A comprehensive and universal method for assessing the performance of differential gene expression analyses

Transcriptional profiling identifies caspase-1 as a T cell-intrinsic regulator of Th17 differentiation

Dendritic cells (DCs) are critical for the differentiation of pathogen-specific CD4 T cells. However, to what extent innate cues from DCs dictate transcriptional changes in T cells remains elusive. Here, we used DCs stimulated with specific pathogens to prime CD4 T cells in vitro and found that these T cells express unique transcriptional profiles dictated by the nature of the priming pathogen. More specifically, the transcriptome of in vitro C. rodentium–primed Th17 cells resembled that of Th17 cells primed following infection in vivo but was remarkably distinct from cytokine-polarized Th17 cells. We identified caspase-1 as a unique gene up-regulated only in pathogen-primed Th17 cells and discovered a critical role for T cell–intrinsic caspase-1, independent of inflammasome, in optimal priming of Th17 responses. T cells lacking caspase-1 failed to induce colitis or confer protection against C. rodentium infection due to suboptimal Th17 cell differentiation in vivo. This study underlines the importance of DC-mediated priming in identifying novel regulators of T cell differentiation.

multiHiCcompare: joint normalization and comparative analysis of complex Hi-C experiments

MOTIVATION

With the development of chromatin conformation capture technology and its high-throughput derivative Hi-C sequencing, studies of the three-dimensional interactome of the genome that involve multiple Hi-C datasets are becoming available. To account for the technology-driven biases unique to each dataset, there is a distinct need for methods to jointly normalize multiple Hi-C datasets. Previous attempts at removing biases from Hi-C data have made use of techniques which normalize individual Hi-C datasets, or, at best, jointly normalize two datasets.

RESULTS

Here, we present multiHiCcompare, a cyclic loess regression-based joint normalization technique for removing biases across multiple Hi-C datasets. In contrast to other normalization techniques, it properly handles the Hi-C-specific decay of chromatin interaction frequencies with the increasing distance between interacting regions. multiHiCcompare uses the general linear model framework for comparative analysis of multiple Hi-C datasets, adapted for the Hi-C-specific decay of chromatin interaction frequencies. multiHiCcompare outperforms other methods when detecting a priori known chromatin interaction differences from jointly normalized datasets. Applied to the analysis of auxin-treated versus untreated experiments, and CTCF depletion experiments, multiHiCcompare was able to recover the expected epigenetic and gene expression signatures of loss of chromatin interactions and reveal novel insights.

AVAILABILITY AND IMPLEMENTATION

multiHiCcompare is freely available on GitHub and as a Bioconductor R package https://bioconductor.org/packages/multiHiCcompare.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

HiCcompare: an R-package for joint normalization and comparison of HI-C datasets

Background

Changes in spatial chromatin interactions are now emerging as a unifying mechanism orchestrating the regulation of gene expression. Hi-C sequencing technology allows insight into chromatin interactions on a genome-wide scale. However, Hi-C data contains many DNA sequence- and technology-driven biases. These biases prevent effective comparison of chromatin interactions aimed at identifying genomic regions differentially interacting between, e.g., disease-normal states or different cell types. Several methods have been developed for normalizing individual Hi-C datasets. However, they fail to account for biases between two or more Hi-C datasets, hindering comparative analysis of chromatin interactions.

Results

We developed a simple and effective method, HiCcompare, for the joint normalization and differential analysis of multiple Hi-C datasets. The method introduces a distance-centric analysis and visualization of the differences between two Hi-C datasets on a single plot that allows for a data-driven normalization of biases using locally weighted linear regression (loess). HiCcompare outperforms methods for normalizing individual Hi-C datasets and methods for differential analysis (diffHiC, FIND) in detecting a priori known chromatin interaction differences while preserving the detection of genomic structures, such as A/B compartments.

Conclusions

HiCcompare is able to remove between-dataset bias present in Hi-C matrices. It also provides a user-friendly tool to allow the scientific community to perform direct comparisons between the growing number of pre-processed Hi-C datasets available at online repositories. HiCcompare is freely available as a Bioconductor R package https://bioconductor.org/packages/HiCcompare/.

Electronic supplementary material

The online version of this article (10.1186/s12859-018-2288-x) contains supplementary material, which is available to authorized users.

