Single-cell transcriptomics reveals gene expression heterogeneity but suffers from stochastic dropout and characteristic bimodal expression distributions in which expression is either strongly non-zero or non-detectable. We propose a two-part, generalized linear model for such bimodal data that parameterizes both of these features. We argue that the cellular detection rate, the fraction of genes expressed in a cell, should be adjusted for as a source of nuisance variation. Our model provides gene set enrichment analysis tailored to single-cell data. It provides insights into how networks of co-expressed genes evolve across an experimental treatment. MAST is available at https://github.com/RGLab/MAST
 .

/pdf/mast-a-flexible-statistical-framework-for-assessing-2nt9ttfvra.pdf

MAST: a flexible statistical framework for assessing transcriptional changes and characterizing heterogeneity in single-cell RNA sequencing data

Coronavirus disease 2019 (COVID-19) represents a global crisis, yet major knowledge gaps remain about human immunity to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). We analyzed immune responses in 76 COVID-19 patients and 69 healthy individuals from Hong Kong and Atlanta, Georgia, United States. In the peripheral blood mononuclear cells (PBMCs) of COVID-19 patients, we observed reduced expression of human leukocyte antigen class DR (HLA-DR) and proinflammatory cytokines by myeloid cells as well as impaired mammalian target of rapamycin (mTOR) signaling and interferon-α (IFN-α) production by plasmacytoid dendritic cells. By contrast, we detected enhanced plasma levels of inflammatory mediators-including EN-RAGE, TNFSF14, and oncostatin M-which correlated with disease severity and increased bacterial products in plasma. Single-cell transcriptomics revealed a lack of type I IFNs, reduced HLA-DR in the myeloid cells of patients with severe COVID-19, and transient expression of IFN-stimulated genes. This was consistent with bulk PBMC transcriptomics and transient, low IFN-α levels in plasma during infection. These results reveal mechanisms and potential therapeutic targets for COVID-19.

https://science.sciencemag.org/content/sci/early/2020/08/10/science.abc6261.full.pdf

Systems biological assessment of immunity to mild versus severe COVID-19 infection in humans.

http://www.immunology.umn.edu/sites/immunology.umn.edu/files/1-s2.0-s0092867414015906-main.pdf

Variation in the Human Immune System Is Largely Driven by Non-Heritable Influences

Meta-analysis of gene expression data sets is increasingly performed to help identify robust molecular signatures and to gain insights into underlying biological processes. The complicated nature of such analyses requires both advanced statistics and innovative visualization strategies to support efficient data comparison, interpretation and hypothesis generation. NetworkAnalyst (http://www.networkanalyst.ca) is a comprehensive web-based tool designed to allow bench researchers to perform various common and complex meta-analyses of gene expression data via an intuitive web interface. By coupling well-established statistical procedures with state-of-the-art data visualization techniques, NetworkAnalyst allows researchers to easily navigate large complex gene expression data sets to determine important features, patterns, functions and connections, thus leading to the generation of new biological hypotheses. This protocol provides a step-wise description of how to effectively use NetworkAnalyst to perform network analysis and visualization from gene lists; to perform meta-analysis on gene expression data while taking into account multiple metadata parameters; and, finally, to perform a meta-analysis of multiple gene expression data sets. NetworkAnalyst is designed to be accessible to biologists rather than to specialist bioinformaticians. The complete protocol can be executed in ∼1.5 h. Compared with other similar web-based tools, NetworkAnalyst offers a unique visual analytics experience that enables data analysis within the context of protein-protein interaction networks, heatmaps or chord diagrams. All of these analysis methods provide the user with supporting statistical and functional evidence.

