Comparison of Pattern Detection Methods in Microarray Time Series of the Segmentation Clock

Objective: To describe the influence of type of dietary fat on the activity of metabolic pathways, as measured by gene expression profiles, and the relation to plasma lipids. Background: Metabolic studies have demonstrated strong associations of dietary fatty acid composition with plasma lipids. Relatively little is known about the pathways and gene expression changes that mediate these relationships in the general population. Methods: We analyzed self-reported dietary intake of fatty acids, plasma lipid levels, and genome-wide gene expression data from Framingham Heart Study Offspring and Third Generation cohort participants. We excluded participants on lipid therapy. Multivariable linear regression models were conducted with plasma lipids as separate outcomes, energy-adjusted residuals of dietary fats as predictors, and adjustment for clinical and dietary covariates. Normalized gene expression from whole blood derived RNA was similarly modeled with additional adjustment for cell count and batch effects. Results: Among 3681 participants, higher polyunsaturated fatty acid (PUFA) intake is associated with lower LDL-C (estimated β [regression coefficient] = -0.5, p=0.002), higher HDL-C (β= 0.4, p<0.0001), and lower triglyceride (β= -0.009, p=0.0003) concentrations after adjustment for age, sex, carbohydrate, protein, and alcohol intake. Higher PUFA intake was associated with differential gene expression of cholesterol efflux transporters (ABCA1/ABCG1), LDL receptor degrader (IDOL), and non-lipoprotein metabolism related transcripts (FDR < 0.05). In contrast, higher saturated fat intake (SFA) showed inverse associations with ABCA1 expression levels and HDL cholesterol ( Figure 1 ). Conclusions: Higher PUFA intake is associated with a less atherogenic lipid profile and higher ABCA1 expression with inverse associations for higher SFA intake. Gene expression analysis reveals important links between dietary fat type, specific cholesterol metabolism pathways, and lipids in a community cohort.

Background: Neuroblastoma is the most common extracranial solid tumor in children, and treatment options for recurrent neuroblastoma are limited. Using molecular profiling to target the molecular vulnerabilities of neuroblastoma with existing therapeutic agents may result in a rational, data-driven approach with potential to improve clinical outcomes. Methods: The primary objective of this pilot study was to evaluate the feasibility of supporting real-time treatment decisions through predictive modeling of genome-wide mRNA gene expression data from neuroblastoma tumor biopsies. Feasibility was defined as completion of tumor biopsy, histopathological evaluation, RNA extraction and quality control, gene expression profiling within a CLIA-certified laboratory, bioinformatic analysis, generation of a drug predicttion report, molecular tumor board review yielding a formulated treatment plan, and independent medical monitor review within a 2-week period. Results: Five patients with multiply relapsed or refractory neuroblastoma were enrolled between April and June 2010. All biopsies passed histopathology and RNA quality control. Generation of gene expression data and its analysis (3-7 days), reports which linked this data into medically actionable drug candidates (1-5 days), molecular tumor board (1-3 days) and independent medical monitor review (1 day) were all completed in real-time. The average time was 10.5 days for all patients. Conclusion: This study shows that it is feasible to create therapeutic treatment plans based on genomic profiling in less than 12 days. This warrants further testing in a Phase I study to determine safety of predicted treatments and evaluate whether the information obtained in these analyses would result in patient benefit.

Background/Aims: Angiotensin II (Ang II)-mediated hypertension is a major risk factor for cardiovascular diseases. Ang II induces changes in vessel structure and function through the activation of genes related to signaling pathways. However, the changes in the gene expression profiles of blood vessels in response to Ang II remain unclear. Methods: Wild-type C57BL/6 mice were infused with Ang II (1500 ng/kg/min) using an osmotic pump for 1, 3, and 7 days. Vascular wall inflammation and remodeling were evaluated by pathological examination. Time-series microarray and quantitative PCR analyses were performed. Bioinformatics analyses were conducted to identify key genes, pathways, and biological processes. Results: After Ang II infusion, blood pressure and aortic remodeling were increased over time. Microarray analysis identified a totally of 3631 differentially expressed genes in aortas at days 1, 3, and 7 of Ang II infusion. These genes were involved in multiple biological processes, including cell adhesion, angiogenesis, cell migration, protein phosphorylation, immune system, and cell cycle, which may play important roles in regulating Ang II-induced arterial injury during hypertension. The genes were classified into 50 profiles by hierarchical cluster analysis, and finally, 14 significant profiles were identified. Among these genes, protein kinase cAMP-activated catalytic subunit alpha (Prkaca), a gene that directly regulated 137 neighboring genes, was located at the center of the gene network in Ang II-infused aortas. Further, Prkaca protein expression and cAMP level were downregulated in a time-dependent manner in Ang II-infused aortas. Conclusions: The combined use of DNA microarrays and cluster and gene network analyses identified Prkaca as a key Ang II-responsive gene that may mediate early vascular injury and hypertension.

