Integration Of Gene Expression Data Research Articles

Predicting protein complexes in protein interaction networks using Mapper and graph convolution networks

Protein complexes are groups of interacting proteins that are central to multiple biological processes. Studying protein complexes can enhance our understanding of cellular functions and malfunctions and thus support the development of effective disease treatments. High-throughput experimental techniques allow the generation of large-scale protein-protein interaction datasets. Accordingly, various computational approaches to predict protein complexes from protein-protein interactions were presented in the literature. They are typically based on networks in which nodes and edges represent proteins and their interactions, respectively. State-of-the-art approaches mainly rely on clustering static networks to identify complexes. However, since protein interactions are highly dynamic in nature, recent approaches seek to model such dynamics by typically integrating gene expression data and identifying protein complexes accordingly. We propose MComplex, a method that utilizes time-series gene expression with interaction data to generate a temporal network which is passed to a generative adversarial network whose generator is a graph convolutional network. This creates embeddings which are then analyzed using a modified graph-based version of the Mapper algorithm to predict corresponding protein complexes. We test our approach on multiple benchmark datasets and compare identified complexes against gold-standard protein complex datasets. Our results show that MComplex outperforms existing methods in several evaluation aspects, namely recall and maximum matching ratio as well as a composite score covering aggregated evaluation measures. The code and data are available for free download from https://github.com/LeonardoDaou/MComplex.

GEDI: An R Package for Integration of Transcriptomic Data from Multiple Platforms for Bioinformatics Applications.

Transcriptomic data is often expensive and difficult to generate in large cohorts relative to genomic data; therefore, it is often important to integrate multiple transcriptomic datasets from both microarray- and next generation sequencing (NGS)-based transcriptomic data across similar experiments or clinical trials to improve analytical power and discovery of novel transcripts and genes. However, transcriptomic data integration presents a few challenges including reannotation and batch effect removal. We developed the Gene Expression Data Integration (GEDI) R package to enable transcriptomic data integration by combining existing R packages. With just four functions, the GEDI R package makes constructing a transcriptomic data integration pipeline straightforward. Together, the functions overcome the complications in transcriptomic data integration by automatically reannotating the data and removing the batch effect. The removal of the batch effect is verified with principal component analysis and the data integration is verified using a logistic regression model with forward stepwise feature selection. To demonstrate the functionalities of the GEDI package, we integrated five bovine endometrial transcriptomic datasets from the NCBI Gene Expression Omnibus. These transcriptomic datasets were from multiple high-throughput platforms, namely, array-based Affymetrix and Agilent platforms, and NGS-based Illumina paired-end RNA-seq platform. Furthermore, we compared the GEDI package to existing tools and found that GEDI is the only tool that provides a full transcriptomic data integration pipeline including verification of both batch effect removal and data integration for downstream genomic and bioinformatics applications. © 2024 The Author(s). Current Protocols published by Wiley Periodicals LLC. Basic Protocol 1: ReadGE, a function to import gene expression datasets Basic Protocol 2: GEDI, a function to reannotate and merge gene expression datasets Basic Protocol 3: BatchCorrection, a function to remove batch effects from gene expression data Basic Protocol 4: VerifyGEDI, a function to confirm successful integration of gene expression data.

Prioritizing disease-related rare variants by integrating gene expression data.

Rare variants, comprising the vast majority of human genetic variations, are likely to have more deleterious impact in the context of human diseases compared to common variants. Here we present carrier statistic, a statistical framework to prioritize disease-related rare variants by integrating gene expression data. By quantifying the impact of rare variants on gene expression, carrier statistic can prioritize those rare variants that have large functional consequence in the patients. Through simulation studies and analyzing real multi-omics dataset, we demonstrated that carrier statistic is applicable in studies with limited sample size (a few hundreds) and achieves substantially higher sensitivity than existing rare variants association methods. Application to Alzheimer's disease reveals 16 rare variants within 15 genes with extreme carrier statistics. We also found strong excess of rare variants among the top prioritized genes in patients compared to that in healthy individuals. The carrier statistic method can be applied to various rare variant types and is adaptable to other omics data modalities, offering a powerful tool for investigating the molecular mechanisms underlying complex diseases.

Accurately deciphering spatial domains for spatially resolved transcriptomics with stCluster.

