Two-way Clustering Research Articles

Prediction of Mutagenicity of Chemicals from Their Calculated Molecular Descriptors: A Case Study with Structurally Homogeneous versus Diverse Datasets.

Variation in high-dimensional data is often caused by a few latent factors, and hence dimension reduction or variable selection techniques are often useful in gathering useful information from the data. In this paper we consider two such recent methods: Interrelated two-way clustering and envelope models. We couple these methods with traditional statistical procedures like ridge regression and linear discriminant analysis, and apply them on two data sets which have more predictors than samples (i.e. n << p scenario) and several types of molecular descriptors. One of these datasets consists of a congeneric group of Amines while the other has a much diverse collection compounds. The difference of prediction results between these two datasets for both the methods supports the hypothesis that for a congeneric set of compounds, descriptors of a certain type are enough to provide good QSAR models, but as the data set grows diverse including a variety of descriptors can improve model quality considerably.

Environmental control of the dominant phytoplankton in Aras Reservoir (Iran): A multivariate approach

This study was undertaken to determine the phytoplankton structure and the environmental variables comprising the driving factors leading to cyanobacterial blooms in Aras Reservoir. Sampling was carried out seasonally at five sampling sites along the main body of the reservoir. Samples were collected for phytoplankton identification and enumeration, chemical analyses and chlorophyll-a (Chl-a) concentrations at each sampling site. Principal component analysis (PCA) and two-way unweighted pair group method with arithmetic mean (UPGMA) were performed to determine the environmental variables affecting phytoplankton community dynamics. Seventy-two species belonging to five divisions were determined during this study. Cyanobacteria contained the highest density (74%) during the study period, with Pseudanabaena limnetica being the most abundant species. The Shannon diversity index was low (0.44–1.87), indicating a high level of cyanobacteria dominance and eutrophic water status. Principal component analysis indicated that total phosphorus and temperature were significantly associated with cyanobacteria growth. Two-way clustering by UPGMA indicated a close relationship among sampling sites during the same season from the perspective of the phytoplankton density. The findings of this study suggested Aras Reservoir could be highly affected by agricultural runoff and untreated sewage loadings. Thus, the proper use of fertilizers and sewage treatment should be taken into account in considering effective conservation and management plans.

Abstract 3992: Value of miRNA profiling in urine to predict significant prostate cancer

Abstract Prostate cancer (PCa) is the most frequently diagnosed malignancy among men in Western Europe and the United States. Due to PSA-testing and early detection strategies, the proportion of men diagnosed with low risk disease according to the D´Amico criteria (PSA ≤ 10ng/ml, Gleason-Score ≤ 6, clinical stage ≤ T2a) has significantly increased from 27.5% in 1990 to 1994 up to 46.4% from 2000 to 2001. Nonetheless, even among patients with low-grade and low-volume prostate cancer, more than 90% currently receive definitive treatment, e.g. surgery or radiotherapy. This in turn has raised concerns about overtreatment that may result in a decreased quality of life, typically related to an impairment of sexual or urinary function following therapies with curative intent. With the aim of reducing the risk of overtreatment in this subgroup, the strategy of active surveillance (AS) has been proposed a reasonable initial “treatment” option, which means to manage low-risk patients expectantly with the intention to treat only if signs of tumor progression emerge. Unfortunately all current models available to predict the probability of insignificant or low-risk PCa are not infallible as a considerable number of presumed low-risk tumors are misclassified. We aimed to provide additional molecular parameters that may be included into current models that help to improve the accuracy of predicting the actual PCa risk category and thus facilitate treatment planning. In this context we are searching for miRNA profiles in urine obtained after digital rectal examination (DRE) of apparently low-risk PCa patients, that in fact suffer from intermediate or high-risk PCa and thus should not be offered an AS strategy. This is a clinically pivotal issue since up to 40% of seemingly low-risk tumors are currently misclassified by clinicopathological criteria alone. Conversely, the identification of signatures that approve low-risk tumors may aid to better counsel patients eligible for AS. DRE-urine total RNA isolates of 16 patients with low-risk (Gleason score ≤ 6) and 14 of high-risk patients (GS ≥ 8) were subjected to microRNA array analysis (Exiqon A/S, 7th Gen) with a present call number of approx. 500. Unsupervised PCA analysis and two-way hierarchical clustering indicate that the samples do not separate by Gleason score. The supervised comparison of both GS groups identified 15 significantly differentially expressed miRNAs (1 upregulated, 14 downregulated), among them miR-106b, miR-20a, miR-21, and let-7a. When we compared miRNA profiles generated from PCa tissue or from blood of prostate cancer patients, our top-50 regulated DRE-urine miRNAs display a concordance of 56% (28/50). This and our experimentally optimized miRNA isolation protocol should further encourage the use of DRE-urine as a suitable non-invasive resource for generating PCa biomarkers. Citation Format: Hans Krause, Hannes Cash, Liam Stelzer, Kurt Miller, Carsten Kempkensteffen. Value of miRNA profiling in urine to predict significant prostate cancer. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 3992. doi:10.1158/1538-7445.AM2015-3992