In a mouse intercross with more than 500 animals and genome-wide gene expression data on six tissues, we identified a high proportion (18%) of sample mix-ups in the genotype data. Local expression quantitative trait loci (eQTL; genetic loci influencing gene expression) with extremely large effect were used to form a classifier to predict an individual’s eQTL genotype based on expression data alone. By considering multiple eQTL and their related transcripts, we identified numerous individuals whose predicted eQTL genotypes (based on their expression data) did not match their observed genotypes, and then went on to identify other individuals whose genotypes did match the predicted eQTL genotypes. The concordance of predictions across six tissues indicated that the problem was due to mix-ups in the genotypes (although we further identified a small number of sample mix-ups in each of the six panels of gene expression microarrays). Consideration of the plate positions of the DNA samples indicated a number of off-by-one and off-by-two errors, likely the result of pipetting errors. Such sample mix-ups can be a problem in any genetic study, but eQTL data allow us to identify, and even correct, such problems. Our methods have been implemented in an R package, R/lineup.

Cyenopyrafen is a recently developed acaricide with a new mode of action as a complex II inhibitor. However, it was recently shown that cross-resistance to cyenopyrafen can occur in resistant field strains of Tetranychus urticae, which might be linked to the previous use of classical METI acaricides. Here, we selected for cyenopyrafen resistance and studied the molecular mechanisms that underlie resistance. Selection for cyenopyrafen resistance confers cross-resistance to the complex II inhibitor cyflumetofen, but also to pyridaben, a frequently used complex I inhibitor. Cyenopyrafen resistance is highly synergised by piperonyl butoxide, and a 15-fold higher P450 activity was detected in the resistant strain. Target-site resistance was not detected. Genome-wide gene expression data, followed by a meta-analysis of previously obtained gene expression data, revealed the overexpression specifically of CYP392A11 and CYP392A12. Cyenopyrafen resistance is strongly linked to the overexpression of two P450s, which probably explains the observed cross-resistance. This information is highly valuable, as the novel complex II inhibitors cyenopyrafen and cyflumetofen are in the process of worldwide registration. The role of both CYP392A11 and CYP392A12 should be further supported by functional expression, but they are very promising candidates as molecular diagnostic markers for monitoring cyenopyrafen susceptibility in the field.

High-throughput techniques are producing large-scale high-dimensional (e.g., 4D with genes vs timepoints vs conditions vs tissues) genome-wide gene expression data. This induces increasing demands for effective methods for partitioning the data into biologically relevant groups. Current clustering and co-clustering approaches have limitations, which may be very time consuming and work for only low-dimensional expression datasets. In this work, we introduce a new notion of co-identification, which allows systematical identification of genes participating different functional groups under different conditions or different development stages. The key contribution of our work is to build a unified computational framework of co-identification that enables clustering to be high-dimensional and adaptive. Our framework is based upon a generic optimization model and a general optimization method termed Maximum Block Improvement. Testing results on yeast and Arabidopsis expression data are presented to demonstrate high efficiency of our approach and its effectiveness.

Carotid intima-media thickness (CIMT) is a surrogate indicator for atherosclerosis and has been shown to predict cardiovascular risk in multiple large studies. Identification of molecular markers for carotid atheroma plaque formation can be critical for early intervention and prevention of atherosclerosis. This study performed transcription factor (TF) network analysis of global gene expression data focusing on two TF genes, ZNF385D and HAND2, whose polymorphisms have been recently reported to show association with CIMT. Genome-wide gene expression data were measured from pieces of carotid endarterectomy collected from 34 hypertensive patients (atheroma plaque of stages IV and above according to the Stary classification) each paired with one sample of distant macroscopically intact tissue (stages I and II). Transcriptional regulation networks or the regulons were reconstructed for ZNF385D (5644 target genes) and HAND2 (781 target genes) using network inference. Their association with the progression of carotid atheroma was examined using gene-set enrichment analysis with extremely high statistical significance for regulons of both ZNF385D and HAND2 (p < 6.95 × 10−7) suggesting the involvement of expression quantitative loci (eQTL). Functional annotation of the regulon genes found heavy involvement in the immune system’s response to inflammation and infection in the development of atherosclerosis. Detailed examination of the regulation and correlation patterns suggests that activities of the two TF genes could have high clinical and interventional impacts on impairing carotid atheroma plaque formation and preventing carotid atherosclerosis.