Spatial transcriptomics provides valuable insights into gene expression within the native tissue context, effectively merging molecular data with spatial information to uncover intricate cellular relationships and tissue organizations. In this context, deciphering cellular spatial domains becomes essential for revealing complex cellular dynamics and tissue structures. However, current methods encounter challenges in seamlessly integrating gene expression data with spatial information, resulting in less informative representations of spots and suboptimal accuracy in spatial domain identification. We introduce stCluster, a novel method that integrates graph contrastive learning with multi-task learning to refine informative representations for spatial transcriptomic data, consequently improving spatial domain identification. stCluster first leverages graph contrastive learning technology to obtain discriminative representations capable of recognizing spatially coherent patterns. Through jointly optimizing multiple tasks, stCluster further fine-tunes the representations to be able to capture complex relationships between gene expression and spatial organization. Benchmarked against six state-of-the-art methods, the experimental results reveal its proficiency in accurately identifying complex spatial domains across various datasets and platforms, spanning tissue, organ, and embryo levels. Moreover, stCluster can effectively denoise the spatial gene expression patterns and enhance the spatial trajectory inference. The source code of stCluster is freely available at https://github.com/hannshu/stCluster.

Retinal microglia express more MHC class I and promote greater T-cell-driven inflammation than brain microglia.

Macrophage function is determined by microenvironment and origin. Brain and retinal microglia are both derived from yolk sac progenitors, yet their microenvironments differ. Utilizing single-cell RNA sequencing (scRNA-seq) data from mice, we tested the hypothesis that retinal and brain microglia exhibit distinct transcriptional profiles due to their unique microenvironments. Eyes and brains from 2-4 month wildtype mice were combined (20 eyes; 3 brains) to yield one biologically diverse sample per organ. Each tissue was digested into single cell suspensions, enriched for immune cells, and sorted for scRNA-seq. Analysis was performed in Seurat v3 including clustering, integration, and differential expression. Multi-parameter flow cytometry was used for validation of scRNA-seq results. Lymphocytic choriomeningitis virus (LCMV) Clone 13, which produces a systemic, chronic, and neurotropic infection, was used to validate scRNA-seq and flow cytometry results in vivo. Cluster analysis of integrated gene expression data from eye and brain identified 6 Tmem119 + P2ry12 + microglial clusters. Differential expression analysis revealed that eye microglia were enriched for more pro-inflammatory processes including antigen processing via MHC class I (14.0-fold, H2-D1 and H2-K1) and positive regulation of T-cell immunity (8.4-fold) compared to brain microglia. Multi-parameter flow cytometry confirmed that retinal microglia expressed 3.2-fold greater H2-Db and 263.3-fold more H2-Kb than brain microglia. On Day 13 and 29 after LCMV infection, CD8+ T-cell density was greater in the retina than the brain. Our data demonstrate that the microenvironment of retina and brain differs, resulting in microglia-specific gene expression changes. Specifically, retinal microglia express greater MHC class I by scRNA-seq and multi-parameter flow cytometry, resulting in a possibly enhanced capability to stimulate CD8+ T-cell inflammation during LCMV infection. These results may explain tissue-specific differences between retina and brain during systemic viral infections and CD8+ T-cell driven autoimmune disease.

Meta-analysis of endometrial transcriptome data reveals novel molecular targets for recurrent implantation failure.

Gene expression analysis of the endometrium has been shown to be a useful approach for identifying the molecular signatures and pathways involved in recurrent implantation failure (RIF). Nevertheless, individual studies have limitations in terms of study design, methodology and analysis to detect minor changes in expression levels or identify novel gene signatures associated with RIF. To overcome this, we conducted an in silico meta-analysis of nine studies, the systematic collection and integration of gene expression data, utilizing rigorous selection criteria and statistical techniques to ensure the robustness of our findings. Our meta-analysis successfully unveiled a meta-signature of 49 genes closely associated with RIF. Of these genes, 38 were upregulated and 11 downregulated in RIF patients' endometrium and believed to participate in key processes like cell differentiation, communication, and adhesion. GADD45A, IGF2, and LIF, known for their roles in implantation, were identified, along with lesser-studied genes like OPRK1, PSIP1, SMCHD1, and SOD2 related to female infertility. Many of these genes are involved in MAPK and PI3K-Akt pathways, indicating their role in inflammation. We also investigated to look for key miRNAs regulating these 49 dysregulated mRNAs as potential diagnostic biomarkers. Along with this, we went to associate protein-protein interactions of 49 genes, and we could recognize one cluster consisting of 11 genes (consisted of 22 nodes and 11 edges) with the highest score (p = 0.001). Finally, we validated some of the genes by qRT-PCR in our samples. In summary, the meta-signature genes hold promise for improving RIF patient identification and facilitating the development of personalized treatment strategies, illuminating the multifaceted nature of this complex condition.

