The DNA virome varies with human genes and environments
Summary
This population-scale study used whole-genome sequencing (WGS) from >900,000 samples (UK Biobank, All of Us, SPARK) to quantify DNA viral fragments in blood and saliva by aligning non-human reads to a curated panel of 31 viral genomes. Viral DNA was generally rare in blood but common and often abundant in saliva for herpesviruses. Viral DNA load differs markedly by age, sex (higher in men), time of day and season, smoking, and genetic ancestry. Genome-wide association studies identified dozens of human loci—most prominently the MHC/HLA region and the ERAP1–ERAP2–LNPEP locus—that shape viral DNA load for EBV, HHV-6B, HHV-7, anelloviruses and others. Ancestry-specific variants (notably the Duffy-null ACKR1 allele) explain a substantial fraction of ancestry differences in EBV and HHV-7 prevalence. EBV strain typing revealed HLA-B*08:01 has opposite associations for EBV type 1 versus type 2. Mendelian randomisation using non-MHC genetic instruments suggests EBV DNA load causally increases Hodgkin’s lymphoma risk (but not multiple sclerosis), and smoking robustly increases EBV load while decreasing HHV-7 load.
Key Points
- WGS from 490,401 UKB blood, 365,918 AoU blood, 48,899 AoU saliva and 12,519 SPARK saliva samples were screened for 31 DNA viruses and retroviruses.
- Herpesvirus DNA is often absent in blood WGS (sparse sampling of latently infected cells) but common and bimodally abundant in saliva, indicating intermittent lytic shedding.
- Viral DNA load correlates with age, sex (generally higher in men), circadian and seasonal patterns, smoking and genetic ancestry; patterns are virus-specific.
- GWAS identified 82 loci for blood viral load (45 for EBV), with the MHC region contributing a large fraction of heritability; ERAP1/ERAP2 and other immune genes also implicated.
- The Duffy-null (ACKR1 rs2814778) genotype explains a major part of African-vs-European differences in EBV and HHV-7 prevalence.
- HLA-B*08:01 shows EBV type-specific effects: protective for type 1 but associated with increased type 2 prevalence.
- Mendelian randomisation indicates EBV DNA load likely causally increases Hodgkin’s lymphoma risk; smoking raises EBV load and may contribute to EBV-positive Hodgkin’s risk.
- Findings suggest interventions that lower EBV load (for example, antivirals) could potentially reduce Hodgkin’s lymphoma risk—warranting experimental follow-up.
Context and relevance
This is one of the largest population-scale virome analyses to date, leveraging routine WGS data to profile resident DNA viruses across ages and ancestries. It connects host genetics, environment and behaviour to variation in persistent viral DNA load, highlighting the MHC/HLA system and peptide-processing genes as central determinants. The work informs epidemiology (ancestry and smoking effects), immunology (HLA and ERAP variation affecting antigen presentation), and oncology (EBV load as a probable causal factor for Hodgkin’s lymphoma). For researchers and clinicians, these results point to measurable biomarkers (viral DNA load) that are shaped by host genetics and modifiable exposures and that have potential downstream clinical consequences.
Author style
Punchy: big cohorts, clear genetic signals, and concrete clinical links. The study isn’t just descriptive — it maps specific HLA and peptide-processing haplotypes to viral load differences, shows ancestry and environment matter, and uses genetics to test causality for cancer risk. If you work on host–pathogen genetics, immune epidemiology or virus-associated malignancy, the detailed loci and MR results are worth a close read.
Why should I read this?
Quick version: they combed through near‑million genomes and found that who you are (genes), how you live (smoking, time of day) and where you come from (ancestry) all sway the tiny amounts of viral DNA floating in your blood and spit. Some of those DNA-load differences look like they actually drive disease risk—Hodgkin’s lymphoma jumped out—so this isn’t just trivia. Read it if you want the fast route to which genes and exposures actually matter for common latent viruses.