Organization that scatters over different regions that perform multi-state business transaction are always interested I identifying novel patterns of interest given augmenting nosiness volumes. On a daily basis, we encounter an unparalleled increase in data from many sources, which adds to the concept of big data in terms of its size, speed, and diversity. These datasets provide significant obstacles to analytics frameworks and processing resources, so making the total study arduous for extracting useful information promptly. Therefore, in order to overcome these types of difficulties, it is crucial to create a very effective framework for analyzing large amounts of data. Therefore, in order to tackle these difficulties by harnessing non-linear connections from extensive and complex information, analytics frameworks are using machine (ML) and (DL) methods. Apache Spark is widely recognized as the most efficient tool for processing large amounts of data. It is particularly useful for solving complex machine learning problems that need several iterations. Spark MLlib, a distributed machine learning library, is used for this purpose. When dealing with research problems in the real world, architectures like Long Short-Term Memory in deep learning are a useful method for addressing practical challenges like as decreased accuracy, long-term sequence dependence, and the problem of disappearing and expanding gradients in traditional deep architectures. This study proposes a very efficient analytics system that combines a progressive machine learning method with Spark-based linear models, Multilayer Perceptron , and Long Short-Term Memory. The accuracy of the system's predictions is improved by the use of a two-stage cascade structure. The architecture that we provide makes it possible to organize statistical analysis of large amounts of data in a way that is not only scalable but also effective. We applied the cascading structure on two different real-world datasets in order to illustrate the effectiveness of our framework. Furthermore, the results of the experiments show that our analytical framework is superior to the approaches that are considered to be state-of-the-art in terms of classification accuracy.

The synchronized oscillation of segmentation clock is required to generate a sharp somite boundary during somitogenesis. However, the molecular mechanism underlying this synchronization in the mouse embryos is not clarified yet. We used both experimental and theoretical approaches to address this key question. Here we show, using chimeric embryos composed of wild-type cells and Delta like 1 (Dll1)-null cells, that Dll1-mediated Notch signalling is responsible for the synchronization mechanism. By analysing Lunatic fringe (Lfng) chimeric embryos and Notch signal reporter assays using a co-culture system, we further find that Lfng represses Notch activity in neighbouring cells by modulating Dll1 function. Finally, numerical simulations confirm that the repressive effect of Lfng against Notch activities in neighbouring cells can sufficiently explain the synchronization in vivo. Collectively, we provide a new model in which Lfng has a crucial role in intercellular coupling of the segmentation clock through a trans-repression mechanism.

ABSTRACTA fundamental requirement for embryonic development is the coordination of signaling activities in space and time. A notable example in vertebrate embryos is found during somitogenesis, where gene expression oscillations linked to the segmentation clock are synchronized across cells in the presomitic mesoderm (PSM) and result in tissue-level wave patterns. To examine their onset during mouse embryo development, we studied the dynamics of the segmentation clock gene Lfng during gastrulation. To this end, we established an imaging setup using selective plane illumination microscopy (SPIM) that enables culture and simultaneous imaging of up to four embryos (‘SPIM- for-4’). Using SPIM-for-4, combined with genetically encoded signaling reporters, we detected the onset of Lfng oscillations within newly formed mesoderm at presomite stages. Functionally, we found that initial synchrony and the first ∼6-8 oscillation cycles occurred even when Notch signaling was impaired, revealing similarities to previous findings made in zebrafish embryos. Finally, we show that a spatial period gradient is present at the onset of oscillatory activity, providing a potential mechanism accounting for our observation that wave patterns build up gradually over the first oscillation cycles.

