Testing For Differential Gene Expression Research Articles

DESpace: spatially variable gene detection via differential expression testing of spatial clusters.

Spatially resolved transcriptomics (SRT) enables scientists to investigate spatial context of mRNA abundance, including identifying spatially variable genes (SVGs), i.e. genes whose expression varies across the tissue. Although several methods have been proposed for this task, native SVG tools cannot jointly model biological replicates, or identify the key areas of the tissue affected by spatial variability. Here, we introduce DESpace, a framework, based on an original application of existing methods, to discover SVGs. In particular, our approach inputs all types of SRT data, summarizes spatial information via spatial clusters, and identifies spatially variable genes by performing differential gene expression testing between clusters. Furthermore, our framework can identify (and test) the main cluster of the tissue affected by spatial variability; this allows scientists to investigate spatial expression changes in specific areas of interest. Additionally, DESpace enables joint modeling of multiple samples (i.e. biological replicates); compared to inference based on individual samples, this approach increases statistical power, and targets SVGs with consistent spatial patterns across replicates. Overall, in our benchmarks, DESpace displays good true positive rates, controls for false positive and false discovery rates, and is computationally efficient. DESpace is freely distributed as a Bioconductor R package at https://bioconductor.org/packages/DESpace.

Genomic Characterization of Cervical Tumors Driven by Rare High-Risk HPV Subtypes Utilizing a Deep Learning Framework

As many as 30-50% of patients with locally advanced cervical cancer (LACC) experience recurrence after chemoradiation therapy (CRT). More sophisticated prognostic schemes and targeted therapeutic options are urgently needed. Human papillomavirus (HPV) is responsible for the majority of invasive cervical carcinomas, most commonly HPV 16 and 18 subtypes. In large part due to their rarity, little is known about the biologic behavior of LACC driven by non-16/18 HPV genotypes. We hypothesized that rare HPV subtypes drive distinct gene expression patterns. To aid in the common dilemma of limited sample procurement, we developed a tool which leverages population data and generative adversarial neural network (GAN) modelling to exploit patterns of gene expression of small phenotypic groups for downstream analysis. Here, we employed a deep learning approach to identify gene expression patterns unique to tumors caused by rare HPV subtypes. Whole transcriptome analysis of tumor biopsies from an institutional LACC tumor bank were employed including 130 HPV16/18 positive, and at least 3 each of rare HPV subtypes 31, 33, 39, 45, 52, 58, and 59. A Wasserstein GAN model (WGAN) was trained for each rare HPV subtype, and used to generate augmented rare HPV population data for downstream differential expression testing between rare HPV subtypes and HPV16/18. K-means clustering analysis was performed using R stats package. Differential gene expression testing between rare HPV subtype populations was performed using the EdgeR R-package. Pathway Enrichment analysis was conducted using EnrichR software. By minimizing intergroup clustering distance, 3 clusters of HPV driven disease were demonstrated. The first cluster demonstrated biology similar to HPV16 and HPV18. Namely, pathway enrichment analysis performed on lists of differentially expressed genes between GAN augmented rare HPV subtypes (HPV45, HPV52, and HPV58) and classical HPV subtypes 16 and 18, yielded no significant statistical differences. The second cluster demonstrated enriched gene expression in proinflammatory IL-1 and IL-6 family cytokines and focal adhesion signaling (HPV31, HPV33, HPV39) (IL-1 regulation of extracellular matrix, p = 5.0e-6; Oncostatin M signaling, p = 2.0e-7; Integrin Signaling, 2.5e-7). The last cluster demonstrated an increased Th2/myeloid signature (HPV59) (T helper cell surface molecules 1.2e-7, myeloid cell signature 8.5e-9). Utilizing a deep learning architecture to model gene expression of rare HPV subtypes we demonstrate 3 groups of HPV driven cervical cancer biology. These groups include a classical subtype, a high IL-1/IL-6 high subtype, and a myeloid associated subtype. Knowledge of such tumor heterogeneity can aid in the future development of targeted treatment plans and patient prognosis.

Genomic and Transcriptomic Characterization of Papillary Microcarcinomas With Lateral Neck Lymph Node Metastases.

