) has been developed to estimate muscle-specific insulin sensitivity based on oral glucose tolerance test (OGTT) data. To date, the score has been implemented with considerable variation in literature and initial positive evaluations were not reproduced in subsequent studies. In this study, we investigate the computation of MISI on oral OGTT data with differing sampling schedules and aim to standardise and improve its calculation. Seven time point OGTT data for 2631 individuals from the Maastricht Study and seven time point OGTT data combined with a hyperinsulinemic-euglycaemic clamp for 71 individuals from the PRESERVE Study were used to evaluate the performance of MISI. MISI was computed on subsets of OGTT data representing four and five time point sampling schedules to determine minimal requirements for accurate computation of the score. A modified MISI computed on cubic splines of the measured data, resulting in improved identification of glucose peak and nadir, was compared with the original method yielding an increased correlation (ρ = 0.576) with the clamp measurement of peripheral insulin sensitivity as compared to the original method (ρ = 0.513). Finally, a standalone MISI calculator was developed allowing for a standardised method of calculation using both the original and improved methods.

A bioinformatics workflow to decipher transcriptomic data from vitamin D studies.

The link between the experimental laboratory studies and bioinformatic approaches aims to find procedures to connect tools from both branches producing workflows that bring together different techniques that are capable of exploiting data at many levels. Thanks to the open access sources and the numerous tools available, it is possible to create various pipelines capable of solving specific problems. Nevertheless the lack of connectivity between them that interconnect different approaches complicates the exploitation of these workflows. Here, we present a detailed description of a workflow composed of different bioinformatics tools that exploits data from large-scale gene expression experiments, contextualizing them at many biological levels. To illustrate the relevance of our workflow for the vitamin D community we applied it to data from myeloid cell models treated with the hormonally active form of vitamin. From raw files of functional genomic studies it is possible to utilize the whole information to obtain a biological insight. Different software and algorithms are included to analyse at pathway, metabolic, ontology and molecular biology level the effects on gene expression. The usage of different databases to analyse gene expression data allows to perform a complete interpretation of functional genomic studies and the implementation of analysis and visualization software tools gives a better understanding of the biological meaning of the results. This review is an example of how to select and bring together several software modules to create one pipeline that processes and analyses genomic data at many biological levels making it open, reproducible and user friendly. Finally, the application of our bioinformatic pipeline revealed that vitamin D modulates crucial metabolic pathways in different myeloid cells that may play an important role in their immune function.

Leukocyte Counts Based on DNA Methylation at Individual Cytosines.

BACKGROUND: White blood cell counts are routinely measured with automated hematology analyzers, by flow cytometry, or by manual counting. Here, we introduce an alternative approach based on DNA methylation (DNAm) at individual CG dinucleotides (CpGs). METHODS: We identified candidate CpGs that were nonmethylated in specific leukocyte subsets. DNAm levels (ranging from 0% to 100%) were analyzed by pyrosequencing and implemented into deconvolution algorithms to determine the relative composition of leukocytes. For absolute quantification of cell numbers, samples were supplemented with a nonmethylated reference DNA. RESULTS: Conventional blood counts correlated with DNAm at individual CpGs for granulocytes (r = -0.91), lymphocytes (r = -0.91), monocytes (r = -0.74), natural killer (NK) cells (r = -0.30), T cells (r = -0.73), CD4+ T cells (r = -0.41), CD8+ T cells (r = -0.88), and B cells (r = -0.66). Combination of these DNAm measurements into the "Epi-Blood-Count" provided similar precision as conventional methods in various independent validation sets. The method was also applicable to blood samples that were stored at 4 °C for 7 days or at -20 °C for 3 months. Furthermore, absolute cell numbers could be determined in frozen blood samples upon addition of a reference DNA, and the results correlated with measurements of automated analyzers in fresh aliquots (r = 0.84). CONCLUSIONS: White blood cell counts can be reliably determined by site-specific DNAm analysis. This approach is applicable to very small blood volumes and frozen samples, and it allows for more standardized and cost-effective analysis in clinical application.

Stem Cell Differentiation as a Non-Markov Stochastic Process.

Pluripotent stem cells can self-renew in culture and differentiate along all somatic lineages in vivo. While much is known about the molecular basis of pluripotency, the mechanisms of differentiation remain unclear. Here, we profile individual mouse embryonic stem cells as they progress along the neuronal lineage. We observe that cells pass from the pluripotent state to the neuronal state via an intermediate epiblast-like state. However, analysis of the rate at which cells enter and exit these observed cell states using a hidden Markov model indicates the presence of a chain of unobserved molecular states that each cell transits through stochastically in sequence. This chain of hidden states allows individual cells to record their position on the differentiation trajectory, thereby encoding a simple form of cellular memory. We suggest a statistical mechanics interpretation of these results that distinguishes between functionally distinct cellular "macrostates" and functionally similar molecular "microstates" and propose a model of stem cell differentiation as a non-Markov stochastic process.

Estimating real cell size distribution from cross-section microscopy imaging.

MOTIVATION: Microscopy imaging is an essential tool for medical diagnosis and molecular biology. It is particularly useful for extracting information about disease states, tissue heterogeneity and cell specific parameters such as cell type or cell size from biological specimens. However, the information obtained from the images is likely to be subjected to sampling and observational bias with respect to the underlying cell size/type distributions. RESULTS: We present an algorithm, Estimate Tissue Cell Size/Type Distribution (EstiTiCS), for the adjustment of the underestimation of the number of small cells and the size of measured cells while accounting for the section thickness independent of the tissue type. We introduce the sources of bias under different tissue distributions and their effect on the measured values with simulation experiments. Furthermore, we demonstrate our method on histological sections of paraffin-embedded adipose tissue sample images from 57 people from a dietary intervention study. This data consists of measured cell size and its distribution over the dietary intervention period at four time points. Adjusting for the bias with EstiTiCS results in a closer fit to the true/expected adipocyte size distribution with earlier studies. Therefore, we conclude that our method is suitable as the final step in estimating the tissue wide cell type/size distribution from microscopy imaging pipeline. AVAILABILITY AND IMPLEMENTATION: Source code and its documentation are available at https://github.com/michaelLenz/EstiTiCS The whole pipeline of our method is implemented in R and makes use of the 'nloptr' package. Adipose tissue data used for this study are available on request. CONTACT: Michael.Lenz@Maastrichtuniversity.nl, Gokhan.Ertaylan@Maastrichtuniversity.nl.