Dissecting the genetic overlap between severe mental disorders and markers of cellular aging: Identification of pleiotropic genes and druggable targets.

Patients with severe mental disorders such as bipolar disorder (BD), schizophrenia (SCZ) and major depressive disorder (MDD) show a substantial reduction in life expectancy, increased incidence of comorbid medical conditions commonly observed with advanced age and alterations of aging hallmarks. While severe mental disorders are heritable, the extent to which genetic predisposition might contribute to accelerated cellular aging is not known. We used bivariate causal mixture models to quantify the trait-specific and shared architecture of mental disorders and 2 aging hallmarks (leukocyte telomere length [LTL] and mitochondrial DNA copy number), and the conjunctional false discovery rate methodto detect shared genetic loci. We integrated gene expression data from brain regions from GTEx and used different tools to functionally annotate identified loci and investigate their druggability. Aging hallmarks showed low polygenicity compared with severe mental disorders. We observed a significant negative global genetic correlation between MDD and LTL (rg = -0.14, p = 6.5E-10), and no significant results for other severe mental disorders or for mtDNA-cn. However, conditional QQ plots and bivariate causal mixture models pointed to significant pleiotropy among all severe mental disorders and aging hallmarks. We identified genetic variants significantly shared between LTL and BD (n = 17), SCZ (n = 55) or MDD (n = 19), or mtDNA-cn and BD (n = 4), SCZ (n = 12) or MDD (n = 1), with mixed direction of effects. The exonic rs7909129 variant in the SORCS3 gene, encoding a member of the retromer complex involved in protein trafficking and intracellular/intercellular signaling, was associated with shorter LTL and increased predisposition to all severe mental disorders. Genetic variants underlying risk of SCZ or MDD and shorter LTL modulate expression of several druggable genes in different brain regions. Genistein, a phytoestrogen with anti-inflammatory and antioxidant effects, was an upstream regulator of 2 genes modulated by variants associated with risk of MDD and shorter LTL. While our results suggest that shared heritability might play a limited role in contributing to accelerated cellular aging in severe mental disorders, we identified shared genetic determinants and prioritized different druggable targets and compounds.

Archetypal clustering reveals physiological mechanisms linking milk yield and fertility in dairy cattle

Fertility in dairy cattle has declined as an unintended consequence of single trait selection for high milk yield. The unfavorable genetic correlation between milk yield and fertility is now well-documented, however, the underlying physiological mechanisms are still uncertain. To understand the relationship between these traits, we developed a method that clusters variants with similar patterns of effects and, after the integration of gene expression data, identifies the genes through which they are likely to act. Biological processes that are enriched in the genes of each cluster were then identified. We identified several clusters with unique patterns of effects. One of the clusters included variants associated with increased milk yield and decreased fertility, where the 'archetypal' variant (i.e., the one with the largest effect) was associated with the gene GC, while others were associated with TRIM32, LRRK2, and U6. These genes have been linked to transcription and alternative splicing, suggesting that these processes are likely contributors to the unfavorable relationship between the 2 traits. Another cluster, with archetypal variant near DGAT1 and including variants associated with CDH2, BTRC, SFRP2, ZFHX3, and SLITRK5, appeared to affect milk yield but have little effect on fertility. These genes have been linked to insulin, adipose tissue, and energy metabolism. A third cluster with archetypal variant near ZNF613 and including variants associated with ROBO1, EFNA5, PALLD, GPC6, and PTPRT were associated with fertility but not milk yield. These genes have been linked to GnRH neuronal migration, embryonic development, and/or ovarian function. The use of archetypal clustering to group variants with similar patterns of effects may assist in identifying the biological processes underlying correlated traits. The method is hypothesis-generating and requires experimental confirmation. However, we have uncovered several novel mechanisms potentially affecting milk production and fertility such as GnRH neuronal migration. We anticipate our method to be a starting point for experimental research into novel pathways which have been previously unexplored within the context of dairy production.