Abstract 3576: Rictor alterations elicit non-canonical signaling mechanisms contributing to tumorigenicity and therapeutic resistance in non-small cell lung cancer (NSCLC)

Abstract Background: Rictor (RPTOR independent companion of MTOR, complex 2) is a highly conserved protein and is a critical component for proper assembly and functionality of the mTORC2 complex. The goal of our current study is to characterize the functional consequences of genomic alterations of RICTOR in advanced refractory NSCLC. Our preliminary data suggest that Rictor alterations have the potential to not only signal canonically through AKT, but also provide cancer cells with alternate, more advantageous oncogenic signaling via non-canonical mechanisms. Methods: We correlated genomic data (DNA hybrid capture based next generation sequencing (NGS), Foundation Medicine, Inc.), gene expression profiling, and clinical outcome in the context of the ongoing BATTLE-2 clinical trial of targeted therapies in chemo-refractory NSCLC (198 cases). We further (1) surveyed early stage NSCLC (230 cases) in The Cancer Genome Atlas (TCGA) database and performed two-way hierarchical clustering comparing gene expression profiling in amplified vs. diploid cases; (2) utilized a single-nucleotide polymorphism array to select RICTOR amplified and diploid NSCLC cell lines; (3) assessed Rictor protein and RNA expression by Western blot and qRT-PCR, respectively; and (4) performed RICTOR knockdown using siRNA followed by migration, invasion, and clonogenic assays. Results: In the BATTLE-2 cases, we identified 15% of RICTOR alterations (9% amplifications, 6% mutations, mutually exclusive) preferentially associated with resistance to all therapies (AKTi+MEKi, erlotinib+AKTi, sorafenib, or erlotinib). In the TCGA we found: (1) 10% of RICTOR amplifications and 3% mutations; (2) significant correlation between amplification and elevated RICTOR gene expression; and (3) a putative functional gene expression signature associated with RICTOR amplification. In diploid cell lines we found concordance between AKT phosphorylation and activation of other downstream mTORC2 targets (i.e. SGK1 and PKCα), but in RICTOR amplified cell lines we witnessed a discordant activation of these pathways, and thus were able to define unique signaling class systems in our cell lines harboring RICTOR alterations. Furthermore, following RICTOR knockdown in our amplified cell lines, a reduction in clonogenic, migratory, and invasive capacity was seen, suggesting that RICTOR amplification may provide a survival advantage in select cancer cells by tipping the signaling balance toward a non-canonical oncogenic pathway (AKT-independent). Conclusion: Rictor alterations may define a new molecular NSCLC subtype with distinct biology that expose unique avenues for therapeutic intervention. Ongoing studies are underway to explore specific therapeutic strategies, non-canonical signaling and Rictor mutations. Supported by: NHI-NCI CA155196 &amp; 2P50CA070907-16A1 Citation Format: Dennis Ruder, Vassiliki Papadimitrakopoulou, Kazuhiko Shien, Neda Kalhor, J. Jack Lee, Waun K. Hong, Ximing Tang, Luc Girard, John D. Minna, Lixia Diao, Jing Wang, Nana E. Hanson, James Sun, Vincent Miller, Garrett Frampton, Roy S. Herbst, Ignacio I. Wistuba, Julie G. Izzo. Rictor alterations elicit non-canonical signaling mechanisms contributing to tumorigenicity and therapeutic resistance in non-small cell lung cancer (NSCLC). [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 3576. doi:10.1158/1538-7445.AM2015-3576

A two-way clustering framework to identify disparities in multimorbidity patterns of mental and physical health conditions among Australians.

