A. EBV-infected vs normal B cells
Public EBV-infected B cell lines (GEUVADIS) scored against normal B cells (DICE). Success criterion: the Disruption Index separates infected from normal cells on held-out samples.
EBV program
Why EBV first
Epstein–Barr virus infects most adults and persists in B cells for life. In 2025 a Stanford team reported that EBV turns a small share of B cells into drivers of lupus autoimmunity — a network-level cascade the RNA QR code is built to read. EBV-infected B cell lines are also among the most sequenced human cells, so the first readout can be run on public data alone.
Source
Younis et al., Science Translational Medicine (2025); Stanford Medicine news release, Nov 12, 2025. Related indications under study by others: multiple sclerosis, rheumatoid arthritis, Crohn's disease, Sjögren's.
I. First readout
Three analyses on existing public data, success criteria fixed before we look
B. Lupus patients over time
Public lupus blood datasets (e.g. GSE65391, GSE88884) scored against controls and against recorded disease activity. Success criterion: the Disruption Index tracks disease activity.
C. Approved-drug screen
Public drug-response profiles (Connectivity Map) scored for restoring the EBV pattern. Success criterion: about twelve approved drugs ranked for a second stage.
No new patients and no laboratory costs: all three analyses run on public data, with held-out samples and pre-registered criteria. The pre-registrations and results will be posted on this page as they are completed. RETROSPECTIVE · PUBLIC DATA (GEO/TCGA) · PRELIMINARY
II. What the program does and does not claim
Measures and scores; does not diagnose or treat
The platform compresses a whole transcriptome into roughly eighty modules and reports two numbers: how far a disease state has moved a cell's gene network (the Disruption Index) and how far an intervention moves it back (the Restoration Index). It identifies candidate compounds for further study. It is not a diagnostic and it does not treat any condition.