RECCIPE: A new framework assessing localized cell-cell interaction on gene expression in multicellular ST data.

Cell-cell interaction (CCI) plays a pivotal role in cellular communication within the tissue microenvironment. The recent development of spatial transcriptomics (ST) technology and associated data analysis methods has empowered researchers to systematically investigate CCI. However, existing methods are tailored to single-cell resolution datasets, whereas the majority of ST platforms lack such resolution. Additionally, the detection of CCI through association screening based on ST data, which has complicated dependence structure, necessitates proper control of false discovery rates due to the multiple hypothesis testing issue in high dimensional spaces. To address these challenges, we introduce RECCIPE, a novel method designed for identifying cell signaling interactions across multiple cell types in spatial transcriptomic data. RECCIPE integrates gene expression data, spatial information and cell type composition in a multivariate regression framework, enabling genome-wide screening for changes in gene expression levels attributed to CCIs. We show that RECCIPE not only achieves high accuracy in simulated datasets but also provides new biological insights from real data obtained from a mouse model of Alzheimer's disease (AD). Overall, our framework provides a useful tool for studying impact of cell-cell interactions on gene expression in multicellular systems.

A Deep Neural Network for Predicting Synergistic Drug Combinations on Cancer.

The exploration of drug combinations presents an opportunity to amplify therapeutic effectiveness while alleviating undesirable side effects. Nevertheless, the extensive array of potential combinations poses challenges in terms of cost and time constraints for experimental screening. Thus, it is crucial to narrow down the search space. Deep learning approaches have gained widespread popularity in predicting synergistic drug combinations tailored for specific cell lines in vitro settings. In the present study, we introduce a novel method termed GTextSyn, which utilizes the integration of gene expression data and chemical structure information for the prediction of synergistic effects in drug combinations. GTextSyn employs a sentence classification model within the domain of Natural Language Processing (NLP), wherein drugs and cell lines are regarded as entities possessing biochemical relevance. Meanwhile, combinations of drug pairs and cell lines are construed as sentences with biochemical relational significance. To assess the efficacy of GTextSyn, we conduct a comparative analysis with alternative deep learning approaches using a standard benchmark dataset. The results from a five-fold cross-validation demonstrate a 49.5% reduction in Mean Square Error (MSE) achieved by GTextSyn, surpassing the performance of the next best method in the regression task. Furthermore, we conduct a comprehensive literature survey on the predicted novel drug combinations and find substantial support from prior experimental studies for many of the combinations identified by GTextSyn.

Novel 2 Gene Signatures Associated With Breast Cancer Proliferation: Insights From Predictive Differential Gene Expression Analysis

The use of proliferation markers provides valuable information about the rate of tumor growth, which can guide treatment decisions. However, there is still a lack of consensus regarding the optimal molecular markers or tests to use in clinical practice. Integrating gene expression data with clinical and histopathologic parameters enhances our understanding of disease processes, facilitates the identification of precise prognostic predictors, and supports the development of effective therapeutic strategies. The purpose of this study was to apply an integrated approach that combines morphologic, clinical, and bioinformatic data to reveal effective regulators of proliferation. Whole-slide images generated from hematoxylin-and-eosin–stained sections of The Cancer Genome Atlas (TCGA) breast cancer (BC) database (n = 1053) alongside their transcriptomic and clinical data were used to identify genes differentially expressed between tumors with high and low mitotic scores. Genes enriched in the cell-cycle pathway were used to predict the protein-protein interaction (PPI) network. Ten hub genes (ORC6, SKP2, SMC1B, CDKN2A, CDC25B, E2F1, E2F2, ORC1, PTTG1, and CDC25A) were identified using CytoHubba a Cytoscape plugin. In a multivariate Cox regression model, ORC6 and SKP2 were predictors of survival independent of existing methods of proliferation assessment including mitotic score and Ki67. The prognostic ability of these genes was validated using the Molecular Taxonomy of Breast Cancer International Consortium, Nottingham cohort, Uppsala cohort, and a combined multicentric cohort. The protein expression of these 2 genes was investigated on a large cohort of BC cases, and they were significantly associated with poor prognosis and patient outcome. A positive correlation between ORC6 and SKP2 mRNA and protein expression was observed. Our study has identified 2 gene signatures, ORC6 and SKP2, which play a significant role in BC proliferation. These genes surpassed both mitotic scores and Ki67 in multivariate analysis. Their identification provides potential opportunities for the development of targeted treatments for patients with BC.

