Which microbes are really in tumors?
Evidence for the cancer microbiome varies widely. Some organisms have been cultured from tumors and shown to affect the disease. Others are faint sequencing signals that fade under reanalysis. This atlas grades every reported association between a microbe and a cancer type by the strength of the evidence behind it, from live cultivation to low-confidence detection.
More filters
Verification status
Every figure below carries its denominator. Unaudited means no researcher has reviewed that entry yet — it is a statement about our review coverage, not about the quality of the underlying paper. Automated consistency checks are listed separately because a machine check is not a human review and never marks an entry verified.
What the graded data shows
These figures are computed from the graded entries and follow the filters you set in the atlas above. They describe how much has been reported and curated — not microbial abundance, effect size, causality, or biological co-occurrence. Each chart has a table equivalent.
How the evidence is graded
Each association is graded by the strongest evidence it has earned. Every entry shows its methods, its reproducibility, and its caveats, including counter-evidence and failed replications. Low-confidence results are labeled rather than hidden, so uncertainty in the field stays visible.
References
A numbered bibliography generated from the atlas source records. References are deduplicated across entries, sorted alphabetically, and linked when DOI or PubMed metadata is available.
How the atlas is built
Four principles guide how associations are graded and presented.
Evidence over enthusiasm
Each association is graded on what has been demonstrated: cultivation, in-situ localization, or reproduction across cohorts. A striking headline does not raise its grade.
Contamination is tracked
Low-biomass tumor sequencing is prone to contamination. Each entry records how decontamination was handled, and that determines whether it can rise above the lowest tiers.
Caveats stay with the claim
Counter-evidence, failed replications, and retractions appear next to the finding, not in a footnote. A strong-looking entry can still carry a warning.
Provenance you can audit
Each entry names its curator, its verification status, and its primary papers, so any grade can be traced back to its source.