Identifying Associations Between Genomic and Clinical Features of SARS-CoV-2 in the New Jersey Area During the Early Stages of the COVID-19 Pandemic

• Activity Number: 514 - Advancements in Multi-Omics Integration Techniques
• Type: Topic Contributed
• Date/Time: Thursday, August 11, 2022 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Tyson Dawson* and Ali Rahnavard
• Companies: The George Washington University and George Washington University
• Keywords: COVID-19; SARS-CoV-2; Transcriptomics; RNA-Seq; Discovery; Epidemiology

Abstract

We developed statistical methods to test how the viral genome of SARS-Cov-2 (the causative agent of COVID-19) evolved and affected patients’ health outcomes during the early phase of the US pandemic in the state of New Jersey. We collected viral genome data and metatranscriptomic data from 116 covid-positive patients across New Jersey from March - June, 2020. We use viral genomic variants, microbial compositional differences, and host differential gene expression along with a novel clustering approach, omeClust, to test for the association of these discrete omics features with clinical variables. Significant associations were identified between viral genome regions and clinical features such as nsp10’s association with the duration patients spent on a ventilator. We highlight the key role that bacterial community alpha diversity plays in disease outcome including the development of fever and shortness of breath. Finally, we identify host transcriptional pathways that are dysregulated in patients with COVID-19. This integrative method allows for a holistic look at the pandemic and can inform our approach to understanding the perpetually evolving COVID-19 situation around the world.