Most papillary microcarcinomas (PMCs) are indolent and subclinical. However, as many as 10% can present with clinically significant nodal metastases. Characterization of the genomic and transcriptomic landscape of PMCs presenting with or without clinically important lymph node metastases. Formalin-fixed paraffin-embedded PMC samples from 40 patients with lateral neck nodal metastases (pN1b) and 71 patients with PMC with documented absence of nodal disease (pN0). To interrogate DNA alterations in 410 genes commonly mutated in cancer and test for differential gene expression using a custom NanoString panel of 248 genes selected primarily based on their association with tumor size and nodal disease in the papillary thyroid cancer TCGA project. The genomic landscapes of PMC with or without pN1b were similar. Mutations in TERT promoter (3%) and TP53 (1%) were exclusive to N1b cases. Transcriptomic analysis revealed differential expression of 43 genes in PMCs with pN1b compared with pN0. A random forest machine learning-based molecular classifier developed to predict regional lymph node metastasis demonstrated a negative predictive value of 0.98 and a positive predictive value of 0.72 at a prevalence of 10% pN1b disease. The genomic landscape of tumors with pN1b and pN0 disease was similar, whereas 43 genes selected primarily by mining the TCGA RNAseq data were differentially expressed. This bioinformatics-driven approach to the development of a custom transcriptomic assay provides a basis for a molecular classifier for pN1b risk stratification in PMC.

Detecting Multivariate Gene Interactions in RNA-Seq Data Using Optimal Bayesian Classification.

Differential gene expression testing is an analysis commonly applied to RNA-Seq data. These statistical tests identify genes that are significantly different across phenotypes. We extend this testing paradigm to multivariate gene interactions from a classification perspective with the goal to detect novel gene interactions for the phenotypes of interest. This is achieved through our novel computational framework comprised of a hierarchical statistical model of the RNA-Seq processing pipeline and the corresponding optimal Bayesian classifier. Through Markov Chain Monte Carlo sampling and Monte Carlo integration, we compute quantities where no analytical formulation exists. The performance is then illustrated on an expression dataset from a dietary intervention study where we identify gene pairs that have low classification error yet were not identified as differentially expressed. Additionally, we have released the software package to perform OBC classification on RNA-Seq data under an open source license and is available at http://bit.ly/obc_package.

Discovery of core biotic stress responsive genes in Arabidopsis by weighted gene co-expression network analysis.

Intricate signal networks and transcriptional regulators translate the recognition of pathogens into defense responses. In this study, we carried out a gene co-expression analysis of all currently publicly available microarray data, which were generated in experiments that studied the interaction of the model plant Arabidopsis thaliana with microbial pathogens. This work was conducted to identify (i) modules of functionally related co-expressed genes that are differentially expressed in response to multiple biotic stresses, and (ii) hub genes that may function as core regulators of disease responses. Using Weighted Gene Co-expression Network Analysis (WGCNA) we constructed an undirected network leveraging a rich curated expression dataset comprising 272 microarrays that involved microbial infections of Arabidopsis plants with a wide array of fungal and bacterial pathogens with biotrophic, hemibiotrophic, and necrotrophic lifestyles. WGCNA produced a network with scale-free and small-world properties composed of 205 distinct clusters of co-expressed genes. Modules of functionally related co-expressed genes that are differentially regulated in response to multiple pathogens were identified by integrating differential gene expression testing with functional enrichment analyses of gene ontology terms, known disease associated genes, transcriptional regulators, and cis-regulatory elements. The significance of functional enrichments was validated by comparisons with randomly generated networks. Network topology was then analyzed to identify intra- and inter-modular gene hubs. Based on high connectivity, and centrality in meta-modules that are clearly enriched in defense responses, we propose a list of 66 target genes for reverse genetic experiments to further dissect the Arabidopsis immune system. Our results show that statistical-based data trimming prior to network analysis allows the integration of expression datasets generated by different groups, under different experimental conditions and biological systems, into a functionally meaningful co-expression network.