Mechanistic analysis of Th2-type inflammatory factors in asthma.

The main pathological features of asthma are widespread chronic inflammation of the airways and restricted ventilation due to airway remodeling, which involves changes in a range of regulatory pathways. While the role of T helper type 2 (Th2)-related inflammatory factors in this process is known, the detailed understanding of how genes affect protein functions during airway remodeling is still lacking. This study aims to fill this knowledge gap by integrating gene expression data and protein function analysis, providing new scientific insights for a deeper understanding of the mechanisms of airway remodeling and for further development of asthma treatment strategies. In this study, the mechanism of Th2-related inflammatory factors in tracheal remodeling was studied through differentially expressed gene (DEG) screening, enrichment analysis, protein-protein interaction (PPI) network construction, machine learning, and the construction of a line graph model. Our study revealed that S100A14, KRT6A, S100A2, ABCA13, UBE2C, RASSF10, PSCA, PLAT, and TIMP1 may be the key genes for airway remodeling; epithelial-mesenchymal transition (EMT)-related genes GEM, TPM4, SLC6A8, and SNTB1 may be involved in airway remodeling due to asthma; IL6 may affect the occurrence of airway remodeling by binding to UBE2C protein or by regulating GEM genes, respectively; IL6 and IL9 may affect the occurrence of airway remodeling by regulating the downstream Toll-like receptor (TLR) signaling pathway and thus IL6 and IL9 may influence the occurrence of tracheal remodeling by regulating downstream TLR signaling pathways. This study further mined the asthma gene microarray database through bioinformatics analysis and identified key genes and important pathways affecting airway remodeling in asthma patients, providing new ideas to uncover the mechanism of airway remodeling due to asthma and then seek new therapeutic targets.

GeneCompete: an integrative tool of a novel union algorithm with various ranking techniques for multiple gene expression data

Background Identifying the genes responsible for diseases requires precise prioritization of significant genes. Gene expression analysis enables differentiation between gene expressions in disease and normal samples. Increasing the number of high-quality samples enhances the strength of evidence regarding gene involvement in diseases. This process has led to the discovery of disease biomarkers through the collection of diverse gene expression data. Methods This study presents GeneCompete, a web-based tool that integrates gene expression data from multiple platforms and experiments to identify the most promising biomarkers. GeneCompete incorporates a novel union strategy and eight well-established ranking methods, including Win-Loss, Massey, Colley, Keener, Elo, Markov, PageRank, and Bi-directional PageRank algorithms, to prioritize genes across multiple gene expression datasets. Each gene in the competition is assigned a score based on log-fold change values, and significant genes are determined as winners. Results We tested the tool on the expression datasets of Hypertrophic cardiomyopathy (HCM) and the datasets from Microarray Quality Control (MAQC) project, which include both microarray and RNA-Sequencing techniques. The results demonstrate that all ranking scores have more power to predict new occurrence datasets than the classical method. Moreover, the PageRank method with a union strategy delivers the best performance for both up-regulated and down-regulated genes. Furthermore, the top-ranking genes exhibit a strong association with the disease. For MAQC, the two-sides ranking score shows a high relationship with TaqMan validation set in all log-fold change thresholds. Conclusion GeneCompete is a powerful web-based tool that revolutionizes the identification of disease-causing genes through the integration of gene expression data from multiple platforms and experiments.

BloodSpot 3.0: a database of gene and protein expression data in normal and malignant haematopoiesis.