Studies on the impact of metabolism on cell fate decisions are seeing a renaissance. However, a key challenge remains to distinguish signaling functions of metabolism from its canonical bioenergetic and biosynthetic roles, which underlie cellular homeostasis. Here, we tackled this challenge using mouse embryonic axis segmentation as an experimental model. First, we found that energetically subminimal amounts of glucose can support ongoing segmentation clock activity, providing evidence that glycolysis exerts a signaling function. Using a dynamical systems approach based on entrainment, we identified fructose 1,6-bisphosphate (FBP) as the potential signaling metabolite. Functionally, we demonstrated that glycolytic flux/FBP control the segmentation clock period and Wnt signaling in an anticorrelated manner. Critically, we showed that the slow segmentation clock phenotype caused by elevated glycolysis is mediated by Wnt signaling rather than cellular bioenergetic and biosynthetic state. Combined, our results demonstrate a modular organization of metabolic functions, revealing a signaling module of glycolysis that can be decoupled from its canonical metabolic functions.

Molecular omics (genomics, transcriptomics, proteomics, metabolomics, etc.) analysis is now playing the significant rule for cancer prediction, where gene expression profile is one of the most popular omicss. Kernel based Support Vector Machines (SVMs) are now widely using for cancer prediction based on the profiles of gene expression. However, the kernel based SVM's performance depends on both kernel and feature selection. Here, in this paper, we would like propose a way for both important kernel and features (important genes) selection for SVM to predict the colon cancer based on gene expression profile. Features (genes) selection has been done by t-statistic. Then a comparative study of cancer prediction accuracy of the proposed kernel based SVM with some popular predictors (Naive Bayes (NB), linear discriminate analysis (LDA) and quadratic linear discriminate analysis (QDA)) had been performed using the selected top 3 genes which are ranked with the help of the t-statistic. The proposed SVM-radial basis is found to be the better colon cancer predictor based on area under the curve (AUC) of the receiver operating characteristics and total accuracy. Thus we may reach to a conclusion that SVM-radial basis and t-statistic based feature selection altogether is an effective and feasible substitute to common techniques.

Narcolepsy type 1 (NT1) is a chronic sleep disorder caused by a loss of orexin-producing cells in the brain and involves autoimmune mechanisms, including the presence of autoreactive T cells. In this study, we performed genome-wide DNA methylation analysis using both CD4+/CD8+ T cells from 42 NT1 patients and 42 controls across discovery and replication cohorts. To identify methylation changes more robustly associated with the disease, we prioritized differentially methylated regions (DMRs) over single-site differentially methylated positions (DMPs). Furthermore, to validate and interpret DMP-level associations, we integrated genome-wide genotype and gene expression data obtained from the same individuals. As a result, the DMR analysis identified 15 reproducible DMRs in CD4+ T cells and 5 in CD8+ T cells, with most DMRs shared between the two cell types. Shared DMRs included regions associated with CCL5 (p= 2.1E-02) and CCR4 (p= 8.3E-03). Integrative analysis with genotype and gene expression data also showed that the DMP related to S100A4, which promotes lymphocyte migration through CCR5 and CXCR3 receptors, was associated with the disease in CD4+ T cells. Pathway analysis of genes identified through both the DMR and integrative analyses indicated enrichment in cell chemotaxis-related pathways, suggesting that aberrant chemokine-mediated cell migration plays a central role in NT1 pathogenesis. Further, NT1-associated methylation changes were predominantly hypomethylation events, significantly enriched in non-promoter, non-CpG island regions (p= 1.74E-102). We further observed that global hypomethylation levels were correlated with hypoSC, a mitotic index estimated from methylation data, highlighting increased T cell proliferation in NT1.

BackgroundCancer cells typically exhibit large-scale aberrant methylation of gene promoters. Some of the genes with promoter methylation alterations play “driver” roles in tumorigenesis, whereas others are only “passengers”.ResultsBased on the assumption that promoter methylation alteration of a driver gene may lead to expression alternation of a set of genes associated with cancer pathways, we developed a computational framework for integrating promoter methylation and gene expression data to identify driver methylation aberrations of cancer. Applying this approach to breast cancer data, we identified many novel cancer driver genes and found that some of the identified driver genes were subtype-specific for basal-like, luminal-A and HER2+ subtypes of breast cancer.ConclusionThe proposed framework proved effective in identifying cancer driver genes from genome-wide gene methylation and expression data of cancer. These results may provide new molecular targets for potential targeted and selective epigenetic therapy.

Global methylation analysis identifies prognostically important epigenetically inactivated tumor suppressor genes in multiple myeloma