Distinct patterns of innate immune activation by clinical isolates of respiratory syncytial virus

Respiratory syncytial virus (RSV) is a major respiratory pathogen of infants and young children. Multiple strains of both subgroup A and B viruses circulate during each seasonal epidemic. Genetic heterogeneity among RSV genomes, in large part due to the error prone RNA-dependent, RNA polymerase, could mediate variations in pathogenicity. We evaluated clinical strains of RSV for their ability to induce the innate immune response. Subgroup B viruses were used to infect human pulmonary epithelial cells (A549) and primary monocyte-derived human macrophages (MDM) from a variety of donors. Secretions of IL-6 and CCL5 (RANTES) from infected cells were measured following infection. Host and viral transcriptome expression were assessed using RNA-SEQ technology and the genomic sequences of several clinical isolates were determined. There were dramatic differences in the induction of IL-6 and CCL5 in both A549 cells and MDM infected with a variety of clinical isolates of RSV. Transcriptome analyses revealed that the pattern of innate immune activation in MDM was virus-specific and host-specific. Specifically, viruses that induced high levels of secreted IL-6 and CCL5 tended to induce cellular innate immune pathways whereas viruses that induced relatively low level of IL-6 or CCL5 did not induce or suppressed innate immune gene expression. Activation of the host innate immune response mapped to variations in the RSV G gene and the M2-1 gene. Viral transcriptome data indicated that there was a gradient of transcription across the RSV genome though in some strains, RSV G was the expressed in the highest amounts at late times post-infection. Clinical strains of RSV differ in cytokine/chemokine induction and in induction and suppression of host genes expression suggesting that these viruses may have inherent differences in virulence potential. Identification of the genetic elements responsible for these differences may lead to novel approaches to antiviral agents and vaccines.

RAD51 interconnects between DNA replication, DNA repair and immunity

RAD51, a multifunctional protein, plays a central role in DNA replication and homologous recombination repair, and is known to be involved in cancer development. We identified a novel role for RAD51 in innate immune response signaling. Defects in RAD51 lead to the accumulation of self-DNA in the cytoplasm, triggering a STING-mediated innate immune response after replication stress and DNA damage. In the absence of RAD51, the unprotected newly replicated genome is degraded by the exonuclease activity of MRE11, and the fragmented nascent DNA accumulates in the cytosol, initiating an innate immune response. Our data suggest that in addition to playing roles in homologous recombination-mediated DNA double-strand break repair and replication fork processing, RAD51 is also implicated in the suppression of innate immunity. Thus, our study reveals a previously uncharacterized role of RAD51 in initiating immune signaling, placing it at the hub of new interconnections between DNA replication, DNA repair, and immunity.

Improving sensitivity of linear regression-based cell type-specific differential expression deconvolution with per-gene vs. global significance threshold

Background

The goal of many human disease-oriented studies is to detect molecular mechanisms different between healthy controls and patients. Yet, commonly used gene expression measurements from blood samples suffer from variability of cell composition. This variability hinders the detection of differentially expressed genes and is often ignored. Combined with cell counts, heterogeneous gene expression may provide deeper insights into the gene expression differences on the cell type-specific level.

Published computational methods use linear regression to estimate cell type-specific differential expression, and a global cutoff to judge significance, such as False Discovery Rate (FDR). Yet, they do not consider many artifacts hidden in high-dimensional gene expression data that may negatively affect linear regression. In this paper we quantify the parameter space affecting the performance of linear regression (sensitivity of cell type-specific differential expression detection) on a per-gene basis.

Results

We evaluated the effect of sample sizes, cell type-specific proportion variability, and mean squared error on sensitivity of cell type-specific differential expression detection using linear regression. Each parameter affected variability of cell type-specific expression estimates and, subsequently, the sensitivity of differential expression detection. We provide the R package, LRCDE, which performs linear regression-based cell type-specific differential expression (deconvolution) detection on a gene-by-gene basis. Accounting for variability around cell type-specific gene expression estimates, it computes per-gene t-statistics of differential detection, p-values, t-statistic-based sensitivity, group-specific mean squared error, and several gene-specific diagnostic metrics.

Conclusions

The sensitivity of linear regression-based cell type-specific differential expression detection differed for each gene as a function of mean squared error, per group sample sizes, and variability of the proportions of target cell (cell type being analyzed). We demonstrate that LRCDE, which uses Welch’s t-test to compare per-gene cell type-specific gene expression estimates, is more sensitive in detecting cell type-specific differential expression at α < 0.05 missed by the global false discovery rate threshold FDR < 0.3.

Electronic supplementary material

The online version of this article (doi:10.1186/s12859-016-1226-z) contains supplementary material, which is available to authorized users.