Multimorbidity is present in more than one quarter of the population in Australia, and its prevalence increases with age. Greater multimorbidity burden among individuals is always associated with poor health-related outcomes, including quality of life, health service utilization and mortality, among others. It is thus significant to identify the heterogeneity in multimorbidity patterns in the community and determine the impact of multimorbidity on individual health outcomes. In this paper, I propose a two-way clustering framework to identify clusters of most significant non-random comorbid health conditions and disparities in multimorbidity patterns among individuals. This framework can establish a clustering-based approach to determine the association between multimorbidity patterns and health-related outcomes and to calculate a multimorbidity score for each individual. The proposed method is illustrated using simulated data and a national survey data set of mental health and wellbeing in Australia.

Unique honey bee (Apis mellifera) hive component-based communities as detected by a hybrid of phospholipid fatty-acid and fatty-acid methyl ester analyses.

Microbial communities (microbiomes) are associated with almost all metazoans, including the honey bee Apis mellifera. Honey bees are social insects, maintaining complex hive systems composed of a variety of integral components including bees, comb, propolis, honey, and stored pollen. Given that the different components within hives can be physically separated and are nutritionally variable, we hypothesize that unique microbial communities may occur within the different microenvironments of honey bee colonies. To explore this hypothesis and to provide further insights into the microbiome of honey bees, we use a hybrid of fatty acid methyl ester (FAME) and phospholipid-derived fatty acid (PLFA) analysis to produce broad, lipid-based microbial community profiles of stored pollen, adults, pupae, honey, empty comb, and propolis for 11 honey bee hives. Averaging component lipid profiles by hive, we show that, in decreasing order, lipid markers representing fungi, Gram-negative bacteria, and Gram-positive bacteria have the highest relative abundances within honey bee colonies. Our lipid profiles reveal the presence of viable microbial communities in each of the six hive components sampled, with overall microbial community richness varying from lowest to highest in honey, comb, pupae, pollen, adults and propolis, respectively. Finally, microbial community lipid profiles were more similar when compared by component than by hive, location, or sampling year. Specifically, we found that individual hive components typically exhibited several dominant lipids and that these dominant lipids differ between components. Principal component and two-way clustering analyses both support significant grouping of lipids by hive component. Our findings indicate that in addition to the microbial communities present in individual workers, honey bee hives have resident microbial communities associated with different colony components.

Contribution to the Macrophytic Typology of Belgian Reference Watercourses

The objective of the study is to create a typology of macrophytes for the reference watercourses of the Meuse River catchment basin in Wallonia as a step towards the implementation of the European Water Framework Directive. The 50 sites studied are the object of a physicochemical and environmental characterization followed by a floristic survey (phanerogams, mosses, liverworts, and macroalgae). Six clusters of watercourses with their characteristic species are highlighted by two-way clustering and indicator species. The abundance of phanerogams in some watercourses of the Arden region is not only linked to light intensity but also depends on the degree of slope and the nature of the geological substrate.

Multivariate Analysis of Grain Yield and Its Attributing Traits in Different Maize Hybrids Grown under Heat and Drought Stress.

This study was carried out to evaluate F1 single cross-maize hybrids in four crop growing seasons (2010–2012). Morphological traits and physiological parameters of twelve maize hybrids were evaluated (i) to construct seed yield equation and (ii) to determine grain yield attributing traits of well-performing maize genotype using a previously unexplored method of two-way hierarchical clustering. In seed yield predicting equation photosynthetic rate contributed the highest variation (46%). Principal component analysis data showed that investigated traits contributed up to 90.55% variation in dependent structure. From factor analysis, we found that factor 1 contributed 49.6% variation (P < 0.05) with primary important traits (i.e., number of leaves per plant, plant height, stem diameter, fresh leaves weight, leaf area, stomata conductance, substomata CO2 absorption rate, and photosynthetic rate). The results of two-way hierarchical clustering demonstrated that Cluster III had outperforming genotype H12 (Sultan × Soneri) along with its most closely related traits (photosynthetic rate, stomata conductance, substomata CO2 absorption rate, chlorophyll contents, leaf area, and fresh stem weight). Our data shows that H12 (Sultan × Soneri) possessed the highest grain yield per plant under environmentally stress conditions, which are most likely to exist in arid and semiarid climatic conditions, such as in Pakistan.

