Attribution begins where silos end.
BG4K builds a research method to detect when independent, co-measured streams share a candidate shared driver — shared latent structure (hypothesis) — and when they don’t — without forcing a story on noise. Not a medical service.
What it is
CSSR/DAS (Cross-System Signal Recognition / Dynamic Adaptive Staging) is a research framing for coordinated multi-silo anomalies across fragmented systems — hypothesis-generation for when independently observed, co-measured streams share a candidate shared driver, and saying so honestly when they don’t.
It is label-free by design: evidence that matters often lives across systems that were never built to talk to each other. The work is about shared-driver detection across silos — not a univariate dashboard, not a single-sensor “winner,” and not a promise that every dataset yields a story.
Research framing only. Not a diagnosis product, not a medical device, not an upload/run service, and not converge-as-a-service. Method honesty after admit — not a public run API.
Who it’s for
Researchers and lab stewards only. US-first · invite-only · ≤ 10 admitted. Path:Apply → manual admit → walkthrough. Bring own research or already de-identified research. Named collaborators preferred. Not a patient funnel, not clinical intake, and not an open beta for everyone.
Research trail
Inspectable prior art and research surface — not a claim that these papers validate the engine. Not clinical advice and not a biomarker product.
ORCID 0009-0005-3995-0251 — Lejla Tetrick.
Lejla Tetrick’s public preprints explore one problem class from different angles: when evidence that matters is split across systems that were never designed to talk to each other, attribution fails — and when independent streams do not share a candidate shared driver, honesty requires saying so. Together they form an inspectable research surface, not a product brochure and not a claim that CSSR/DAS is “validated” by any single paper. These deposits are research materials for critique — not clinical advice, not a biomarker product, and not disease-product marketing.
Unveiling the Global Disparity in Chronic Mycotoxin Exposure
Zenodo preprint — mycotoxin disparity. DOI 10.5281/zenodo.15660663.
View DOI / ZenodoA Novel miRNA Triadic Biomarker Panel for Chronic Inflammatory Response Syndrome
Zenodo preprint — CIRS miRNA panel. DOI 10.5281/zenodo.17205924.
View DOI / ZenodoThe Retina as a Convergence Sentinel for Cross-System Signal Architecture
Zenodo preprint — retina as convergence sentinel. DOI 10.5281/zenodo.20278639.
View DOI / Zenodo
Full list on Publications.
Methods receiptMH370 oceanographic package — DOI10.5281/zenodo.20263776— methods / reproducibility for expert critique only. Not a location claim, not a claim to have located the aircraft, and not official findings.
Next step
Apply to the research pilot, or contact for a collaborator walkthrough. No upload box. No paid checkout on this site.