An Integrated Host-Microbe Gene Classifier to Predict SARS-CoV-2 and Severe Disease in Children with Respiratory Viral Coinfections

PROJECT SUMMARY Respiratory viral coinfections (RVCIs) are more common in children than adults and have become increasingly important due to the emergence of SARS-COV-2. However, the distinction between coinfection and codetection remains unclear. Further, the association between finding multiple viruses using our current clinical methods and the effect on clinical outcomes is nebulous. Understanding positive viral testing has become crucial with the ongoing spread of SARS-CoV-2 and the treatment algorithms that are expensive or deleterious to patients with other viruses. Currently, clinicians struggle to identify the dominant virus inducing the host immune response in RVCIs in children. Our group has developed gene classifiers to identify adults with SARS-CoV-2 compared to other viruses using nasopharyngeal swabs. Further, we are the first to develop a host/microbe classifier for pediatric patients on the ventilator that will distinguish lower respiratory tract infections using lower airway sampling. Here, our objective is to identify patient features, coinfecting viruses, microbial contributions, and host responses that enhance disease severity in SARS-CoV-2-infected children. As the only pediatric hospital in the state, Arkansas Children's Hospital is an ideal site for studying SARS-CoV-2 RVCIs. We aim to leverage our unique multi-institutional collaborative team with extensive experience applying metagenomic next-generation sequencing to simultaneously evaluate viral and host genetic material from clinically obtained nasopharyngeal specimens. This approach identifies host-virus interactions and allows us to assess their impact on immune responses and disease severity during RVCIs. We hypothesize that a combination of patient characteristics, viral features, and host immune responses will predispose a child to more severe disease. We also expect to identify an immunologic fingerprint for SARS-CoV-2 that can be used to identify it as the "infecting virus" in children with codetections. The impact of this study is significant, and the multidisciplinary team that this proposal brings together is experienced. By employing epidemiologic, -omic, and computational approaches, we will identify immunologic fingerprints of SARS-CoV-2 infections in children, which can help identify the "infecting virus" when multiple viruses are detected. Further, this study will provide distinction regarding the clinical implications of codetections of viruses, including SARS-CoV-2, in children. The findings will support future clinical trials evaluating treatment options based on the immunologic fingerprints that we identify.