BloodSpot is a specialised database integrating gene expression data from acute myeloid leukaemia (AML) patients related to blood cell development and maturation. The database and interface has helped numerous researchers and clinicians to quickly get an overview of gene expression patterns in healthy and malignant haematopoiesis. Here, we present an update to our framework that includes protein expression data of sorted single cells. With this update we also introduce datasets broadly spanning age groups, which many users haverequested, with particular interest for researchers studying paediatric leukaemias. The backend of the database has been rewritten and migrated to a cloud-based environment to accommodate the growth, and provide a better user-experience for our many international users. Users can now enjoy faster transfer speeds and a more responsive interface. In conclusion, the continuing popularity of the database and emergence of new data modalities has prompted us to rewrite and futureproof the back-end, including paediatric centric views, as well as single cell protein data, allowing us to keep the database updated and relevant for the years to come. The database is freelyavailable at www.bloodspot.eu.

Integrating gene expression data into a genome-scale metabolic model to identify reprogramming during adaptive evolution.

The development of a method for identifying latent reprogramming in gene expression data resulting from adaptive laboratory evolution (ALE) in response to genetic or environmental perturbations has been a challenge. In this study, a method called Metabolic Reprogramming Identifier (MRI), based on the integration of expression data to a genome-scale metabolic model has been developed. To identify key genes playing the main role in reprogramming, a MILP problem is presented and maximization of an adaptation score as a criterion indicating a pattern of using metabolism with maximum utilization of gene expression resources is defined as an objective function. Then, genes with complete expression usage and significant expression differences between wild-type and evolved strains were selected as key genes for reprogramming. This score is also applied to evaluate the compatibility of expression patterns with maximal use of key genes. The method was implemented to investigate the reprogramming of Escherichia coli during adaptive evolution caused by changing carbon sources. cyoC and cydB responsible for establishing proton gradient across the inner membrane were identified to be vital in the E. coli reprogramming when switching from glucose to lactate. These results indicate the importance of the inner membrane in reprogramming of E. coli to adapt to the new environment. The method predicts no reprogramming occurs during the evolution for growth on glycerol.

PathTracer Comprehensively Identifies Hypoxia-Induced Dormancy Adaptations in Mycobacterium tuberculosis.

Mining large-scale data to discover biologically relevant information remains a challenge despite the rapid development of bioinformatics tools. Here, we have developed a new tool, PathTracer, to identify biologically relevant information flows by mining genome-wide protein-protein interaction networks following integration of gene expression data. PathTracer successfully mines interactions between genes and traces the most perturbed paths of perceived activities under the conditions of the study. We further demonstrated the utility of this tool by identifying adaptation mechanisms of hypoxia-induced dormancy in Mycobacterium tuberculosis (Mtb).

ADVANCING THE UNDERSTANDING OF CLINICAL SEPSIS USING GENE EXPRESSION-DRIVEN MACHINE LEARNING TO IMPROVE PATIENT OUTCOMES.

Sepsis remains a major challenge that necessitates improved approaches to enhance patient outcomes. This study explored the potential of machine learning (ML) techniques to bridge the gap between clinical data and gene expression information to better predict and understand sepsis. We discuss the application of ML algorithms, including neural networks, deep learning, and ensemble methods, to address key evidence gaps and overcome the challenges in sepsis research. The lack of a clear definition of sepsis is highlighted as a major hurdle, but ML models offer a workaround by focusing on endpoint prediction. We emphasize the significance of gene transcript information and its use in ML models to provide insights into sepsis pathophysiology and biomarker identification. Temporal analysis and integration of gene expression data further enhance the accuracy and predictive capabilities of ML models for sepsis. Although challenges such as interpretability and bias exist, ML research offers exciting prospects for addressing critical clinical problems, improving sepsis management, and advancing precision medicine approaches. Collaborative efforts between clinicians and data scientists are essential for the successful implementation and translation of ML models into clinical practice. Machine learning has the potential to revolutionize our understanding of sepsis and significantly improve patient outcomes. Further research and collaboration between clinicians and data scientists are needed to fully understand the potential of ML in sepsis management.

Metabolic modeling of sex-specific liver tissue suggests mechanism of differences in toxicological responses.