Spliceosome Mutations Are Frequent In Patients With Myelodysplastic Syndromes Who Failed Hypomethylating Therapy: Possible Implications Of Spliceosome Inhibitors As Alternative Treatments

Myelodysplastic syndromes (MDS) are a heterogeneous group of blood cancers characterized by bone marrow (BM) failure, peripheral blood cytopenias, dysplasia, chromosomal abnormalities and an increased risk for transformation to acute myeloid leukemia (AML). Patients (pts) with higher risk disease are primarily treated with pharmacologic treatments like hypomethylating therapy (HMT) (5-azacytidine and decitabine). 5-azacytidine (AZA) and decitabine (DAC) can result in overall response rates of 36% with a median duration of response of 15 months and 17-21% with a median duration of response of 10 months, respectively. Pts refractory to HMT have poor outcomes with a median overall survival of ∼4 months. Spliceosome gene mutations are frequently found in certain subtypes of MDS specifically SF3B1 (∼28%), U2AF1 (6-12%) and SRSF2 (6-12%). The prognostic value of spliceosome mutations in different MDS subtypes has been largely investigated while the impact of these mutations on treatment response is still unknown. We aim to investigate the frequency of three commonly mutated spliceosome genes (SF3B1, U2AF1, and SRSF2) in pts who failed HMT in order to define mutational frequency and evaluate the feasibility of targeted therapy with next generation spliceosome inhibitors. We screened a cohort of 120 pts (MDS, 70; MDS/MPN, 33; MDS/sAML, 17; median age: 69; male/female: 85/35) that underwent HMT (AZA: 58; DAC: 21; AZA/DAC: 7; AZA/REV: 25; DAC/REV: 4; AZA/DAC/REV: 5). Forty-eight percent of pts failed HMT therapy as refractory or relapse. We performed Sanger sequencing on BM/peripheral blood DNA for known pathways involved in MDS pathogenesis including methylation (TET2, DNMT3A, IDH1/2), histone (ASXL1, UTX, EZH2), signaling (CBL, N/KRAS), transcription (RUNX1, TP53, JAK2), and RNA splicing (SF3B1, U2AF1, SRSF2). Data analysis was available for 90 pts. We detected a total of 131 mutations in different pathways. In total, spliceosome mutations were observed in 28/90 (31%) of pts. When we analyzed the presence of the mutations in relation to the rate of response, we found that pts who failed HMT have frequent spliceosome mutations: 17/58 (29%). We have reported that molecular mutations in TET2 and DNMT3A can predict response to treatment to HMT (Traina F, Blood (ASH Annual Meeting Abstracts), Nov 2011; 118: 461). Indeed, the frequency of mutations in methylation genes was lower in the group of pts who failed HMT (11/58; 18.9%) compared to pts who achieved hematological response (11/32; 34%). Spliceosome inhibitors have been proposed for targetted therapy in MDS. The presence of spliceosome mutations in pts who failed HMT can open a new era of investigation leading to the possibility of using spliceosome inhibitors in pts who fail conventional therapy. We performed RNA-sequencing analysis on BM cells of pts who failed HMT compared to pts who achieved hematological response (n=2 vs 2) in order to define any specific gene signature explaining the differences in response to HMT. We performed differential gene expression testing on 11,459 expressed genes. In total, 158 genes were differentially expressed at FDR < .2 in responders compared to not responders. We identified several interesting genes involved in tumorigenesis and epigenetic regulation such as YPEL3, and ST14, which were up-regulated in responders vs not responders (FC: 4 and 7.5; P<.00001). In sum, spliceosome mutations are frequent in pts who failed HMT and may affect downstream pathways most probably in the epigenetic network leading to differences in susceptibility to HMT and can open the possibility of treatment with spliceosome inhibitors. Disclosures:No relevant conflicts of interest to declare.

Computational analysis of bacterial RNA-Seq data

Recent advances in high-throughput RNA sequencing (RNA-seq) have enabled tremendous leaps forward in our understanding of bacterial transcriptomes. However, computational methods for analysis of bacterial transcriptome data have not kept pace with the large and growing data sets generated by RNA-seq technology. Here, we present new algorithms, specific to bacterial gene structures and transcriptomes, for analysis of RNA-seq data. The algorithms are implemented in an open source software system called Rockhopper that supports various stages of bacterial RNA-seq data analysis, including aligning sequencing reads to a genome, constructing transcriptome maps, quantifying transcript abundance, testing for differential gene expression, determining operon structures and visualizing results. We demonstrate the performance of Rockhopper using 2.1 billion sequenced reads from 75 RNA-seq experiments conducted with Escherichia coli, Neisseria gonorrhoeae, Salmonella enterica, Streptococcus pyogenes and Xenorhabdus nematophila. We find that the transcriptome maps generated by our algorithms are highly accurate when compared with focused experimental data from E. coli and N. gonorrhoeae, and we validate our system’s ability to identify novel small RNAs, operons and transcription start sites. Our results suggest that Rockhopper can be used for efficient and accurate analysis of bacterial RNA-seq data, and that it can aid with elucidation of bacterial transcriptomes.