DNA polymerase-α regulates the activation of type I interferons through cytosolic RNA:DNA synthesis

Aberrant nucleic acids generated during viral replication are the main trigger for antiviral immunity, and mutations that disrupt nucleic acid metabolism can lead to autoinflammatory disorders. Here we investigated the etiology of X-linked reticulate pigmentary disorder (XLPDR), a primary immunodeficiency with autoinflammatory features. We discovered that XLPDR is caused by an intronic mutation that disrupts the expression of POLA1, which encodes the catalytic subunit of DNA polymerase-α. Unexpectedly, POLA1 deficiency resulted in increased production of type I interferons. This enzyme is necessary for the synthesis of RNA:DNA primers during DNA replication and, strikingly, we found that POLA1 is also required for the synthesis of cytosolic RNA:DNA, which directly modulates interferon activation. Together this work identifies POLA1 as a critical regulator of the type I interferon response.

Differential outcome of TRIF-mediated signaling in TLR4 and TLR3 induced DC maturation

Recognition of pathogen-associated molecular patterns by Toll-like receptors (TLRs) on dendritic cells (DCs) leads to DC maturation, a process involving up-regulation of MHC and costimulatory molecules and secretion of proinflammatory cytokines. All TLRs except TLR3 achieve these outcomes by using the signaling adaptor myeloid differentiation factor 88. TLR4 and TLR3 can both use the Toll-IL-1 receptor domain-containing adaptor inducing IFN-β (TRIF)-dependent signaling pathway leading to IFN regulatory factor 3 (IRF3) activation and induction of IFN-β and -α4. The TRIF signaling pathway, downstream of both of these TLRs, also leads to DC maturation, and it has been proposed that the type I IFNs act in cis to induce DC maturation and subsequent effects on adaptive immunity. The present study was designed to understand the molecular mechanisms of TRIF-mediated DC maturation. We have discovered that TLR4-TRIF-induced DC maturation was independent of both IRF3 and type I IFNs. In contrast, TLR3-mediated DC maturation was completely dependent on type I IFN feedback. We found that differential activation of mitogen-activated protein kinases by the TLR4- and TLR3-TRIF axes determined the type I IFN dependency for DC maturation. In addition, we found that the adjuvanticity of LPS to induce T-cell activation is completely independent of type I IFNs. The important distinction between the TRIF-mediated signaling pathways of TLR4 and TLR3 discovered here could have a major impact in the design of future adjuvants that target this pathway.

B-Cell and Monocyte Contribution to Systemic Lupus Erythematosus Identified by Cell-Type-Specific Differential Expression Analysis in RNA-Seq Data

Systemic lupus erythematosus (SLE) is an autoimmune disease characterized by complex interplay among immune cell types. SLE activity is experimentally assessed by several blood tests, including gene expression profiling of heterogeneous populations of cells in peripheral blood. To better understand the contribution of different cell types in SLE pathogenesis, we applied the two methods in cell-type-specific differential expression analysis, csSAM and DSection, to identify cell-type-specific gene expression differences in heterogeneous gene expression measures obtained using RNA-seq technology. We identified B-cell-, monocyte-, and neutrophil-specific gene expression differences. Immunoglobulin-coding gene expression was altered in B-cells, while a ribosomal signature was prominent in monocytes. On the contrary, genes differentially expressed in the heterogeneous mixture of cells did not show any functional enrichment. Our results identify antigen binding and structural constituents of ribosomes as functions altered by B-cell- and monocyte-specific gene expression differences, respectively. Finally, these results position both csSAM and DSection methods as viable techniques for cell-type-specific differential expression analysis, which may help uncover pathogenic, cell-type-specific processes in SLE.

Transgenic expression of microRNA-181d augments the stress-sensitivity of CD4(+)CD8(+) thymocytes

Physiological stress resulting from infections, trauma, surgery, alcoholism, malnutrition, and/or pregnancy results in a substantial depletion of immature CD4⁺CD8⁺ thymocytes. We previously identified 18 distinct stress-responsive microRNAs (miRs) in the thymus upon systemic stress induced by lipopolysaccharide (LPS) or the synthetic glucocorticoid, dexamethasone (Dex). MiRs are short, non-coding RNAs that play critical roles in the immune system by targeting diverse mRNAs, suggesting that their modulation in the thymus in response to stress could impact thymopoiesis. MiR-181d is one such stress-responsive miR, exhibiting a 15-fold down-regulation in expression. We utilized both transgenic and gene-targeting approaches to study the impact of miR-181d on thymopoiesis under normal and stress conditions. The over-expression of miR-181d in developing thymocytes reduced the total number of immature CD4⁺CD8⁺ thymocytes. LPS or Dex injections caused a 4-fold greater loss of these cells when compared with the wild type controls. A knockout mouse was developed to selectively eliminate miR-181d, leaving the closely spaced and contiguous family member miR-181c intact. The targeted elimination of just miR-181d resulted in a thymus stress-responsiveness similar to wild-type mice. These experiments suggest that one or more of three other miR-181 family members have overlapping or compensatory functions. Gene expression comparisons of thymocytes from the wild type versus transgenic mice indicated that miR-181d targets a number of stress, metabolic, and signaling pathways. These findings demonstrate that selected miRs enhance stress-mediated thymic involution in vivo.