Male subjects in animal and human studies are disproportionately used for toxicological testing. This discrepancy is evidenced in clinical medicine where females are more likely than males to experience liver-related adverse events in response to xenobiotics. While previous work has shown gene expression differences between the sexes, there is a lack of systems-level approaches to understand the direct clinical impact of these differences. Here, we integrate gene expression data with metabolic network models to characterize the impact of transcriptional changes of metabolic genes in the context of sex differences and drug treatment. We used Tasks Inferred from Differential Expression (TIDEs), a reaction-centric approach to analyzing differences in gene expression, to discover that several metabolic pathways exhibit sex differences including glycolysis, fatty acid metabolism, nucleotide metabolism, and xenobiotics metabolism. When TIDEs is used to compare expression differences in treated and untreated hepatocytes, we find several subsystems with differential expression overlap with the sex-altered pathways such as fatty acid metabolism, purine and pyrimidine metabolism, and xenobiotics metabolism. Finally, using sex-specific transcriptomic data, we create individual and averaged male and female liver models and find differences in the pentose phosphate pathway and other metabolic pathways. These results suggest potential sex differences in the contribution of the pentose phosphate pathway to oxidative stress, and we recommend further research into how these reactions respond to hepatotoxic pharmaceuticals.

Lipidomic QTL in Diversity Outbred mice identifies a novel function for α/β hydrolase domain 2 (Abhd2) as an enzyme that metabolizes phosphatidylcholine and cardiolipin.

We and others have previously shown that genetic association can be used to make causal connections between gene loci and small molecules measured by mass spectrometry in the bloodstream and in tissues. We identified a locus on mouse chromosome 7 where several phospholipids in liver showed strong genetic association to distinct gene loci. In this study, we integrated gene expression data with genetic association data to identify a single gene at the chromosome 7 locus as the driver of the phospholipid phenotypes. The gene encodes α/β-hydrolase domain 2 (Abhd2), one of 23 members of the ABHD gene family. We validated this observation by measuring lipids in a mouse with a whole-body deletion of Abhd2. The Abhd2KO mice had a significant increase in liver levels of phosphatidylcholine and phosphatidylethanolamine. Unexpectedly, we also found a decrease in two key mitochondrial lipids, cardiolipin and phosphatidylglycerol, in male Abhd2KO mice. These data suggest that Abhd2 plays a role in the synthesis, turnover, or remodeling of liver phospholipids.

21-OR: Insulin and Glucagon Regulate a Novel Bidirectional Long Noncoding RNA/Protein-Coding Gene Pair That Control Cell Sensitivity to Metabolic Stress

An imbalance of glucagon and insulin actions in type 2 diabetes (T2D) results in the dysregulation of glucose and lipid metabolism and metabolic stress throughout the body. Integrating gene expression data in multiple models of T2D and metabolic syndromes, we have identified a novel gene pair, that share the same bidirectional promoter, are coordinately upregulated by glucagon and downregulated by insulin in the liver in a variety of physiological and pathophysiological states. This gene pair consists of the long non-coding (lnc) RNA Gm15663 and the protein coding gene Tmtc2, an important regulator of endoplasmic reticulum (ER) Ca2+ flux. Gm15663/Tmtc2 are also upregulated when glucagon predominates insulin, such as during fasting, after high-fat or western diet and in ob/ob mice. Conversely, Gm15663/Tmtc2 are downregulated during feeding, when insulin predominates glucagon. This transcriptional regulation is mediated in part by the transcription factor Foxo1. Foxo1 binds with two promoter-distal enhancers, creating a spatial chromatin looping interaction with Gm15663/Tmtc2 promoter, which is involved in transcriptional regulation of this gene pair. Knockdown of Tmtc2 in hepatocytes results in a higher cytosolic Ca2+ level and lower ER Ca2+ level. This renders the cell more sensitive to both metabolic and ER stress, and impairs cell respiration, which is in part mediated by the increased cytosolic Ca2+ level and the activation of CaM kinase II (CamKII). In a unique mechanism, Gm15663 and Tmtc2 also cis-regulate the transcription of each other to maintain the homeostasis of their own expression levels. Thus, Gm15663/Tmtc2 gene pair are dynamically regulated by glucagon and insulin and serve as novel regulators of cell sensitivity to metabolic stress. This gene pair is also linked to a common subtype of T2D in GWAS studies, making them potential targets to treat or modify the development of T2D in these individuals. Disclosure X.Liu: None. C.Kahn: None.