Trend Analysis in Ecological Status and Macrophytic Characterization of Watercourses: Case of the Semois-Chiers Basin, Belgium Wallonia

In order to analyze the distribution and evolution of the aquatic vegetation and ecological status of the Semois-Chiers basin (Semois sub-basin and Chiers sub-basin), macrophyte surveys were conducted at 48 sites in 2007 and 2013. Environmental parameters were also measured in order to characterize the waterbodies in terms of physico-chemical properties and anthropogenic pressure. The two-way clustering and indicator value (INDVAL) methods were used to assess groups of sites according to their macrophytic composition and species communities. The results showed a clear difference between streams in the Lorraine area (calcareous watercourses) and in the Ardennes (siliceous). Within each natural region, those with natural vegetation of high ecological status were separated from those dominated by resistant species. The Macrophytical Biological Index for Rivers (IBMR) was calculated for the sites visited in 2010 and 2013 and the results show a trend towards an increase of IBMR values of polluted sites. For the latter, the Wilcoxon test was performed to assess the significance of the difference in quality between 2010 and 2013. This showed a statistically significant difference (p-value = 0.035). Our results showed similarities with previous data, as well as some differences. The differences observed might indicate a gradual change in the composition of the vegetation in the study area, which was caused by changes in environmental conditions. They could also reflect a lack of information about the ecology of certain groups of plants, mainly bryophytes and macroalgae that were not considered in previous studies. Despite the measures implemented under the EU’s Water Framework Directive (WFD), the current vegetation of the Semois river differs little from that observed in 1996. The headwaters of a Semois river, described in previous studies as polytrophic and devoid of vegetation, show a slight improvement, with the appearance of macrophytic species. In some parts of the Chiers sub-basin, however, resistant species observed in 1999 persist.

Using Mathematical Method to Solve Gene Identification Research

For the identification of the gene sequences of the different types of biological key to construct the gene sequence screening model based on two -way clustering algorithm. First, the establishment of the FCM algorithm based on the primary model solution similar to clustering samples using two-way clustering algorithm optimized to filter out the key gene sequence. The problem of inaccurate forecasts for the experience of the threshold, the introduction of boots with a sampling algorithm based threshold model obtained cluster of clusters. Confidence level α = 0.05 under the highest confidence, in order to solve the species optimal threshold value selected. Checksum achieve the classification of genes coding interval 90% of the validity and accuracy of 88%, a 50% increase compared to the experience threshold algorithm. As for the random noise covering part of intron fluctuations, interfere with gene identification, the wavelet transform function is introduced into the DNA coding region prediction to filter the genes noise. Therefore, In order to solve drawbacks of coding region prediction imprecise, we establish a DNA sequence coding region prediction model based on wavelet transform. Using this model, the detection rate reached to 81%, 27% increase from the neural network method, the prediction accuracy reached to75%, 36% higher than the Fourier analysis.

Bifonazole in vitro activity and its azole-parallel resistance in clinical yeast isolates

The susceptibility/resistance profile of bifonazole (BFZ) in 170 dermatophyte strains including azole parallel-resistance in 324 clinical yeast isolates was determined, additionally with impact on patient-relevant factors. Overall susceptibility to four azoles tested in parallel was 70%, with differences to both, the azoles, and species-specific for isolates from patients with superficial or invasive/systemic infections. 86% of the C. glabrata (n=166) isolates were susceptible to bifonazole, 76% were BFZ-susceptible to fluconazole-resistant C. glabrata (n=184) isolates, whereas 45% of the bifonazole-resistant strains (n=82) were susceptible to FLC. However, compared to voriconazole most of the other non-C. albicans Candida, and non-Candida species were less susceptible (< 50%) to bifonazole. As the other azoles tested, BFZ showed bimodular MIC-distribution. Susceptibility pattern analysis (SPA) demonstrated that isolates from antifungal agent pretreated patients had zero to significant less complete susceptible isolates (SP: SSSS) compared to non-treated patients. Furthermore, SPA revealed zero to fourfold parallel-resistance, species-specifically distributed, most prominently in C. glabrata and C. parapsilosis. Evaluation of azole susceptibility and two-way hierarchical clustering revealed a high grade of diversity and heterogeneity among the clinical C. glabrata isolates. A modified MIC assessment system was introduced to achieve a more realistic, well-arranged, and therapy oriented reporting of MIC in vitro data.