Signature MicroRNA expression patterns identified in humans with 22q11.2 deletion/DiGeorge syndrome

Patients with 22q11.2 deletion syndrome have heterogeneous clinical presentations including immunodeficiency, cardiac anomalies, and hypocalcemia. The syndrome arises from hemizygous deletions of up to 3Mb on chromosome 22q11.2, a region that contains 60 genes and 4 microRNAs. MicroRNAs are important post-transcriptional regulators of gene expression, with mutations in several microRNAs causal to specific human diseases. We characterized the microRNA expression patterns in the peripheral blood of patients with 22q11.2 deletion syndrome (n=31) compared to normal controls (n=22). Eighteen microRNAs had a statistically significant differential expression (p<0.05), with miR-185 expressed at 0.4× normal levels. The 22q11.2 deletion syndrome cohort exhibited microRNA expression hyper-variability and group dysregulation. Selected microRNAs distinguished patients with cardiac anomalies, hypocalcemia, and/or low circulating T cell counts. In summary, microRNA profiling of chromosome 22q11.2 deletion syndrome/DiGeorge patients revealed a signature microRNA expression pattern distinct from normal controls with clinical relevance.

Literature aided determination of data quality and statistical significance threshold for gene expression studies

Background

Gene expression data are noisy due to technical and biological variability. Consequently, analysis of gene expression data is complex. Different statistical methods produce distinct sets of genes. In addition, selection of expression p-value (EPv) threshold is somewhat arbitrary. In this study, we aimed to develop novel literature based approaches to integrate functional information in analysis of gene expression data.

Methods

Functional relationships between genes were derived by Latent Semantic Indexing (LSI) of Medline abstracts and used to calculate the function cohesion of gene sets. In this study, literature cohesion was applied in two ways. First, Literature-Based Functional Significance (LBFS) method was developed to calculate a p-value for the cohesion of differentially expressed genes (DEGs) in order to objectively evaluate the overall biological significance of the gene expression experiments. Second, Literature Aided Statistical Significance Threshold (LASST) was developed to determine the appropriate expression p-value threshold for a given experiment.

Results

We tested our methods on three different publicly available datasets. LBFS analysis demonstrated that only two experiments were significantly cohesive. For each experiment, we also compared the LBFS values of DEGs generated by four different statistical methods. We found that some statistical tests produced more functionally cohesive gene sets than others. However, no statistical test was consistently better for all experiments. This reemphasizes that a statistical test must be carefully selected for each expression study. Moreover, LASST analysis demonstrated that the expression p-value thresholds for some experiments were considerably lower (p < 0.02 and 0.01), suggesting that the arbitrary p-values and false discovery rate thresholds that are commonly used in expression studies may not be biologically sound.

Conclusions

We have developed robust and objective literature-based methods to evaluate the biological support for gene expression experiments and to determine the appropriate statistical significance threshold. These methods will assist investigators to more efficiently extract biologically meaningful insights from high throughput gene expression experiments.

Anthrax lethal toxin-induced gene expression changes in mouse lung

A major virulence factor of Bacillus anthracis is the anthrax Lethal Toxin (LeTx), a bipartite toxin composed of Protective Antigen and Lethal Factor. Systemic administration of LeTx to laboratory animals leads to death associated with vascular leakage and pulmonary edema. In this study, we investigated whether systemic exposure of mice to LeTx would induce gene expression changes associated with vascular/capillary leakage in lung tissue. We observed enhanced susceptibility of A/J mice to death by systemic LeTx administration compared to the C57BL/6 strain. LeTx-induced groups of both up- and down-regulated genes were observed in mouse lungs 6 h after systemic administration of wild type toxin compared to lungs of mice exposed to an inactive mutant form of the toxin. Lungs of the less susceptible C57BL/6 strain showed 80% fewer differentially expressed genes compared to lungs of the more sensitive A/J strain. Expression of genes known to regulate vascular permeability was modulated by LeTx in the lungs of the more susceptible A/J strain. Unexpectedly, the largest set of genes with altered expression was immune specific, characterized by the up-regulation of lymphoid genes and the down-regulation of myeloid genes. Transcripts encoding neutrophil chemoattractants, modulators of tumor regulation and angiogenesis were also differentially expressed in both mouse strains. These studies provide new directions for the investigation of vascular leakage and pulmonary edema induced by anthrax LeTx.