circRNA Signatures Distinguishing COVID-19 Outcomes and Acute Respiratory Distress Syndrome: A Longitudinal, Two-Timepoint, Precision-Weighted Analysis of a Public RNA-Seq Cohort

TL;DR

This study identifies circular RNA patterns that distinguish outcomes in severe COVID-19 patients and those with ARDS using a two-timepoint RNA-Seq analysis.

Contribution

A novel precision-weighted, longitudinal analysis method for identifying stable circRNA signatures in severe COVID-19.

Findings

Nine significant and four suggestive circRNA candidates distinguished non-survivors from survivors.

Some circRNAs showed differences between survivors and ARDS controls.

Combining data from two timepoints improved detection of stable circRNA signals.

Abstract

Background: Although circular RNAs are increasingly implicated in host responses, their longitudinal behaviors to predict outcomes in severe COVID-19 remain unclear. The purpose of this study is to distinguish the circRNA signature associated with COVID-19 outcome. Method: Public total RNA-seq data from GEO (GSE273149) were used to assess circRNA differences among COVID-19 non-survivors, COVID-19 survivors, and patients with acute respiratory distress syndrome (ARDS) serving as severity-matched disease controls at two timepoints: Early (Day 3) and Late (Days 7 to 10). Differential expression was assessed after quality filtering, with the results reported as significant (FDR < 0.05) or suggestive (0.05–0.10); |log2FC| ≥ 1 was used as a guide for interpretation. Early and Late effects were combined using a two-timepoint, precision-weighted approach to prioritize time-consistent signals. Results: A distinction between non-survivors and survivors was observed, with nine significant and four suggestive candidates identified in the combined analysis; in addition, some candidates indicated a difference between survivors and ARDS controls. Early and Late effects primarily occurred in the same direction, and several circRNAs that were borderline at one timepoint became significant when the two timepoints were combined. Conclusion: This time-resolved, precision-weighted analysis of public RNA-seq data reveals stable circRNA differences between key clinical groups (patients with severe COVID-19 and those with ARDS), improving detection and interpretability relative to single-timepoint tests and yielding a concise set of candidates suitable for mechanistic follow-up and potential biomarker development.

Figures