Semisupervised Generative Autoencoder for Single-Cell Data.

Single-cell transcriptomics offers a tool to study the diversity of cell phenotypes through snapshots of the abundance of mRNA in individual cells. Often there is additional information available besides the single-cell gene expression counts, such as bulk transcriptome data from the same tissue, or quantification of surface protein levels from the same cells. In this study, we propose models based on the Bayesian deep learning approach, where protein quantification, available as CITE-seq counts, from the same cells is used to constrain the learning process, thus forming a SemI-SUpervised generative Autoencoder (SISUA) model. The generative model is based on the deep variational autoencoder (VAE) neural network architecture.

Genomic and molecular characterization of preterm birth.

Preterm birth (PTB) complications are the leading cause of long-term morbidity and mortality in children. By using whole blood samples, we integrated whole-genome sequencing (WGS), RNA sequencing (RNA-seq), and DNA methylation data for 270 PTB and 521 control families. We analyzed this combined dataset to identify genomic variants associated with PTB and secondary analyses to identify variants associated with very early PTB (VEPTB) as well as other subcategories of disease that may contribute to PTB. We identified differentially expressed genes (DEGs) and methylated genomic loci and performed expression and methylation quantitative trait loci analyses to link genomic variants to these expression and methylation changes. We performed enrichment tests to identify overlaps between new and known PTB candidate gene systems. We identified 160 significant genomic variants associated with PTB-related phenotypes. The most significant variants, DEGs, and differentially methylated loci were associated with VEPTB. Integration of all data types identified a set of 72 candidate biomarker genes for VEPTB, encompassing genes and those previously associated with PTB. Notably, PTB-associated genes RAB31 and RBPJ were identified by all three data types (WGS, RNA-seq, and methylation). Pathways associated with VEPTB include EGFR and prolactin signaling pathways, inflammation- and immunity-related pathways, chemokine signaling, IFN-γ signaling, and Notch1 signaling. Progress in identifying molecular components of a complex disease is aided by integrated analyses of multiple molecular data types and clinical data. With these data, and by stratifying PTB by subphenotype, we have identified associations between VEPTB and the underlying biology.

Identification of PGN_1123 as the Gene Encoding Lipid A Deacylase, an Enzyme Required for Toll-Like Receptor 4 Evasion, in Porphyromonas gingivalis.

Removal of one acyl chain from bacterial lipid A by deacylase activity is a mechanism used by many pathogenic bacteria to evade the host's Toll-like receptor 4 (TLR4)-mediated innate immune response. In Porphyromonas gingivalis, a periodontal pathogen, lipid A deacylase activity converts a majority of the initially synthesized penta-acylated lipid A, a TLR4 agonist, to tetra-acylated structures, which effectively evade TLR4 sensing by being either inert or antagonistic at TLR4. In this paper, we report successful identification of the gene that encodes the P. gingivalis lipid A deacylase enzyme. This gene, PGN_1123 in P. gingivalis 33277, is highly conserved within P. gingivalis, and putative orthologs are phylogenetically restricted to the Bacteroidetes phylum. Lipid A of ΔPGN_1123 mutants is penta-acylated and devoid of tetra-acylated structures, and the mutant strain provokes a strong TLR4-mediated proinflammatory response, in contrast to the negligible response elicited by wild-type P. gingivalis Heterologous expression of PGN_1123 in Bacteroides thetaiotaomicron promoted lipid A deacylation, confirming that PGN_1123 encodes the lipid A deacylase enzyme.IMPORTANCE Periodontitis, commonly referred to as gum disease, is a chronic inflammatory condition that affects a large proportion of the population. Porphyromonas gingivalis is a bacterium closely associated with periodontitis, although how and if it is a cause for the disease are not known. It has a formidable capacity to dampen the host's innate immune response, enabling its persistence in diseased sites and triggering microbial dysbiosis in animal models of infection. P. gingivalis is particularly adept at evading the host's TLR4-mediated innate immune response by modifying the structure of lipid A, the TLR4 ligand. In this paper, we report identification of the gene encoding lipid A deacylase, a key enzyme that modifies lipid A to TLR4-evasive structures.

Using Incomplete Trios to Boost Confidence in Family Based Association Studies.

Most currently available family based association tests are designed to account only for nuclear families with complete genotypes for parents as well as offspring. Due to the availability of increasingly less expensive generation of whole genome sequencing information, genetic studies are able to collect data for more families and from large family cohorts with the goal of improving statistical power. However, due to missing genotypes, many families are not included in the family based association tests, negating the benefits of large scale sequencing data. Here, we present the CIFBAT method to use incomplete families in Family Based Association Test (FBAT) to evaluate robustness against missing data. CIFBAT uses quantile intervals of the FBAT statistic by randomly choosing valid completions of incomplete family genotypes based on Mendelian inheritance rules. By considering all valid completions equally likely and computing quantile intervals over many randomized iterations, CIFBAT avoids assumption of a homogeneous population structure or any particular missingness pattern in the data. Using simulated data, we show that the quantile intervals computed by CIFBAT are useful in validating robustness of the FBAT statistic against missing data and in identifying genomic markers with higher precision. We also propose a novel set of candidate genomic markers for uterine related abnormalities from analysis of familial whole genome sequences, and provide validation for a previously established set of candidate markers for Type 1 diabetes. We have provided a software package that incorporates TDT, robustTDT, FBAT, and CIFBAT. The data format proposed for the software uses half the memory space that the standard FBAT format (PED) files use, making it efficient for large scale genome wide association studies.