Quantification of Phytophthora pisi DNA and RNA transcripts during in planta infection of pea

Phytophthora pisi is a novel species of Phytophthora that causes root rot in pea and is a potentially devastating pathogen for the cultivation of pea in many parts of the world. The focus of the current study was to: (a) develop an in planta infection system of pea seedling roots; (b) develop a quantitative PCR method to measure P. pisi and pea DNA; and (c) test for differential gene expression of selected putative pathogenicity factors in P. pisi during infection of pea. P. pisi zoospores were attracted to the tips of the pea seedling roots and encysted within 30 min. After 6 h the pathogen had reached about five cortical cell layers using both inter- and intracellular growth, and water-soaked lesions were observed visually after 20 h. DNA-based detection of P. pisi and pea showed a gradual accumulation of P. pisi DNA in pea root tips up to 48 h post infection. Quantitative reverse transcriptase PCR showed induction of genes encoding putative enzyme inhibitors during the early phase of infection, between 2 h and 6 h post infection. Genes encoding putative cysteine protease, glucanase, pleiotropic drug transporter and crinkler proteins were induced during the late phase of infection. The induced P. pisi cysteine protease belonged to a cathepsin-L-like group from the C1A subfamily of peptidases and displayed a unique modular structure that included a MD-2-related lipid-recognition domain.

Quantifying differential gene connectivity between disease states for objective identification of disease-relevant genes

BackgroundNetwork modeling of whole transcriptome expression data enables characterization of complex epistatic (gene-gene) interactions that underlie cellular functions. Though numerous methods have been proposed and successfully implemented to develop these networks, there are no formal methods for comparing differences in network connectivity patterns as a function of phenotypic trait.ResultsHere we describe a novel approach for quantifying the differences in gene-gene connectivity patterns across disease states based on Graphical Gaussian Models (GGMs). We compare the posterior probabilities of connectivity for each gene pair across two disease states, expressed as a posterior odds-ratio (postOR) for each pair, which can be used to identify network components most relevant to disease status. The method can also be generalized to model differential gene connectivity patterns within previously defined gene sets, gene networks and pathways. We demonstrate that the GGM method reliably detects differences in network connectivity patterns in datasets of varying sample size. Applying this method to two independent breast cancer expression data sets, we identified numerous reproducible differences in network connectivity across histological grades of breast cancer, including several published gene sets and pathways. Most notably, our model identified two gene hubs (MMP12 and CXCL13) that each exhibited differential connectivity to more than 30 transcripts in both datasets. Both genes have been previously implicated in breast cancer pathobiology, but themselves are not differentially expressed by histologic grade in either dataset, and would thus have not been identified using traditional differential gene expression testing approaches. In addition, 16 curated gene sets demonstrated significant differential connectivity in both data sets, including the matrix metalloproteinases, PPAR alpha sequence targets, and the PUFA synthesis pathway.ConclusionsOur results suggest that GGM can be used to formally evaluate differences in global interactome connectivity across disease states, and can serve as a powerful tool for exploring the molecular events that contribute to disease at a systems level.

Markov Chain Monte Carlo With Mixtures of Mutually Singular Distributions

Markov chain Monte Carlo (MCMC) methods for Bayesian computation are mostly used when the dominating measure is the Lebesgue measure, the counting measure, or a product of these. Many Bayesian problems give rise to distributions that are not dominated by the Lebesgue measure or the counting measure alone. In this article we introduce a simple framework for using MCMC algorithms in Bayesian computation with mixtures of mutually singular distributions. The idea is to find a common dominating measure that allows the use of traditional Metropolis-Hastings algorithms. In particular, using our formulation, the Gibbs sampler can be used whenever the full conditionals are available. We compare our formulation with the reversible jump approach and show that the two are closely related. We give results for three examples, involving testing a normal mean, variable selection in regression, and hypothesis testing for differential gene expression under multiple conditions. This allows us to compare the three methods considered: Metropolis-Hastings with mutually singular distributions, Gibbs sampler with mutually singular distributions, and reversible jump. In our examples, we found the Gibbs sampler to be more precise and to need considerably less computer time than the other methods. In addition, the full conditionals used in the Gibbs sampler can be used to further improve the estimates of the model posterior probabilities via Rao-Blackwellization, at no extra cost.