Molecular and morphometric analyses reveal discrete grouping of pomegranate (Punica granatum) genebank accessions away from cultivars

Management of germplasm in the field genebanks is greatly assisted by genetic analysis. Estimation of genetic diversity and assessment of genetic relationships among 45 accessions of pomegranate (Punica granatum L.) that included cultivars and germplasm collections were carried out using nine morphometric, 241 ISSR and six SSR markers. The average genetic distance values based on ANOVA sum of squares (Ward) were 3.94 and 5.10 for morphometric and DNA markers respectively. Hierarchical clustering based on genetic distances grouped the accessions into at least three distinct clusters; the two-way clustering showing the contribution of individual markers in genetic grouping. Discrete grouping of field genebank accessions (with IC numbers) away from cultivars was evident based both on quantitative traits data as well as DNA marker data. The findings suggested the possibility of broadening genetic base of cultivated varieties by augmenting the breeding programmes in India with diverse as well as trait- specific pomegranate germplasm.

Ontogeny of hepatic energy metabolism genes in mice as revealed by RNA-sequencing.

The liver plays a central role in metabolic homeostasis by coordinating synthesis, storage, breakdown, and redistribution of nutrients. Hepatic energy metabolism is dynamically regulated throughout different life stages due to different demands for energy during growth and development. However, changes in gene expression patterns throughout ontogeny for factors important in hepatic energy metabolism are not well understood. We performed detailed transcript analysis of energy metabolism genes during various stages of liver development in mice. Livers from male C57BL/6J mice were collected at twelve ages, including perinatal and postnatal time points (n = 3/age). The mRNA was quantified by RNA-Sequencing, with transcript abundance estimated by Cufflinks. One thousand sixty energy metabolism genes were examined; 794 were above detection, of which 627 were significantly changed during at least one developmental age compared to adult liver. Two-way hierarchical clustering revealed three major clusters dependent on age: GD17.5–Day 5 (perinatal-enriched), Day 10–Day 20 (pre-weaning-enriched), and Day 25–Day 60 (adolescence/adulthood-enriched). Clustering analysis of cumulative mRNA expression values for individual pathways of energy metabolism revealed three patterns of enrichment: glycolysis, ketogenesis, and glycogenesis were all perinatally-enriched; glycogenolysis was the only pathway enriched during pre-weaning ages; whereas lipid droplet metabolism, cholesterol and bile acid metabolism, gluconeogenesis, and lipid metabolism were all enriched in adolescence/adulthood. This study reveals novel findings such as the divergent expression of the fatty acid β-oxidation enzymes Acyl-CoA oxidase 1 and Carnitine palmitoyltransferase 1a, indicating a switch from mitochondrial to peroxisomal β-oxidation after weaning; as well as the dynamic ontogeny of genes implicated in obesity such as Stearoyl-CoA desaturase 1 and Elongation of very long chain fatty acids-like 3. These data shed new light on the ontogeny of homeostatic regulation of hepatic energy metabolism, which could ultimately provide new therapeutic targets for metabolic diseases.

SKIN PROTEOMIC COMPARISON AMONG FOUR GRAPE CULTIVARS WITH DIFFERENT ANTHOCYANIN CONTENTS

Grape skin is characterized by a very intense metabolism during ripening involving many biosynthetic activities such as those relating to secondary compounds like the anthocyanins. In order to improve our knowledge about the physiology of this tissue, we performed a comparison among the proteomes of the ripe skin of ?Riesling Italico?, ?Pinot Gris?, ?Pinot Noir? and ?Croatina?, genotypes that accumulate anthocyanins to different extents. In order to examine quantitative changes in proteomic maps, data underwent a two-way ANOVA test using cultivars and years as factors. The 297 significant variations underwent two different clustering procedures. By means of the two-way hierarchical clustering, it was possible to observe that the gels of the four cultivars were arranged in an order that clearly reflected anthocyanin levels and emphasized the similarities between the two Pinots, from one side, and the divergence of the proteome of ?Croatina?, from the other. It was then possible to group the 297 spots, according to their trends, in six clusters through the k-means procedure. One of the clusters included 85 spots showing a variation profile that could be associated with anthocyanin levels. Some were identified through LC-ESI-MS/MS and their role was discussed.