Big biomedical data as the key resource for discovery science.

Modern biomedical data collection is generating exponentially more data in a multitude of formats. This flood of complex data poses significant opportunities to discover and understand the critical interplay among such diverse domains as genomics, proteomics, metabolomics, and phenomics, including imaging, biometrics, and clinical data. The Big Data for Discovery Science Center is taking an "-ome to home" approach to discover linkages between these disparate data sources by mining existing databases of proteomic and genomic data, brain images, and clinical assessments. In support of this work, the authors developed new technological capabilities that make it easy for researchers to manage, aggregate, manipulate, integrate, and model large amounts of distributed data. Guided by biological domain expertise, the Center's computational resources and software will reveal relationships and patterns, aiding researchers in identifying biomarkers for the most confounding conditions and diseases, such as Parkinson's and Alzheimer's.

Gene-pair expression signatures reveal lineage control.

The distinct cell types of multicellular organisms arise owing to constraints imposed by gene regulatory networks on the collective change of gene expression across the genome, creating self-stabilizing expression states, or attractors. We curated human expression data comprising 166 cell types and 2,602 transcription-regulating genes and developed a data-driven method for identifying putative determinants of cell fate built around the concept of expression reversal of gene pairs, such as those participating in toggle-switch circuits. This approach allows us to organize the cell types into their ontogenic lineage relationships. Our method identifies genes in regulatory circuits that control neuronal fate, pluripotency and blood cell differentiation, and it may be useful for prioritizing candidate factors for direct conversion of cell fate.

A functional genomic yeast screen to identify pathogenic bacterial proteins.

Many bacterial pathogens promote infection and cause disease by directly injecting into host cells proteins that manipulate eukaryotic cellular processes. Identification of these translocated proteins is essential to understanding pathogenesis. Yet, their identification remains limited. This, in part, is due to their general sequence uniqueness, which confounds homology-based identification by comparative genomic methods. In addition, their absence often does not result in phenotypes in virulence assays limiting functional genetic screens. Translocated proteins have been observed to confer toxic phenotypes when expressed in the yeast Saccharomyces cerevisiae. This observation suggests that yeast growth inhibition can be used as an indicator of protein translocation in functional genomic screens. However, limited information is available regarding the behavior of non-translocated proteins in yeast. We developed a semi-automated quantitative assay to monitor the growth of hundreds of yeast strains in parallel. We observed that expression of half of the 19 Shigella translocated proteins tested but almost none of the 20 non-translocated Shigella proteins nor approximately 1,000 Francisella tularensis proteins significantly inhibited yeast growth. Not only does this study establish that yeast growth inhibition is a sensitive and specific indicator of translocated proteins, but we also identified a new substrate of the Shigella type III secretion system (TTSS), IpaJ, previously missed by other experimental approaches. In those cases where the mechanisms of action of the translocated proteins are known, significant yeast growth inhibition correlated with the targeting of conserved cellular processes. By providing positive rather than negative indication of activity our assay complements existing approaches for identification of translocated proteins. In addition, because this assay only requires genomic DNA it is particularly valuable for studying pathogens that are difficult to genetically manipulate or dangerous to culture.

Yeast functional genomic screens lead to identification of a role for a bacterial effector in innate immunity regulation.

Numerous bacterial pathogens manipulate host cell processes to promote infection and ultimately cause disease through the action of proteins that they directly inject into host cells. Identification of the targets and molecular mechanisms of action used by these bacterial effector proteins is critical to understanding pathogenesis. We have developed a systems biological approach using the yeast Saccharomyces cerevisiae that can expedite the identification of cellular processes targeted by bacterial effector proteins. We systematically screened the viable yeast haploid deletion strain collection for mutants hypersensitive to expression of the Shigella type III effector OspF. Statistical data mining of the results identified several cellular processes, including cell wall biogenesis, which when impaired by a deletion caused yeast to be hypersensitive to OspF expression. Microarray experiments revealed that OspF expression resulted in reversed regulation of genes regulated by the yeast cell wall integrity pathway. The yeast cell wall integrity pathway is a highly conserved mitogen-activated protein kinase (MAPK) signaling pathway, normally activated in response to cell wall perturbations. Together these results led us to hypothesize and subsequently demonstrate that OspF inhibited both yeast and mammalian MAPK signaling cascades. Furthermore, inhibition of MAPK signaling by OspF is associated with attenuation of the host innate immune response to Shigella infection in a mouse model. These studies demonstrate how yeast systems biology can facilitate functional characterization of pathogenic bacterial effector proteins.