Combining Shapley value and statistics to the analysis of gene expression data in children exposed to air pollution

BackgroundIn gene expression analysis, statistical tests for differential gene expression provide lists of candidate genes having, individually, a sufficiently low p-value. However, the interpretation of each single p-value within complex systems involving several interacting genes is problematic. In parallel, in the last sixty years, game theory has been applied to political and social problems to assess the power of interacting agents in forcing a decision and, more recently, to represent the relevance of genes in response to certain conditions.ResultsIn this paper we introduce a Bootstrap procedure to test the null hypothesis that each gene has the same relevance between two conditions, where the relevance is represented by the Shapley value of a particular coalitional game defined on a microarray data-set. This method, which is called Comparative Analysis of Shapley value (shortly, CASh), is applied to data concerning the gene expression in children differentially exposed to air pollution. The results provided by CASh are compared with the results from a parametric statistical test for testing differential gene expression. Both lists of genes provided by CASh and t-test are informative enough to discriminate exposed subjects on the basis of their gene expression profiles. While many genes are selected in common by CASh and the parametric test, it turns out that the biological interpretation of the differences between these two selections is more interesting, suggesting a different interpretation of the main biological pathways in gene expression regulation for exposed individuals. A simulation study suggests that CASh offers more power than t-test for the detection of differential gene expression variability.ConclusionCASh is successfully applied to gene expression analysis of a data-set where the joint expression behavior of genes may be critical to characterize the expression response to air pollution. We demonstrate a synergistic effect between coalitional games and statistics that resulted in a selection of genes with a potential impact in the regulation of complex pathways.

Fisher's combined p-value for detecting differentially expressed genes using Affymetrix expression arrays

BackgroundCurrently, most tests of differential gene expression using Affymetrix expression array data are performed using expression summary values representing each probe set on a microarray. Recently testing methods have been proposed which incorporate probe level information. We propose a new approach that uses Fisher's method of combining evidence from multiple sources of information. Specifically, we combine p-values from probe level tests of significance.ResultsThe combined p method and other competing methods were compared using three spike-in datasets (where probe sets corresponding differentially spiked transcripts are known) and array data from a biological study validated with qRT-PCR. Based on power and false discovery rates computed for the spike-in datasets, we demonstrate that the combined p method compares favorably with other methods. We find that probe level testing methods select many of the same genes as differentially expressed. We illustrate the use of the combined p method for diagnostic purposes using examples.ConclusionCombined p is a promising alternative to existing methods of testing for differential gene expression. In addition, the combined p method is particularly well suited as a diagnostic tool for exploratory analysis of microarray data.

Identifying differential expression in multiple SAGE libraries: an overdispersed log-linear model approach

BackgroundIn testing for differential gene expression involving multiple serial analysis of gene expression (SAGE) libraries, it is critical to account for both between and within library variation. Several methods have been proposed, including the t test, tw test, and an overdispersed logistic regression approach. The merits of these tests, however, have not been fully evaluated. Questions still remain on whether further improvements can be made.ResultsIn this article, we introduce an overdispersed log-linear model approach to analyzing SAGE; we evaluate and compare its performance with three other tests: the two-sample t test, tw test and another based on overdispersed logistic linear regression. Analysis of simulated and real datasets show that both the log-linear and logistic overdispersion methods generally perform better than the t and tw tests; the log-linear method is further found to have better performance than the logistic method, showing equal or higher statistical power over a range of parameter values and with different data distributions.ConclusionOverdispersed log-linear models provide an attractive and reliable framework for analyzing SAGE experiments involving multiple libraries. For convenience, the implementation of this method is available through a user-friendly web-interface available at .