Furby: fuzzy force-directed bicluster visualization.

Cluster analysis is widely used to discover patterns in multi-dimensional data. Clustered heatmaps are the standard technique for visualizing one-way and two-way clustering results. In clustered heatmaps, rows and/or columns are reordered, resulting in a representation that shows the clusters as contiguous blocks. However, for biclustering results, where clusters can overlap, it is not possible to reorder the matrix in this way without duplicating rows and/or columns. We present Furby, an interactive visualization technique for analyzing biclustering results. Our contribution is twofold. First, the technique provides an overview of a biclustering result, showing the actual data that forms the individual clusters together with the information which rows and columns they share. Second, for fuzzy clustering results, the proposed technique additionally enables analysts to interactively set the thresholds that transform the fuzzy (soft) clustering into hard clusters that can then be investigated using heatmaps or bar charts. Changes in the membership value thresholds are immediately reflected in the visualization. We demonstrate the value of Furby by loading biclustering results applied to a multi-tissue dataset into the visualization. The proposed tool allows analysts to assess the overall quality of a biclustering result. Based on this high-level overview, analysts can then interactively explore the individual biclusters in detail. This novel way of handling fuzzy clustering results also supports analysts in finding the optimal thresholds that lead to the best clusters.

Genome-Scale Profiling of DNA Methylation in Sporadic Pituitary Macroadenomas: Association with Tumor Invasion and Histopathological Subtype

Pituitary adenomas (PAs) are commonly-occurring neoplasms that cause a variety of neurological and endocrine effects. Although known causal contributors include heredity, hormonal influence and somatic mutations, the pathophysiologic mechanisms driving tumorigenesis and invasion of sporadic PAs remain unknown. We aimed to investigate whether variation in DNA methylation is associated with PA invasion and functional phenotype. Genome-wide DNA methylation profiles were assessed in 24 sporadic pituitary macroadenomas, according to histopathological subtype and Knosp invasion scores. Methylation data were analyzed using two tailed t-test and one-way ANOVA statistics. Two-way hierarchical clustering analysis was performed using Pearson's correlation and average linkage for the gene/sample tree. Although no significant differences in global methylation levels were observed between invasive and noninvasive PAs, two CpGs annotated as enhancers in invasion-associated genes (FLT1 and SLIT3) were significantly hypomethylated in invasive PAs. Compared with hormonally-active somatotroph adenomas, nonfunctional PAs displayed global hypermethylation across CpG sites (β-value 0.47 versus 0.42, p = 0.005). The most significant site of differential methylation was within the promoter region of the potassium voltage-gated channel, shaker-related subfamily, β member 2 (KCNAB2) (FDR = 5.11 × 10^-10). Pathway analysis in global promoter-associated CpGs showed that non-functional PAs are most closely associated with the ion channel activity signal pathway (FDR < 0.25). DNA methylation analysis may provide a clinically useful molecular correlate to standard PA immunostaining classification. Differential methylation of cell motility related genes (FLT1 and SLIT3) may be associated with PA invasion. Hypermethylation of KCNAB2 and downstream ion channel activity signal pathways may contribute to the endocrine-inactive status of non-functional PAs.

Module function and two-way clustering analysis of Epstein-Barr virus-related nasopharyngeal cancer

This study sought to identify and characterize the function of genes as diagnostic markers for Epstein-Barr virus (EBV)-related nasopharyngeal cancer (NPC). The gene expression profile of GSE13597 was downloaded from the Gene Expression Omnibus database, which included 28 EBV-related NPC gene expression profile data sets, 25 disease samples, and 3 control samples. Data were pre-processed, and differentially expressed genes were screened using the R language. The co-expression coefficient was calculated to construct a co-expression network using Cytoscape. ClusterONE was used to perform module analysis to find enriched gene families. The BiCAT software was used to perform a two-way clustering analysis of differentially expressed gene expression profiles based on the co-expression networks and to verify the enriched gene families, followed by functional enrichment analysis of these gene families. The MCM gene family was found to be enriched in EBV-related NPC. This gene family is essential for eukaryotic DNA replication. Functional analysis of differentially expressed genes in the co-expression network revealed that the enriched biological processes and pathways were mainly involved in the cell cycle. EBV-related NPC is likely associated mainly with the process of cell reproduction, providing a strong basis for the prevention, diagnosis, and treatment of EBV-related NPC and a direction for targeted chemotherapy.