NetworkAnalyst for statistical, visual and network-based meta-analysis of gene expression data

Single-cell transcriptomic profiling enables the unprecedented interrogation of gene expression heterogeneity in rare cell populations that would otherwise be obscured in bulk RNA sequencing experiments. The stochastic nature of transcription is revealed in the bimodality of single-cell transcriptomic data, a feature shared across single-cell expression platforms. There is, however, a paucity of computational tools that take advantage of this unique characteristic. We present a new methodology to analyze single-cell transcriptomic data that models this bimodality within a coherent generalized linear modeling framework. We propose a two-part, generalized linear model that allows one to characterize biological changes in the proportions of cells that are expressing each gene, and in the positive mean expression level of that gene. We introduce the cellular detection rate, the fraction of genes turned on in a cell, and show how it can be used to simultaneously adjust for technical variation and so-called “extrinsic noise” at the single-cell level without the use of control genes. Our model permits direct inference on statistics formed by collections of genes, facilitating gene set enrichment analysis. The residuals defined by such models can be manipulated to interrogate cellular heterogeneity and gene-gene correlation across cells and conditions, providing insights into the temporal evolution of networks of co-expressed genes at the single-cell level. Using two single-cell RNA-seq datasets, including newly generated data from Mucosal Associated Invariant T (MAIT) cells, we show how model residuals can be used to identify significant changes across biologically relevant gene sets that are missed by other methods and characterize cellular heterogeneity in response to stimulation.

/pdf/mast-a-flexible-statistical-framework-for-assessing-2pdbdc32tw.pdf

MAST: A flexible statistical framework for assessing transcriptional changes and characterizing heterogeneity in single-cell RNA-seq data

Many vaccines induce protective immunity via antibodies. Systems biology approaches have been used to determine signatures that can be used to predict vaccine-induced immunity in humans, but whether there is a 'universal signature' that can be used to predict antibody responses to any vaccine is unknown. Here we did systems analyses of immune responses to the polysaccharide and conjugate vaccines against meningococcus in healthy adults, in the broader context of published studies of vaccines against yellow fever virus and influenza virus. To achieve this, we did a large-scale network integration of publicly available human blood transcriptomes and systems-scale databases in specific biological contexts and deduced a set of transcription modules in blood. Those modules revealed distinct transcriptional signatures of antibody responses to different classes of vaccines, which provided key insights into primary viral, protein recall and anti-polysaccharide responses. Our results elucidate the early transcriptional programs that orchestrate vaccine immunity in humans and demonstrate the power of integrative network modeling.

/pdf/molecular-signatures-of-antibody-responses-derived-from-a-4jr01332fd.pdf

Molecular signatures of antibody responses derived from a systems biology study of five human vaccines

Although pertussis disease is vaccine preventable, Washington State experienced a substantial rise in pertussis incidence beginning in 2011. By June 2012, the reported cases reached 2,520 (37.5 cases per 100,000 residents), a 1,300% increase compared with the same period in 2011. We assessed the molecular epidemiology of this statewide epidemic using 240 isolates collected from case patients reported from 19 of 39 Washington counties during 2012 to 2013. The typing methods included pulsed-field gel electrophoresis (PFGE), multilocus variable number tandem repeat analysis (MLVA), multilocus sequence typing (MLST), and pertactin gene (prn) mutational analysis. Using the scheme PFGE-MLVA-MLST-prn mutations-Prn deficiency, the 240 isolates comprised 65 distinct typing profiles. Thirty-one PFGE types were found, with the most common types, CDC013 (n = 51), CDC237 (n = 44), and CDC002 (n = 42), accounting for 57% of them. Eleven MLVA types were observed, mainly comprising type 27 (n = 183, 76%). Seven MLST types were identified, with the majority of the isolates typing as prn2-ptxP3-ptxA1-fim3-1 (n = 157, 65%). Four different prn mutations accounted for the 76% of isolates exhibiting pertactin deficiency. PFGE provided the highest discriminatory power (D = 0.87) and was found to be a more powerful typing method than MLVA and MLST combined (D = 0.67). This study provides evidence for the continued predominance of MLVA 27 and prn2-ptxP3-ptxA1 alleles, along with the reemergence of the fim3-1 allele. Our results indicate that the Bordetella pertussis population causing this epidemic was diverse, with a few molecular types predominating. The PFGE, MLVA, and MLST profiles were consistent with the predominate types circulating in the United States and other countries. For prn, several mutations were present in multiple molecular types.

Andrea Milton

Papers

Molecular signatures of antibody responses derived from a systems biology study of five human vaccines

Molecular Epidemiology of the Pertussis Epidemic in Washington State in 2012