Orbiting a Genomically Characterized Organoid Biobank in 3D
This 3D companion renders the same population-scale biomarker-discovery workflow as an orbitable point cloud: each sphere is one patient-derived organoid line, positioned along a genomic-profile axis, colored by its mutation cluster (TP53, KRAS, MSI-high, or a wildtype panel), and lifted along the vertical axis by its measured drug-response magnitude. As the correlation-strength control increases, a live regression plane — fitted from the same candidate-cluster-vs-response data driving the on-screen statistics — visibly tilts, and once the correlation clears both the significance and effect-size stage, the candidate cluster is ringed with an emerald halo. A second, smaller point cloud floats further back in depth, representing an independent validation cohort whose attenuated correlation either confirms or undercuts the candidate biomarker, exactly mirroring the five-stage discovery-to-validation pipeline of the 2D original.
Why a population-level correlation is not a treatment recommendation
Every number on this page is illustrative of research methodology, not of any real biobank's dataset. A biobank-scale genomic correlation — however strong the regression plane's tilt looks — is a hypothesis-generating signal that must survive multiple-testing correction, independent organoid-cohort replication, and eventually correlation with real clinical outcomes before it could ever inform an individual treatment decision. The "candidate biomarker" and "validation" stages in this simulation exist specifically to make that distinction visible: a strong correlation in the discovery cloud is the start of a multi-year validation pipeline, not its conclusion.