February 21, 2023

Predicting Clinically Significant Microbial Biomarkers via Differential Sequence Assembly

5th Call, TPdF

Short Summary

Sepsis is the body’s extreme response to a microbial infection, resulting from a complex interaction between the infection and the patient. For such conditions, general-purpose treatment strategies can fail to successfully treat the patient, instead requiring a personalized treatment approach. Analyzing DNA sequences from sepsis-causing microbes sampled from patients opens up a window of opportunity to non-invasively study microbe-patient interactions. In this study within the framework of the Personalized Swiss Sepsis Study (https://sepsis-network.ch/), we strive to improve the mechanistic understanding of sepsis and to predict novel DNA biomarkers to aid in forecasting adverse patient outcomes.

Goals

Our goals in this study within the framework of the Personalized Swiss Sepsis Study are 1) to improve the mechanistic understanding of sepsis by studying changes in the populations of microbes present during infection and 2) to predict novel DNA biomarkers (i.e., sequence patterns) to use alongside vital sign measurements for forecasting adverse patient outcomes. In the process, our third goal is 3) to develop an optimal experimental protocol for future biomedical DNA analysis studies on microbes to guide this study and future studies of its kind. This proposal collaborates with clinical experts, ensuring that all steps taken throughout the study will maintain clinical relevance and present feasible treatment opportunities.
We will uphold principles of accessibility and reproducibility throughout the study to ensure that other groups can benefit from our developed methods. Strict separation of patient data-dependent components from the rest of the project allows following data privacy guidelines.

Significance

48.9 million cases of sepsis and 11 million deaths were reported worldwide in 2017 by the WHO. With a mortality rate of up to 26% and a rapidly decreasing chance of survival within hours when left untreated, improving strategies to ensure timely detection and treatment of sepsis is a major goal for the medical research community. We are confident that developing our DNA analysis framework into this new application area will enable clinically relevant analyses across large cohorts with direct translation into patient benefit. We further expect to spark new and challenging theoretical research questions that help further advance the field of computational biomedicine.

Background

After decades of research into complex clinical syndromes, it is now clear that personalized treatment is an indispensable tool for ensuring better treatment success in any health care system. Sepsis is the body’s extreme response to a microbial infection. It results from a complex interaction between the infection and the patient. For such conditions, general-purpose treatment strategies can fail to successfully treat the patient and jeopardize the treatment prospects of future patients by promoting the evolution of treatment-resistant strains. Analyzing DNA sequences from sepsis-causing microbes sampled from patients opens up a window of opportunity to non-invasively study microbe-patient interactions. With the recent development of highly optimized algorithms for jointly analyzing these data from large patient cohorts and the expanding capabilities of high-performance computing infrastructures, studying these interactions is now in reach. Applying our experience from years of development on the MetaGraph project by the Biomedical Informatics Group at ETH Zürich, we are now setting out to leverage its powerful capabilities to develop novel quantitative techniques for improving sepsis management.