Preprint Population-scale sequencing resolves correlates and determinants of latent Epstein-Barr Virus infection 2025 Nyeo et al

[Andy]

Abstract​

Epstein-Barr Virus (EBV) is an endemic herpesvirus implicated in autoimmunity, cancer, and neurological disorders. Though primary infection typically resolves with subclinical symptoms, long-term complications can arise due to immune dysregulation or viral latency, in which EBV DNA is detectable in blood for decades. Despite the ubiquity of this virus, we have an incomplete understanding of the highly variable responses to EBV that range from asymptomatic infection to a trigger for severe disease.

Here, we demonstrate that existing whole genome sequencing (WGS) data contains ample non-human DNA sequences to reconstruct a molecular biomarker of latent EBV infection consistent with orthogonal phenotypes, including viral serology. Using the UK Biobank (n = 490,560) and All of Us (n = 245,394), we uncover reproducible complex trait associations that nominate latent blood-derived EBV DNA as a respiratory, autoimmune, and cardiovascular disease biomarker.

Further, we evaluate the genetic determinants of persistent EBV DNA via genome-wide and exome-wide association studies, uncovering protein-altering variants from 147 genes. Single-cell and pathway-scale enrichment analyses implicate variable antigen processing and presentation as a primary genetic determinant of latent EBV persistence, with gene programs expressed in B cells and antigen-presenting cells. Using predicted viral epitope presentation affinities, we implicate genetic variation in MHC class II as a key modulator of EBV DNA persistence.

Our analyses demonstrate how existing WGS data can derive novel molecular biomarkers, which may generalize to dozens of viruses comprising the blood virome.

Open access

 

[jnmaciuch]

This is really interesting--if HLA alleles come up in DecodeME, my first question would be whether those alleles are known to confer functional susceptibility to latency (or just viral susceptibility in general).

Associations with EBV have definitely come up repeatedly for other chronic illnesses, a few of which they note in the intro. Something I'd be really curious about is if the known HLA alleles associations with those other illnesses are partially attributable to viral susceptibility (rather than the role of T cells in maintaining the disease as is always assumed).

 

[tralfamadorian97]

This looks like a very cool paper, illustrating the power of whole genome sequencing data. It is impressive that the authors were able to use a new technique to reproduce previously known associations between disease traits and EBV status.

I also wonder if it will be possible to perform analogous virome studies using the data generated from SequenceME. The key difference is that this paper uses data derived from blood samples, whereas the SequenceME data will be derived from saliva.

 

[jnmaciuch]

I also wonder if it will be possible to perform analogous virome studies using the data generated from SequenceME. The key difference is that this paper uses data derived from blood samples, whereas the SequenceME data will be derived from saliva.

It’s theoretically possible! For EBV at least, since it can infect and establish latency in the epithelial cells of the mouth as well as circulatory B cells. What I don’t know for sure is if you’d be able to get a sufficient quantity of infected cells from a sample swab.

If you’re simply trying to detect the presence of any EBV proviral DNA then my guess is that it could be enough, but you probably couldn’t do anything more detailed than that.

 

[Jonathan Edwards]

I am unimpressed by an abstract that says nothing specific about what was found - just general guff about how great they think their study is.

There has been an obsession with MHC diversity encoding disease susceptibility through antigen presentation of peptides in grooves, and vast amounts have been written on EBV under that assumption. I am not aware of much to support it. The most obvious link between HLA and disease is B27. That has a major protective effect against several viruses as I understand it so presumably not through presentation of any one special peptide.

The link to DR4 in RA is most likely related to citrulline binding, although nobody is quite sure how. Citrullination affects a lot of proteins so it is not an issue of any specific peptide from a specific organism.

DQ associations are interesting because rather few T cells show DQ restricted TCR-nased responses. Nobody knows where DQ is fitting in. Danny Altmann wrote a couple of reviews on this.

The situation for EBV latency is complicated because the latency is mostly DNA incorporated into B cell clones being handed on from clone to clone over time. How that relates to diseases like MS is puzzling but maybe the most likely role for MHC would be in influencing ongoing deletion of such clones by CD57+ T cells. The DQ allele expressed by the B cell might alter its susceptibility to deletion and not necessarily through presentation of a specific peptide since DQ can also contribute to more 'innate' interactions.

I have no had time to look at the full text but I would worry that this a paper that finds what it set out to find based on rather blinkered paradigms.

​

​