The Genomic Advances in Sepsis GAinS RNA seq

ABSTRACT: Sepsis is a clinically defined syndrome for which there are no targeted treatments, in part due to extensive variation in the individual host immune response. From transcriptomic data, we identified disease endotypes, sepsis response signatures (SRSs), that are associated with differential early mortality (Davenport et al, Lancet Respiratory Medicine, 2016; Burnham et al, American Journal of Respiratory and Critical Care Medicine, 2017) and response to treatment in a clinical trial (Antcliffe and Burnham et al, American Journal of Respiratory and Critical Care Medicine, 2018) highlighting the value of molecular data for developing a personalised medicine approach to treatment in sepsis. We identified and validated these SRS groups from microarray gene expression data in a cohort of 550 patients with sepsis as a result of community acquired pneumonia (CAP) or faecal peritonitis (FP). These patients are a subset of the Genomic Advances in Sepsis (GAinS) study, an established biobank of >1,300 patients with samples for functional genomics and detailed clinical information. Genotyping and proteomic datasets are currently being completed for the full cohort and plasma nucleic acids from the majority of CAP patients have been sequenced to identify the causative pathogen. We will be preforming RNA sequencing to establish informative biomarkers of the SRS endotypes and assign SRS group membership across the GAinS cohort. This will allow us to address the following research questions: (1) What are the most robust and informative SRS transcriptomic biomarkers? We have a 7-gene signature for assigning SRS group from microarray data but we need to test its robustness across gene expression platforms (notably RNA-seq as the most commonly used approach) and therefore propose to perform RNA-seq on a subset of patients with microarray data (n=100) allowing selection of the most informative gene signature set. (2) What is the nature of differential gene expression in the SRS endotypes? RNA-seq data will allow us to determine alternatively spliced isoforms and noncoding RNAs that may be specific to SRS endotype and are not currently known from our microarray analysis. It would also allow determination of whether there are transcriptomic signatures that differentiate the pathogens causing sepsis. (3) How dynamic is SRS membership over time? By profiling and assigning SRS membership for bio-banked samples taken at different times after admission we can determine this, with implications for the feasibility of therapeutic intervention if endotype states are reversible. (4) Which genetic variants are associated with SRS group membership and can a polygenic risk score for SRS be determined to allow SRS identification in other external datasets? By assigning SRS across the full cohort we will be maximally powered for such analysis. (5) Are there potential protein biomarkers of SRS endotype? Assignment of SRS group membership will allow us to combine this information with proteomic data being generated for the full cohort. We therefore propose performing RNA-sequencing on 922 samples (run on 3 flow cells on the NovaSeq), comprising 100 paired samples with microarray data (aim 1) and 822 samples from 603 patients enrolled in GAinS for whom transcriptomic data has not yet been generated (enabling aims 2-5). This project will be led by Katie Burnham (Davenport group) and Emma Davenport, and Julian Knight at the University of Oxford (Chief Investigator on the GAinS study) to build on an established and successful collaboration that has led to the work described above on transcriptomics and eQTL mapping in the context of sepsis. In terms of broader impact, this work would complement two aspects of the BioAID project; firstly, the identification of SRS groups at an earlier stage in disease progression and secondly the identification of transcriptomic profiles associated with aetiology such as causative pathogen (Prelim I3253, Sandhu and Davenport). It would also lay the foundation for a proposed work package that Emma would lead as part of a Wellcome Collaborative Award application (Julian Knight lead PI). This collaborative grant would focus on immunophenotyping the endotype associated with greater mortality and therefore determining the optimal gene set to robustly assign SRS group membership will be a necessity. This data is part of a pre-publication release. For information on the proper use of pre-publication data shared by the Wellcome Trust Sanger Institute (including details of any publication moratoria), please see http://www.sanger.ac.uk/datasharing/

PROVIDER: EGAS00001003772 | EGA |

REPOSITORIES: EGA