Leveraging multi-omics to maximize the scientific value of pediatric sepsis biorepository and advance patient endotyping. PROJECT SUMMARY: Sepsis is major pediatric health problem and kills more children than cancer in the U.S each year. Primarily driven by a dysfunctional host response to an infection, a subset of patients with persistent or progressive multiple organ dysfunctions disproportionately contribute to sepsis morbidity and mortality. Yet, there are no disease modifying therapies currently available beyond early antibiotics and organ support. Biological heterogeneity among patients has significantly impeded scientific progress and advances in patient care. Although precision medicine approaches have been used to begin to sift through patient-level differences, we fundamentally lack a comprehensive understanding of disease mechanisms. Thus, there is a crucial need to maximize the use of existing pediatric sepsis biorepositories to unravel causal pathways, facilitate rapid identification of biologically relevant patient subclasses more likely to benefit from targeted therapies. To bridge this gap, we seek to utilize state-of-the-art multi-omics approaches to explore the scientific value of our vast collection of biospecimens from critically ill children with sepsis. While gene-expression profiling has been used to identify biologically relevant endotypes, it is increasingly evident that interrogation of single layer of molecular data is likely insufficient. Recent studies suggest that the epigenomic changes, including differential DNA methylation at CpG islands, closely regulate gene-expression in human sepsis. Through this phased innovation award, we seek to determine whether integrated analyses of methylomic and transcriptomic datasets at scale can provide a comprehensive understanding of mechanisms and inform patient endotyping. We further seek to determine whether clinical data linked with patient biospecimens can be used to predict endotype membership, with operational implications for predictive enrichment in future clinical trials. Milestone-driven developmental activities in the R21 phase will focus on stringent quality control of DNA and RNA samples within our biorepository to determine suitability for high throughput methods. We will then generate pilot DNA methylation profiling and RNA sequencing data for study planning and to demonstrate feasibility of integrated analyses. In the R33 phase, we will scale efforts to generate robust methylomic and transcriptomic datasets. We will leverage the bioinformatic capabilities of the investigator team to derive and validate novel multi-omic endotypes and determine their clinical significance. Finally, we will develop a classifier model to predict endotype membership using clinical data within the derivation cohort and test its generalizability in a large electronic health record-based dataset of >15,000 critically ill children with sepsis and multiple organ dysfunctions. Through the successful execution of this proposal, we seek to generate a rich dataset to drive future mechanistic research in human sepsis and develop an actionable framework for rapid and equitable identification of pediatric sepsis subclasses who may benefit from targeted sepsis therapeutics.

Leveraging multi-omics to maximize the scientific value of pediatric sepsis biorepository and advance patient endotyping. PROJECT NARRATIVE: Children with sepsis—a life threatening condition due to an abnormal patient response to an infection— represent a mixed group with varying underlying biology. Our incomplete knowledge of disease mechanisms within subgroups of septic patients has prevented the development of the right drugs for the right patients. To address this problem, we seek to use our vast collection of blood samples from children with sepsis to find out whether an understanding of how genes are turned on and off can be used to improve sepsis outcomes.