Data, methods and limits
See what is calculated and which conclusions the data does not support.
Which cell models
The fibroblast mode is limited to human fibroblasts. Oncology calculations target established human cell lines in 2D monoculture with a completed model record. Comparisons require a genuinely comparable nonmalignant model. Arbitrary tissues, primary cells, organoids and clinical samples are not validated scenarios.
How calculations work
Paired ratios are calculated within independent replicates, not across all wells. SD describes spread; SEM is not substituted for SD. GR requires baseline and control counts. Division and death events require direct observation. Bliss and HSA are different combination reference models, not proof of clinical synergy.
What has been checked
Arithmetic and reproducibility checks, synthetic controls and published-data comparisons exist for selected modes. This is not overall clinical or biological validation. Full division/death event accounting is currently checked on synthetic data. Some author tables remain in the local validation package pending redistribution-rights clarification.
Scientific sources and catalog
The PRISM catalog is a historical snapshot of published responses in nine lines, with source attribution. Ranking reflects measured signals in those conditions; it does not predict effects in patients or a new cell line.
Where data resides
A table is sent over HTTPS to a server in Russia for calculation and is not written to a CellFate database. Projects and autosaves reside in your browser. Requests are limited to 8 MB. Clearing browser storage removes the local library; export JSON. Projects from 127.0.0.1 do not automatically move to the domain. Do not upload patient personal data.