Evidence Tracing

Honest introspection: recency-weighting in EBV persistence evidence

Where the recency bias is defensible. Spatial and single‑cell methods answer the question 'in which cells' directly — they localize viral programs to identified cells in intact tissue, which bulk serology fundamentally cannot. The 2025 IM tonsil spatial atlas mapped EBNA2+ B cells near epithelium and LMP1+ B cells deeper in tissue, while the 2025 MS CSF study (47,217 T cells, 18,149 CD8+) clonally resolved EBV-reactive GZMK+/GZMH+ CD8 effectors 2025 MS CSF EBV TCR multi-omics study. Serology cannot distinguish reservoir from effector cells — a limitation, not a flaw.

Where the bias is a genuine blindspot. Older bulk studies carried mechanistic weight that recent spatial work has not replaced. The 2007 monocyte-dissemination study used sorted-cell QC-PCR (78 donors, 95–99% purity) to quantify EBV carriage in circulating monocytes and show CCR2-blockade reduced EBV+ keratinocytes ~5-fold — causal perturbation evidence that modern descriptive spatial atlases lack 2007 CCR2 blockade monocyte transmigration. Similarly, conditional EBNA3A systems (2003) established that EBNA3A is non-substitutable for LCL growth maintenance — mechanistic necessity that no recent single-cell descriptive study provides 2003 EBNA3A conditional LCL study.

What would change my weighting. If recent spatial studies repeatedly fail to replicate older bulk findings in vivo (e.g., no monocyte EBV signal in tissue-resident myeloid cells), I should downgrade the older work. Conversely, a single high-quality perturbation experiment in an intact system outweighs many descriptive atlases. The correct stance is method-complementarity, not recency: descriptive localization (new) + causal perturbation (old) = convergent validity.

Confidence note: High confidence that memory B cells are the canonical reservoir; moderate confidence that monocytes are a real but secondary transient reservoir (bulk-sorted evidence is strong but ex vivo models limit causality). Serology-only studies of EBV persistence are indeed weak for cell localization — my prior against them is justified on that axis, but I should not extend it to older bulk molecular (PCR/sorted-cell) studies, which are a different evidence class.