Adapting Interrelated Two-Way Clustering Method for Quantitative Structure-Activity Relationship (QSAR) Modeling of Mutagenicity/Non- Mutagenicity of a Diverse Set of Chemicals

Interrelated Two-way Clustering (ITC) is an unsupervised clustering method developed to divide samples into two groups in gene expression data obtained through microarrays, selecting important genes simultaneously in the process. This has been found to be a better approach than conventional clustering methods like K-means or selforganizing map for the scenarios when number of samples is much smaller than number of variables (n«p). In this paper we used the ITC approach for classification of a diverse set of 508 chemicals regarding mutagenicity. A large number of topological indices (TIs), 3-dimensional, and quantum chemical descriptors, as well as atom pairs (APs) has been used as explanatory variables. In this paper, ITC has been used only for predictor selection, after which ridge regression is employed to build the final predictive model. The proper leave-one-out (LOO) method of cross-validation in this scenario is to take as holdout each of the 508 compounds before predictor thinning and compare the predicted values with the experimental data. ITC based results obtained here are comparable to those developed earlier.

Oligonucleotide Microarray Identifies Genes Differentially Expressed during Tumorigenesis of DMBA-Induced Pancreatic Cancer in Rats

The extremely dismal prognosis of pancreatic cancer (PC) is attributed, at least in part, to lack of early diagnosis. Therefore, identifying differentially expressed genes in multiple steps of tumorigenesis of PC is of great interest. In the present study, a 7,12-dimethylbenzanthraene (DMBA)-induced PC model was established in male Sprague-Dawley rats. The gene expression profile was screened using an oligonucleotide microarray, followed by real-time quantitative polymerase chain reaction (qRT-PCR) and immunohistochemical staining validation. A total of 661 differentially expressed genes were identified in stages of pancreatic carcinogenesis. According to GO classification, these genes were involved in multiple molecular pathways. Using two-way hierarchical clustering analysis, normal pancreas, acute and chronic pancreatitis, PanIN, early and advanced pancreatic cancer were completely discriminated. Furthermore, 11 upregulated and 142 downregulated genes (probes) were found by Mann-Kendall trend Monotone test, indicating homologous genes of rat and human. The qRT-PCR and immunohistochemistry analysis of CXCR7 and UBe2c, two of the identified genes, confirmed the microarray results. In human PC cell lines, knockdown of CXCR7 resulted in decreased migration and invasion. Collectively, our data identified several promising markers and therapeutic targets of PC based on a comprehensive screening and systemic validation.

High-resolution chemical dissection of a model eukaryote reveals targets, pathways and gene functions

Due to evolutionary conservation of biology, experimental knowledge captured from genetic studies in eukaryotic model organisms provides insight into human cellular pathways and ultimately physiology. Yeast chemogenomic profiling is a powerful approach for annotating cellular responses to small molecules. Using an optimized platform, we provide the relative sensitivities of the heterozygous and homozygous deletion collections for nearly 1800 biologically active compounds. The data quality enables unique insights into pathways that are sensitive and resistant to a given perturbation, as demonstrated with both known and novel compounds. We present examples of novel compounds that inhibit the therapeutically relevant fatty acid synthase and desaturase (Fas1p and Ole1p), and demonstrate how the individual profiles facilitate hypothesis-driven experiments to delineate compound mechanism of action. Importantly, the scale and diversity of tested compounds yields a dataset where the number of modulated pathways approaches saturation. This resource can be used to map novel biological connections, and also identify functions for unannotated genes. We validated hypotheses generated by global two-way hierarchical clustering of profiles for (i) novel compounds with a similar mechanism of action acting upon microtubules or vacuolar ATPases, and (ii) an un-annotated ORF, YIL060w, that plays a role in respiration in the mitochondria. Finally, we identify and characterize background mutations in the widely used yeast deletion collection which should improve the interpretation of past and future screens throughout the community. This comprehensive resource of cellular responses enables the expansion of our understanding of eukaryotic pathway biology.