D91-03 Multi-layered Single-cell Genomics Dissects Germline and Somatic Genetic Mechanisms Driving COVID-19 Severity

Rationale Genome-wide association studies (GWAS) have revealed that host genetic background critically influences COVID-19 severity. In addition to germline variants, emerging evidence suggests that somatic mutations, such as mosaic chromosomal alterations (mCAs), contribute to severe outcomes. However, how these inherited and acquired genetic variants shape immune cell function at the molecular level remains unclear. To address this gap, we aimed to delineate the molecular mechanisms linking host genetic variation to immune-cell dysfunction by integrating germline and somatic mutation profiles with single-cell transcriptomic and immune-repertoire data. Methods We constructed immune single-cell atlas with multi-layer omics from 235 Japanese individuals, including COVID-19 patients and healthy subjects. Dataset consists of single-cell transcriptomics from >1.5 million peripheral blood mononuclear cells, host genetics including germline and somatic genetic variations and plasma proteomics. Utilizing whole-genome sequencing (WGS) and SNP genotyping, we detected individuals with mCAs and loss of the Y chromosome (LOY). Results We examined single-cell expression quantitative trait loci (sc-eQTL) effects of the COVID-19 risk germline variants to understand their underlying regulatory mechanisms. The sc-eQTL analyses revealed that COVID-19-associated variants (such as the IFNAR2 variant rs13050728) had context (COVID-19)-specific and cell-type (monocyte)-specific eQTL effects. In addition to single germline mutations, differential expression analyses captured context-specific and monocyte-specific transcriptional and proteomic alterations linked to polygenic risk of severe COVID-19. Collectively, these multimodal and integrative data analyses demonstrated that host genetic risk for severe COVID-19 is preferentially enriched within innate immune lineages. By projecting pre-identified somatic mutations from genomic data onto single-cell RNA-seq, we successfully distinguished somatically mutated cells from normal cells at single-cell resolution within each individual. We observed highly cell type- and context-specific accumulation of mCAs. In one COVID-19 patient, recombinant antibodies derived from clonally expanded B cells harboring 17q gain showed no reactivity to major SARS-CoV-2 antigens, indicating a loss of neutralizing capacity and a functional defect in adaptive immunity. LOY was enriched in myeloid cells. Differential transcriptomic and abundance analyses between LOY and normal monocytes, as well as between LOY-high and LOY-low COVID-19 patients, suggested that LOY monocytes may drive immune dysregulation through altered T-cell composition. Conclusions Our single-cell-level deconvolution approach, leveraging genomics and single-cell transcriptomics data, provides novel insights into how diverse genetic variations shape immune cell function. This integrative multi-omics framework advances the mechanistic understanding of COVID-19 severity and provides a foundation for precision immunogenomics in infectious diseases. This abstract is funded by: None